No. 112 NAC) VN SEN SORE ol [1 NINERS Ie} Vital and Health Statistics Health Service Areas for the United States November 1991 CENTERS FOR DISEASE CONTROL Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggestion Citation Makuc DM et al. Health service areas for the United States. National Center for Health Statistics. Vital Health Stat (2)112. 1991. Library of Congress Cataloging-in-Publication Data Health service areas for the United States. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research; no. 112) (DHHS publication; no. (PHS) 91-1386) By Diane M. Makuc . . . [et al.]. “October 1991.” Includes bibliographical references. ISBN 0-8406-0452-1 1. Health service areas — United States. 2. Health surveys — United States. I. Makuc, Diane M. Il. National Center for Health Statistics (U.S.) Ill. Series. IV. Series: DHHS publication; no. (PHS) 91-1386. [DNLM: 1. Health Care Rationing — United States — statistics. 2. Health Resources — supply & distribution — United States — statistics. 3. Health Services Needs and Demand — United States — statistics. W2 A N148vb no. 112] RA409.U45 no. 112 [RABO4] 362.1'0723 s—dc20 [362.1'0973] DNLM/DLC for Library of Congress 91-30831 CIP Vital and Health Statistics Health Service Areas =, “ing Series 2. Mek gy Data Evaluation and Methods Research No. 112 The objectives of this report are to document methods used to identify health service areas for the United States and to describe and evaluate these areas. A health service area is defined as one or more counties that are relatively self- contained with respect to the provision of routine hospital care. Service areas that include more than one county are characterized by travel between the counties for routine hospital care. EE EF TR EE Im, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control National Center for Health Statistics Hyattsville, Maryland November 1991 DHHS Publication No. (PHS) 92-1386 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Robert A. Israel, Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Acting Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Division of Analysis Jennifer Madans, Ph.D., Acting Director Jennifer Madans, Ph.D., Acting Chief, Demographic Analysis Staff Diane M. Makuc, Dr.P.H., Chief, Analytical Coordination Branch Sandra T. Rothwell, Chief, Longitudinal Analysis Branch Division of Epidemiology and Health Promotion Ronald W. Wilson, Director Patricia M. Golden, Special Assistant Diane K. Wagener, Ph.D., Chief, Environmental Studies Branch Ronald W. Wilson, Acting Chief, Health Status Measurement Branch Acknowledgments The authors gratefully acknowledge the assistance of the following people: Patricia Knapp and Brenda Barber pro- vided computer programming assistance. Donald Beu created the maps of the service areas. Daryl Rosenberg, Michael McMullen, and Rose Connerton provided the Medicare data. Donald Malec reviewed the report. a 1 " r , . x - 1 2 - ® wa ER » = x - . - ae - 7 - A gly Pa - . Fw CE ) I SRL 4 Nv : . . B ; “ 4 ‘ vo - So - > . . oa A - 5 . . . - A ER , - : - * . . . . . v r - F - . . : 3 . . # ros ow og = . - oo oo id =F, i . - PO . = SI - cn a 8 . z ~ : % s . oo . wo i a & : CY Cn = & ‘ , 3 B w . Pe S ; oF . - . I ) - . = . " CI - Be 3 oe " : =e N oo RE % . J a y & on z E . | 4 L » - . . I iL = SI » T = oe C3 . “ . E - % . 0» i et, Ea i . a . . - - - . . = & ny ! . ow . 3 - = ¥ - ’ La . i za . oe Fi , &mT- = Se LE . : 3 & . x + LE = - a A - a 3 " - # 2 3 - x - «4 . - ” a “= ® x - . E : . . Contents BaCKGIOUNM ic crore sto mio 5050508 os 50 60003 100 E0010 080 060 i HL 6 ow 50 200 6.0 0 0 0 6 0 0 Bk 1 Health care commiting areas (TICCASY io cui win isings osm soos in 400 0015 ¢ Bl toki won wii 0090 458 312 2 Sidhe cw win 0 010 0 1 LADO IEEE GEEAR oo v vv norm ums 15 toe santos sas wiuiiel 070 BIR. 50 400 ere! Se IE 213 16 Ruan vie £1 RTO RT E08 ew i en 2 MEINOAS. ooo ann masses nen iy SRE 4 5g fe SESE SEE wy SP ERE EEE wwe RE EO Be sp GR 0 3 DOA SOUTCE 27 5:4 5 00 vs 0 4 H4 TDR TSAR 58 0 30 30 HE FR UE 00 03000 0 SE (3 15900 0 ie wg WR RE 00 ls ims © ST IE 0 3 CHOSEET UIVAIYRES.. . isi 5 300m 0 ere oo 90 4 55 8038 4.60 58 ct rs fr 6 05 0B 5 oe 4 0 6 35 BL 0 6 3 0 0 90 RL at #0 HE 8 > DD StINGE HCHEUTES sv viv em smumm mast sm ss sive 1 0 FES 60 0 ete ev 1.150 3 0 0 no we 0 mi 20 3 0 4 Precluster analysis tHRIOME cu. mw nem tm mies smn smn 20 37 nie #00 £10 2 810 BUS 610 40 8000 108 £100 55 30 020 030 914 0 BU HEE 60H A 5 Metropolitan statistical areas (VISAS) ..uvusimrcninrinsprassorensss as ss®mess ones sony sas suwnsvson sys 5 Counties WHI 10 HOSPIAIE 0 visa 5 om seems ims miners ie fh nme stew ime yi RS ors won mop 4 R80 Bo on won i 5 Cottics WHI SIPONG LOE + .cumsss simmons sma Lams «ems ms ess Ses nis ss « oaee myn se 5s ve ws wanes peda 5 Geographic TNS Of ZORIVEIE . .c.oivun sn emus maui an osm sr 80 BR 23 wk 21% 0 i 050 50 0 BEF 0 ot 6% 8 0 I 8 0 i i 5 INTIS OF CIB EES 5 05 5 view or vv 5525 915.5 158 0 10 5 8 0 0 Fh io 9 op 90 E08 00100 Ft pe 0 8 5036 08 vm oc 0 25 6 Posteluster analysis METNOES «nur vims nus mre ments mius mass nr yumm Fac ms wna Bale aa ness ms veme®sas 6 Comparison of AUSTAIIVE SOIIEIONE 1x5 wih + 4 ssw sine 150758 £4 sos wee wii 0 wide ie saw wo wim sms 8 how wip we gm sR 8 6 8 Area size and demographic CharacteriSHIES , vy russnran nesses £ arms £E 480 24838 CAFES ERARH 1 0.0% A3 LBC HDT CERES 2 8 MS As in linked and UnHOKEO SOMILIONS ous 50 5.55 0 fos 5508 500 88 E05 00.00 woh a8 5145 500 6 Boi £10 I Hl i hh RAS 9 THEO COTE TEMOUTORE ; worn 100 012 1 508 $18 5000 1 + io memos oF 1h 68 stains ss gts 18 G2 61 8 bigot» esse 510 6 & hr sign ve od 10 Travel Tor HEAL COIR. cou vr msmmansmnr cs mms mnie meme meivme mtn seme esos us @minsnassme: spss sss 11 SUTRIIATES «0 vrei 5 50 20500 530s me 90 ip 2 11 402 0 £ Sd Ta Sir 4 es 03 50 UE 4 8 0 sry 7 0 FSR 4 i 0 3 ws 0000 Fe Fak emi oe 1 0 iE TB A 15 BR CTONTRE, ov v0 10 200 AT ri ti a BA er go 0 BAR AB tots ir A TEL HE ny aww FEI ad once rr Se fn 16 Appendixes 1 Alternative. diStANCE: YIYSHSUIES ov sm 504530 50 mo mms 04 ahs 018 8 Fe bret + 3 8 8 FRE $0 210.390 4 mom sm 05 8 00 #0 3 we rs vg mk 4 18 II. Maps of health service areas for the 800-area unlinked solution. ..................oouun. Him sma AE 19 HII. Listings of health SErviOe AU@AS . «wien sm sms em ans sme so om as sa Waves fas s sons SHEE Ran 4048 0H $n 48 200 00008 28 IV. ObSielric SErVIOE BUSA: cvs vo nrrrt 401 S0 00500 At Siro 8 8 FASEB Rh 0 5 ED RES BRR 08 50408 RD EIREE 8 Wh Soh 3 75 Text tables A. Selected characteristics of service areas by type of area: United States ..................cooiiiiiiin... 8 B. Number of counties and population of service areas by type and metropolitan status of area: United States.. 9 C. Availability of health care resources by type of area: United States ...........coiiiiiiiiiiiinnneenenn. 10 D. Availability of health care resources by type and metropolitan status of area: United States................ 11 E. Percent distribution of areas and population according to travel for routine Medicare hospital stays by type of area: United Sates, JI88 «ocr irs mani: #7055 nein wana s pin ies oie ms #9030 4 5008 50m se oes 12 F. Percent distribution of areas and population according to travel for routine Medicare hospital stays by type and metropolitan status of area: United States, 1988................ coo, 12 G. Percent of short-stay hospital stays outside area of residence by age, type of county, and type Of 2180. LInUed 'SIHEES, TB. . ou wv vim v hm bmi i oki si eid 309 $0518 895 Wh mio wes 505508. 0 5 1 ws 9m 5 50s BER HR mein 2 i3 H. Percent of routine Medicare hospital stays outside area of residence by type of county and type of aren Uned Sates, 1088. «uvnc own vwmis wie 805 ow wo ome srain B0 oem aos ws B50 S00 5 wwe w wpe 5oeh 13 State abbreviations AL AK AZ AR CA CO CT DE FL GA HI ID IL IN IA KS KY LA Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Hawaii Idaho Illinois Indiana Towa Kansas Kentucky Louisiana ME Maine MD Maryland MA Massachusetts MI Michigan MN Minnesota MS Mississippi MO Missouri MT Montana NE Nebraska NV Nevada NH New Hampshire NJ New Jersey NM New Mexico NY New York NC North Carolina ND North Dakota OH Ohio OK Oklahoma OR PA RI SC SD TN TX UT VT VA WA wv WI wY DC Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming District of Columbia vi Symbols ——— Data not available 0.0 Category not applicable Quantity zero Quantity more than zero but less than 0.05 Quantity more than zero but less than 500 where numbers are rounded to thousands Figure does not meet standard of reliability or precision (estimate has relative standard error of more than 30 percent) Figure suppressed to comply with confidentiality requirements Health Service Areas for the United States by Diane M. Makuc, Dr.P.H., Office of Analysis and Epidemiology, National Center for Health Statistics; Bengt Haglund, Dr.Med.Sc., Uppsala University; Deborah D. Ingram, Ph.D., Joel C. Kleinman, Ph.D., and Jacob J. Feldman, Ph.D., Office of Analysis and Epidemiology, National Center for Health Statistics Background The objectives of this report are to document methods used to identify health service areas for the United States, and to describe and evaluate these areas. We define a health service area as one or more counties that are relatively self-contained with respect to the provision of routine hospital care. Service areas that include more than one county are characterized by travel between the coun- ties for routine hospital care. We have assigned every county in the coterminous United States to a single service area that must have at least one hospital. The health service areas reflect current travel patterns between coun- ties for routine hospital care. This project was motivated in part by a decision at the National Center for Health Statistics (NCHS) to consider health service areas as possible primary sampling units (PSUs) for the National Health Interview Survey (HIS). Another motivation was the need to use health service areas as units of analysis to measure the availability of health care resources (for example, per capita physicians and hospital beds), to study geographic variation in health care use, and to study the relationship between health care resources, health care utilization, and health status (1). Before this project the most recently defined national health service areas were constructed using data that are now more than 20 years old (2). Thus, the new service areas will be a useful tool in conducting health services research. The concept of a service area has been used in analyses of NCHS data to study the relationship between health care resources and use of health care (3,4). For example, data from the HIS have been used to study the relationship between the supply of surgeons and surgery rates within the HIS PSUs under the assumption that the PSUs were approximate health service areas (3). Another study analyzed the relationship between physician supply and use of physician services based on utilization data from the HIS and birth records (4). Individuals were categorized according to the physician supply in their service area of residence and a categorical data analysis was carried out. Health service areas have been defined in a number of ways (5-8). Some of the methods that have been used to define service areas include: (a) geopolitical boundaries, for example, counties and metropolitan statistical areas, (b) physical distance between hospitals, for example, 15- mile radius, (c) patient-origin analysis, and (d) price analysis. Geopolitical boundaries have been used as ser- vice areas primarily because they are expedient and com- patible with many data systems, thus increasing the possibilities for data analysis. Physical distance has been considered on the basis that most hospital patients are admitted by office-based physicians who travel a limited distance to hospitals. Service areas based on patient-origin data use empirical data on travel patterns among commu- nities for health care. The community can be a ZIP Code, census tract, town, county, metropolitan statistical area (MSA), or other geographic areas. Service areas may be formed using different algorithms with patient-origin data. Further, different sources of patient-origin data may be used, for example, hospital discharge records, birth records, or death records. Health care commuting areas (HCCAS) Although substantial literature exists concerning the definition of health service areas for States or local areas, there have been few attempts to identify service areas for the entire United States. One such attempt was health care commuting areas (HCCAs), defined by grouping counties based on 1968 national natality data, 1969 na- tional mortality data, and 1970 census data on the journey to work (2). The natality data file contains data on the county of mother’s residence and the county of delivery, providing information on travel patterns for obstetrical care. Similarly, the mortality data file includes the county of residence of the decedent as well as the county of death, providing information on travel patterns for hospi- tal stays of decedents who die in hospitals. However, this analysis was not restricted to deaths that occurred in hospitals. An algorithm was developed to group counties into areas so that travel outside areas was minimized. The algorithm grouped counties to minimize a commuting ratio (CR). CR is defined as the ratio of the total demand for health care services occurring in the group to the demand occurring in the group by group residents only. 1 The algorithm began by identifying a centroid, the county with the largest number of occurrences (for example, births). If the CR for the centroid and the county that had the highest percent of residents commuting into the cen- troid was less than for the centroid county alone, then the two counties were grouped together. The CR was recalcu- lated adding the county with the highest percent commut- ing into the present group. If the CR decreased the county was added. The process was repeated until there were no counties that reduced the CR of the existing group. The county with the largest number of occurrences that was not included in the first group became the next centroid and the group was formed for that centroid in a similar manner, with the added constraint that the original cen- troid county could not be added to the new group. The process that was used to form the first two groups contin- ued until there were no counties left ungrouped. Noncen- troid counties could be linked with more than one group. Thus, the final step was to determine the best group for each county, that is, the group for which the county reduced the CR the most. Eight alternative definitions of service areas were developed based on assigning different relative weights to the natality, mortality, and employment data and different threshold or minimal commuting levels. The total number of service areas for the United States under the eight definitions ranged from 385 to 950. The preferred algo- rithm assigned relatively more weight to the natality data and less to mortality and employment, resulting in the delineation of 780 areas for the nation. A disadvantage of this approach is that it requires complex computer pro- gramming. In addition, the algorithm was not documented in sufficient detail to replicate it precisely. An analysis of data from the 1978 HIS compared how well the HCCAs and two other types of county aggrega- tions performed as primary care physician service areas (9). The 780 HCCAs were compared with 183 areas developed for economic analyses by the Bureau of Eco- nomic Analysis and 492 basic trading areas developed by Rand McNally. The results of this analysis suggested that the HCCAs were the most appropriate primary care physician service areas. The HCCAs were the smallest in size and population and exhibited an amount of outside area travel for care similar to that of the two larger types of areas. They also exhibited the greatest variability in physician supply. A similar analysis using 1983 NHIS data showed that the percent of visits that occurred outside the HCCA of residence remained about the same between 1978 and 1983, indicating that overall, service areas did not change substantially over this period. Labor market areas A more recent attempt to define service areas for the entire United States occurred in 1987 when the U.S. Department of Agriculture defined 382 labor market areas and 875 subareas to be used for statistical and planning purposes in research on rural America (10). These areas were based on 1980 Census journey to work data for counties. The relative strength of the commuting ties between each pair of counties was measured by the following ratio: the total number of commuters between two counties divided by the resident labor force of the smaller of the two counties. A matrix of these “distance measures” was used in hierarchical cluster analysis to form groups of counties with strong shared commuting ties. In addition to the cluster analysis, other criteria used to identify the labor market areas were (a) a minimum population of 100,000 due to Census confidentiality stan- dards and (b) maximum geographic detail. The first crite- rion was achieved by combining some of the areas identified in the cluster analysis, while the second was achieved by identifying market subareas within larger labor market areas. Methods Data source County was selected as the basic geographic unit of analysis because the county was the smallest geographic unit for which most national data sources are currently available. However, counties differ substantially in size and meaning across the country. In New England, towns and townships are smaller and more meaningful geo- graphic units than counties, but data were not available for these geographic units. At the State and local level much work on service areas has used ZIP Code or census tract as the basic unit of analysis (11,12). Use of a smaller geographic unit than the county is desirable because travel patterns may differ for different sections of the county. In metropolitan areas the use of counties may mask impor- tant travel patterns for health care. Nevertheless, at the national level it was not practical to use geographic units smaller than the county at the present time. Three data sources with county-level data for the entire United States were considered — natality, mortality, and Medicare. We selected 1988 Medicare data on short-stay hospital stays as the most appropriate data source because it provides information on all types of hospital services and includes information about type of stay so that stays for specialized care could be eliminated. The drawback of Medicare data is that this information is only available for persons 65 years of age and over (17 percent of hospital discharges in 1988 based on the National Hospital Discharge Survey). However, data from the NHIS indicate that travel for ambulatory medical care does not differ substantially by age (13). Further, 1989 HIS data show that travel patterns for hospital stays also do not differ substantially by age (table G). From the initial Medicare file containing data for 10 million hospital stays, 2 million records were eliminated for deaths, disabled and end-stage renal disease Medicare beneficiaries, hospital stays outside short-stay general hos- pitals, county of beneficiary missing or invalid, and resi- dents outside the coterminous United States. Overall, 28 percent of hospital stays occurred outside the county of residence. Records were eliminated for the 1 million stays with Diagnosis Related Group (DRG) codes for which more than 35 percent of hospital stays occurred outside the county of residence. The rationale for this was to eliminate stays that were most likely to be for specialized care. Thus, the final Medicare data base included 7 million hospital stays with a median of 853 stays per county of residence. The number of Medicare hospital stays was greater than 100 for 97 percent of the counties. Only 48 counties had fewer than 50 stays. Natality data were also considered as a possible data source. Birth certificates provide data on travel for obstet- ric care, and these data were used in the development of HCCAs. However, only about 70 percent of general med- ical and surgical hospitals have obstetric units (14) and 24 percent of counties do not have obstetric facilities, whereas only 16 percent of counties do not have short-stay general hospitals (15). Further, travel patterns for obstet- ric stays may differ from other types of stays because mothers may seek hospitals with birthing rooms or other types of facilities not routinely available. Although we decided not to use natality data to define the health service areas, national obstetric service areas were defined in a separate analysis based on natality data for 1984-86 using the same methods described for the Medicare data (see Appendix IV). Mortality data were eliminated as a possible data source because only about 60 percent of deaths occur in hospitals, and 70 percent of these are for persons 65 years of age and over, the same population covered by the Medicare data (16). Cluster analysis We used agglomerative hierarchical cluster analysis to group counties into service areas (17,18). Hierarchical cluster analysis generates a hierarchical classification for a set of observations (for example, counties) based on a measure of distance between observations. The distance measure can be any function that defines the relationship between observations. The hierarchical classification is defined by an ordered sequence of partitions of the observations. In the initial partition each cluster contains a single observation, that is, each observation starts out as its own cluster. The final partition consists of a single cluster that includes all observations. The partition at each level of clustering is formed by joining a single pair of clusters with the smallest distance measure. The sequen- tial process of forming clusters can be carried out using different methods of redefining the distance measure after the first two observations have been joined to form a cluster. We selected the average linkage method as appro- priate for construction of health service areas. In average 3 linkage the distance between two clusters is the average distance between all pairs of observations from the two clusters. The distance between two clusters C(K) and C(L) is defined by Distance (K,L) = £2 D;/(N(K)*N(L)) where Dj; is the distance measure between observations i andj, N(K) is the number of observations in C(K), and N(L) is the number of observations in C(L). When N(K)=N(L)=1, then Distance (K,L) = D;. The ratio- nale for selecting the average linkage method is that it considers the distance between all pairs of observations in two clusters and is conceptually simple to under- stand. The average linkage method was also used in the development of labor market areas. Examples of some other conceptually simple methods that were rejected are single linkage and complete linkage. In single linkage the distance between two clusters is the mini- mum distance between an observation in one cluster and an observation in the other cluster. This approach tends to produce elongated clusters which are not appropriate for service areas. On the other hand, with complete linkage, the distance between two clusters is the maximum distance between an observation in one cluster and an observation in the other cluster. This approach tends to produce clusters with equal diame- ters and can be distorted by moderate outliers (19). The CLUSTER procedure in SAS was used to carry out the agglomerative hierarchical cluster analy- sis using the average linkage algorithm (19). The TREE procedure in SAS was used to identify the desired level of clustering. Distance measures The input data for the cluster analysis was the follow- ing lower triangular matrix of distance measures (D;;). D,; Ds, Ds, D,, D, 5® 4 Dini The distances measure the strength of the flow be- tween pairs of counties (i and j) for routine hospital care. It should be noted that distance in this context has nothing to do with physical distance. Several alter- native distance measures were considered, using the following notation: F; = Flow of hospital stays to county i from county j, i.e., the number of hospital stays occurring in county i by residents of county j. For each pair of counties this flow can occur in both directions. P; = Total production of hospital stays (number of occur- rences) in county i. C; = Total consumption of hospital stays by residents of county i (number of hospitalizations for the residents of county i). H; = The consumption of hospital stays in county i by the residents of that county (home county consumption). Note that H; = Fj; The flow between pairs of counties can be related in different ways to P, C, and H to construct many different distance measures (see Appendix I). After considering a number of alternatives, we selected the same measure that was used by the Department of Agriculture to construct labor market areas (10). The measure is defined as follows: D; =1-(F; + Fy/lG; i = 1- (Fy + FplC; Note that D; = Dj. If D; < 0 then set D;; = 0.001. The distance between two counties ranges from 0.001 to 1 and is a function of the total flow of hospital stays between two counties (regardless of the direction) and the consumption of hospital stays in the smaller county, that is, the county whose residents have fewer hospital stays. The ratio of flow to consumption is subtracted from 1, so that the distance is larger for pairs of counties with less interac- tion. The flow between counties is divided by consumption in the smaller county so that large counties do not domi- nate the analysis. We want the distance between two counties to be small in the case when residents of a small county are highly dependent on the hospitals in a large county for care. If the consumption by residents of the large county were included in the denominator then the strength of the relationship between a small and large county would be reduced. A threshold value for the flow between two counties was defined such that if F;;/C; < threshold value then Fj; was set to zero. The rationale for setting a threshold level was to eliminate highly unusual flow patterns from the analysis. For example, such unusual patterns could arise through hospital stays during out-of-town travel. We arbi- trarily set the threshold value to 2 percent. This level was also used to define HCCAs (2). To calculate the distance matrix, the Medicare hospi- tal discharge data were arranged in a matrix with the rows representing the county of occurrence and the columns representing the county of residence as follows: if P, > P; otherwise. Fy, Fy, “ee F,, F,, Fy, vue F,, F,,; F,, che F,. The off-diagonal elements of the matrix were the flows from county j to county i (F;) and from county i to county j (F};), and the diagonal elements were the home county consump- tion (F; = H;). The row sums were the total number of Medicare hospital stays occurring in each county (F,. = P,) and the column sums were the total number of Medicare hospital stays by residents of the county (F.; = C;). Precluster analysis methods The distance matrix for all 3,073 counties in the coterminous United States could not be analyzed simulta- neously due to computer space limitations. Thus, the United States was divided into six overlapping regions, such that the borders of most States were internal to one region. The six regions were defined as follows: I. Connecticut, Delaware, Kentucky, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, North Carolina, Ohio, Pennsylvania, Rhode Island, Tennessee, Vermont, Virginia, Washington, DC, and West Virginia II: Alabama, Florida, Georgia, Kentucky, Mississippi, Missouri, North Carolina, South Carolina, Tennessee, and Virginia III: Illinois, Indiana, Iowa, Kentucky, Michigan, Minne- sota, Missouri, Ohio, Pennsylvania, West Virginia, and Wisconsin IV: Alabama, Arizona, Arkansas, Colorado, Kansas, Lou- isiana, Mississippi, New Mexico, Oklahoma, Tennes- see, and Texas V: Arkansas, Colorado, Iowa, Kansas, Minnesota, Mis- souri, Montana, Nebraska, North Dakota, Oklahoma, South Dakota, and Wyoming VI: Arizona, California, Colorado, Idaho, Kansas, Mon- tana, Nebraska, Nevada, New Mexico, North Dakota, Oklahoma, Oregon, South Dakota, Utah, Washing- ton, and Wyoming The use of six overlapping regions complicates the analysis considerably because the number of clusters must be specified for each of the six regions and decisions about the best solution for counties that appear in more than one region must be made. Metropolitan statistical areas (MSASs) Alternative solutions treated MSAs in two ways: (a) counties within an MSA were constrained to be within the same service area and (b) no constraints were placed on MSA counties. In both approaches nonmetropolitan counties could join the MSA counties to form service areas and counties in different MSAs were allowed to join the same service area. The rationale for constraining counties within an MSA to be in the same service area is that MSAs are formed based on strong economic and other linkages, and it was expected that these linkages might also apply to health care. On the other hand, we developed solutions with no constraints because certain MSAs encompass large geographic areas that may include multiple service areas. Further, linkages between counties for health care may differ from those for other types of services. There are 317 MSAs in the coterminous United States that include 729 counties. Of the 317 MSAs, 51 percent have 1 county, 21 percent have 2, 11 percent have 3, and 7 percent have 4 counties. The remaining 33 MSAs have 5 to 12 counties, with the exception of 1 MSA (Atlanta, Georgia) that has 18 counties. Counties with no hospitals After examining initial results from the cluster analy- sis, we found that certain small counties with no hospitals had an undue influence on the results. This could arise because counties are clustered based on the rank order of the distance measures without regard to the stability of the measure. A county whose residents have few hospital stays would have unstable and possibly very small distance measures, thereby clustering early and affecting the aver- age distance measures for all subsequent linkages with its cluster. To help alleviate this problem, counties with no hospital stays occurring within the county (zero- production) were excluded from the cluster analysis. After the cluster analysis was completed, each zero-production county was assigned to the cluster where the greatest number of its residents’ hospital stays occurred. The only exception to this procedure was for solutions that con- strained all counties within an MSA to be within the same service area. In those solutions, counties within an MSA were kept in the cluster analysis after being linked to the other counties within their MSA, regardless of whether or not they had any hospitals. No hospital stays occurred within 503 counties, 54 of which were in MSAs. Counties with strong ties After linking counties within an MSA and eliminat- ing counties with no production of hospital stays, counties with very strong ties were linked prior to the cluster analysis. Two counties were linked if the flow from county i to county j was more than 2 times greater than the flow from county i to any other county and the total consumption for county i was at least 50 hospital stays. However, if more than one county met this criterion with county j, then no linkage was formed. Part of the rationale for prelinking counties with strong ties was to reduce the number of observations in the cluster analysis, due to computer space limitations. However, few pairs of counties met these criteria and so these linkages had little effect on the number of observations or the clusters that were formed. Geographic units of analysis In the analysis that linked counties within MSAs, the initial 3,073 counties in the coterminous United States were reduced to 2,212 geographic units by link- ing the 568 counties within multiple-county MSAs and eliminating 449 nonmetro counties with no hospital stays. As previously discussed, a separate distance matrix was created for each of the six regions. The initial number of counties in each region, how that number was reduced by linking counties within MSAs, 5 linking counties with strong ties, and eliminating coun- ties with zero-production are shown below. Region I II III Iv Vv VI Counties 804 953 980 929 913 807 Counties/MSAS 615 802 842 839 859 769 Zero- production 121 181 150 112 154 131 Linked pairs 7 10 15 19 18 15 Analytic units 487 611 677 708 687 623 In the analysis that does not link counties within MSAs the initial 3,073 counties in the coterminous United States were reduced to 2,570 after eliminating the 503 counties with no hospital stays. The number of geographic units in the analysis for each region are shown below. Region I II III Vv Vv VI Counties 804 953 980 929 913 807 Zero- production 147 213 168 121 162 132 Linked pairs 10 12 16 19 20 17 Analytic units 647 728 796 789 731 658 Numbers of clusters In carrying out a cluster analysis a decision must be made concerning the appropriate number of clusters. In making this decision we considered previous research on national health service areas. An evaluation of the previ- ously defined HCCAs indicated that these areas per- formed well as health service areas (9). We decided to develop a solution with 800 clusters, approximately the same number as HCCAs, under the assumption that the total number of service areas has not changed substan- tially between 1970 and 1988. This seems to be a reason- able assumption because there have not been major changes in the availability of hospital beds during this period (the number of community hospital beds per 1,000 population in the United States declined slightly from 4.3 to 3.9). The number of clusters in each of the 6 regions for the 800-cluster solution was based on the division of the HCCAs among the 6 regions: I 220, II 220, III 260, IV 240, V 240, and VI 240. A 1400-cluster solution was also developed to provide a number of clusters closer to the current number of PSUs in the National Health Interview Survey. The 1400 clusters were divided among the 6 regions in the same proportion as the 800 clusters as follows: I 385, II 385, III 455, IV 420, V 420, and VI 420. Postcluster analysis methods The cluster results for the six overlapping regions had to be combined to form a national solution. Be- cause the regions were defined using State borders, it was necessary to consider States in combining the six solutions. For each State, clusters were selected from the region with the most complete coverage of the State’s borders. Forty-one States had borders that were completely internal to only one region. The region selected for each of these States was: I: Connecticut, Delaware, Maine, Maryland, Massachu- setts, New Hampshire, New Jersey, New York, Penn- sylvania, Rhode Island, Vermont, Virginia, Washington, DC, and West Virginia II: Alabama, Florida, Georgia, North Carolina, and South Carolina III: Illinois, Indiana, Ohio, Michigan, Wisconsin IV: Mississippi, Louisiana, and Texas V: Kansas, Nebraska, North Dakota, and South Dakota VI: Arizona, California, Colorado, Idaho, Montana, Ne- vada, Oregon, Utah, Washington, and Wyoming For the remaining eight States, the two regions that internalized the greatest physical distance of each State’s borders were identified. Clusters from the region with the most complete border coverage were selected unless a comparable cluster from the second region included coun- ties from a greater number of States. The rationale for this was that in those instances the second region provided better coverage of selected parts of the State’s borders. The first and second choice regions for each of the eight States are: Kentucky, I and III; Tennessee, II and IV; Iowa, V and III ; Minnesota, V and III; Missouri, VI and III; Arkansas, IV and V; New Mexico, VI and IV; and Oklahoma, IV and V. This algorithm can be illustrated by considering New Mexico. All of New Mexico’s borders are internal to region VI except for the border with Texas, a State that is not part of region VI. Thus, the region VI solution provides no information about the flow for hospital care between New Mexico and Texas. To obtain that informa- tion the region IV solution must be used. In those clusters where New Mexico counties are linked with Texas coun- ties in the region IV solution, the solution would be selected according to the algorithm described above be- cause the region IV clusters would include two States, Texas and New Mexico, whereas the region VI clusters would only include one State, New Mexico. The next step in completing the national solution was to ensure that each county (except zero-production coun- ties) was assigned to one and only one cluster. Counties with multiple solutions and counties without a solution were identified and a unique solution was assigned to each of these counties. The solution for which the county had the smallest percent of outside area travel for care was selected when a county had more than one solution. The rules defined above to select clusters from the two re- gional solutions for these eight States resulted in a small number of counties not being assigned to any cluster. In these instances either the alternative regional solution was used or the county without a solution was added to an existing cluster. Finally, each zero-production county was assigned to the cluster where the greatest number of its hospital stays occurred. Six of the zero- production counties were not contiguous to the cluster where the greatest number of hospital stays occurred in at least one of the four solutions (Meade, Kentucky; Shelby, Missouri; Harding, New Mexico; Dallam, Texas; Rains, Texas; and Wayne, Utah). These coun- ties were reassigned to the cluster where the second largest number of stays occurred, so that noncontigu- ous clusters were not formed. A check was made for other clusters containing non- contiguous counties. Some of the noncontiguous clusters that were found were due to miscoding of county of hospital stay on the Medicare file. After editing the Medicare file and reassigning counties to clusters based on the corrected data, 16 counties or groups of counties remained noncontiguous to their clusters in at least 1 of the 4 solutions. Solutions for four of these counties or groups were not changed as they were separated from their clusters only by water, and there was substantial commuting across the water for hospital care (Mackinac, Michigan; Chowan, and Dare, North Carolina; and Mid- dlesex, Virginia). In two instances water separates MSA counties that were constrained to stay together in the linked solutions (Marin, California and St. Tammany, Louisiana). Nine noncontiguous counties or groups were reassigned to a contiguous cluster based on examination of the travel patterns for the noncontiguous counties and/or their assignment in contiguous solutions. For example, Cheyenne, Colorado was not contiguous to its cluster in the 800-unlinked solution only. Thus, results from the 800-linked solution were examined to reassign Cheyenne, Colorado to a contiguous cluster in the 800-unlinked solution. Other counties that were reassigned to new contiguous clusters in one or more solutions were Inyo, Mono, Los Angeles, and Orange, California; Rio Blanco and Kit Carson, Colorado; East Carroll and Morehouse, Louisiana; Barnstable, Massachusetts; Clatsop, Oregon; Montour and Schuylkie, Pennsylvania; Reeves, Texas; and Pacific, Washington. The health service areas for the 800-unlinked solution were sent to each State for review. This provided a check on the validity of our methods across the United States. Overall, the health service areas produced by the cluster analysis were consistent with the knowledge of State experts regarding travel patterns in their States. Suggested changes were made to the health service areas for 27 counties in 9 States. Some of the possible reasons why the cluster analysis might not have worked well in certain areas are that the ranking of distance measures may not have been optimal for counties with few hospital stays or perhaps a different level of clustering may have been more appropriate for some areas. Changes suggested by State experts were evaluated with respect to their effect on the percent of stays by county residents outside the service area. Only changes that reduced the percent of stays by county residents commuting outside the service areas were incorporated in order to maintain consistent methodology across the nation. Comparison of alternative solutions Four alternative solutions were generated: 800 clus- ters and MSAs unconstrained, 800 clusters with MSA counties linked within the same service area, 1400 clusters and MSAs unconstrained, and 1400 clusters and MSA counties linked. Appendix II presents maps for the 800- unlinked solution. Appendix III lists the service areas for all four solutions and presents some summary statistics at the county and service area levels for the 800-unlinked and 1400-unlinked solutions. Data presented for the four alternative solutions in- clude service area size and demographic characteristics, health care resources, and travel across service area bor- ders for care. For comparison, similar data are shown for health care commuting areas (HCCAs), counties, and National Health Interview Survey primary sampling units (HIS PSUs). (In New England there are multiple HIS PSUs in counties because the PSUs are assigned to townships. The entire county was assigned the PSU num- ber that occurred most frequently in the county.) The numbers of service areas are 802 and 1,415 in the 2 unlinked solutions and 790 and 1,400 in the 2 linked solutions. In both instances the numbers of areas differ slightly from 800 and 1,400 because the number of clusters had to be specified at the regional rather than national level. In addition, some adjustments to the solutions were made to eliminate noncontiguous clusters and to take into account the review by State experts. Area size and demographic characteristics The distribution of the number of counties per area differs substantially among the four solutions (table A). The percentage of single county areas ranges from 11 percent for the 800-unlinked solution to 50 percent for the 1400-linked solution. The linked solutions yielded more single county areas than the unlinked solutions, and the 1400-area solutions yielded a much higher percentage of single county areas than the 800-area solutions. The per- centage of areas with 10 or more counties is also greater among the linked than unlinked solutions (6 percent compared with 3 percent for the 800-area solutions). Among the 1,855 HIS PSUs, 59 percent include only 1 Table A. Selected characteristics of service areas by type of area: United States [Data are for the coterminous United States] Type of area 800 800 1400 1400 Characteristic of area unlinked linked unlinked linked HCCA HIS PSU County NUMBBrofareas. . . «i: vi cvswemusios as 802 790 1,415 1,400 775 1,855 3,073 Number of counties per area Percent of areas Be 5 4 5 wis mre wr are et Pani le ge 8 WEE 8 114 16.6 41.2 49.9 22.7 58.7 100.0 ZB 4 oat ER RE ne 41.6 41.5 44.6 35.4 38.1 37.1 sore BB, ena a EAE REREE RY EE 28.4 22.2 10.7 9.0 18.5 25 BB. vncinusvnmrmnwndns hn nes 16.3 13.7 3.2 4.1 13.8 12 TON MOIS... «ws so sm ames ye screws 25 6.1 0.4 1.6 7.0 0.5 Boundary crossing Stateboundary .................. 15.7 15.9 6.8 7.4 179 17 - Regional boundary . . c «cv vv vu eve on 27 3.0 1.3 1.6 4.1 0.2 - MSAs in areas NOMBA. ovis mantmsnmemramems em 54.4 64.3 68.3 77.9 64.9 84.6 76.3 MSA: 501515 3 3 om soy wm mh bl Eh eB 39.4 31.8 30.3 21.6 28.1 14.2 23.7 SMSASOIrMOMB. . « tas cusvssmsvais 6.2 3.9 1.3 0.5 7.0 1.2 - 1987 population Population in thousands By DBICARIIBL vv vont: 4 G80 ls 0 0 14 13 7 7 10 3 3 250 PRICENtilB, csc r cvs n mrs ny 50 41 21 18 33 14 11 BOth percentile. . .— .» vs snvmw enone 126 93 52 41 77 33 23 TBI BIBS, +» vss ww ven win ® + aig 279 238 143 113 225 73 54 OB percentile, . .... ives svnenns wa 1,161 1,276 726 679 1,285 388 316 Table B. Number of counties and population of service areas by type and metropolitan status of area: United States [Data are for the coterminous United States] Type and metropolitan status of area 800-unlinked 800-linked 1400-unlinked 1400-linked Characteristic of area Metro Nonmetro Metro Nonmetro Metro Nonmetro Metro Nonmetro Numberofareas.. . wi. weims soos ss 366 436 282 508 448 967 310 1090 Number of counties per area Percent distribution Yori 5. 4 5: 5 ham di i os 6 8 6 55 15.8 7.1 21.9 20.3 50.9 18.1 59.0 DB wo pwn 0 By ew 33.1 48.6 25.5 50.4 525 41.0 35.8 35.3 AB ii imi apie iamews we oF ees 32.5 252 24.5 20.9 19.6 6.5 23.6 4.9 BRD crim sven RoR BE Arye mm 24.3 9.9 26.6 6.5 6.5 1.7 15.5 0.8 VO.OT TIONS 55.50. 5nd i 3% Wists avaey 4.6 0.5 16.3 0.4 13 - 71 —- 1987 population Population in thousands Sthpercentile. . . ............. 110 9 134 9 59 6 115 6 25th percentile... oc « wv vovins wuss 192 28 208 26 135 16 176 15 50th percentile. . . ............. 296 56 362 51 233 31 286 30 78thpercentile. .'.. civ vnsws 2a 570 95 705 84 466 59 646 55 O5th percentile. . « » cave suis ms fa 2,116 186 2,427 165 1,634 127 2,203 113 Note: Metropolitan service areas include at least one metropolitan county. county and 0.5 percent include 10 or more counties. Among the 775 HCCAs the distribution of numbers of counties per area is similar to that for the 800-area linked solution, although there are somewhat more single county HCCAs (23 percent compared with 17 percent). The percent of areas that cross State boundaries is 16 percent for both 800-area solutions, 7 percent for both 1400-area solutions, and 18 percent for the HCCAs. Only 2 percent of HIS PSUs cross State borders. Crossing of the 4 census region borders occurs for 1-3 percent of the service areas and 4 percent of HCCAs. The percentage of areas that include only nonmetro- politan counties is greater for the 1400-area than 800-area solutions and greater for the linked than unlinked solu- tions, ranging from 54 percent for the 800-unlinked solu- tion to 78 percent for the 1400-linked solution. Thus, more nonmetropolitan counties were joined with metropolitan counties in the unlinked solutions than the linked solu- tions, resulting in a higher median population for the unlinked solutions. In all 4 solutions a small number of service areas include at least 1 county from each of 2 or 3 MSA:s, ranging from 0.5 percent for the 1400-linked solu- tion to 6 percent for the 800-unlinked solution, and 7 percent for the HCCAs. Two service areas in the 800- unlinked solution and 4 service areas in the 800-linked solution include counties from 3 MSAs. Further, in the two unlinked solutions, some multiple-county MSAs split into more than one service area. For example, in the 800-unlinked solution counties from 42 MSAs split into 2 service areas and counties from an additional 22 MSAs split into 3 or more service areas. The 1987 median population per area ranged from 41,000 for the 1400- linked solution to 126,000 for the 800-unlinked solution. The 1987 median population was 33,000 for HIS PSUs and 77,000 for HCCAs. Table B compares characteristics of service areas that include at least one metropolitan county with those that do not. Areas that include both metro and nonmetro counties have been grouped with areas that include only metro counties because in both instances the most popu- lated county is metropolitan. The distribution of the number of counties per area differs between metro and nonmetro service areas. The metro service areas include a larger number of counties than the nonmetro areas, par- ticularly for the 800-linked solution where 16 percent of the 282 metro service areas have 10 or more counties. The percent of single county service areas is higher for non- metro than metro service areas. More than half of the nonmetro areas in the 1400-area solutions are single county areas. The 1987 median population for the metro areas ranged from 233,000 for the 1400-unlinked solution to 362,000 for the 800-linked solution. For nonmetro areas the median population was 30-31 thousand for the 1400- area solutions and 51-56 thousand for the 800-area solutions. MSAs in linked and unlinked solutions This section describes the 800-area solutions for spe- cific MSAs (New York City, Chicago, San Francisco, and Washington, DC) to examine differences between the linked and unlinked solutions in greater detail. In each instance the 800-linked solution yields a single service area for the MSA, whereas the 800-unlinked solution splits the MSA into as many as 6 service areas. The New York MSA consists of the following eight counties: Bronx, Kings (Brooklyn), New York (Manhat- tan), Putnam, Queens, Richmond (Staten Island), Rock- land, and Westchester. In the 800-linked solution these 8 counties joined with a 2-county MSA, Nassau and Suffolk. In contrast, the 800-unlinked solution splits the New York MSA into the following 5 service areas that range in size from 1 to 4 counties: (a) Rockland; (b) Richmond and Kings; (¢) Bronx and New York; (d) Queens, Nassau, and 9 Table C. Availability of health care resources by type of area: United States [Data are for the coterminous United States] Type of area 800 800 1400 1400 Type of resource unlinked linked unlinked linked HCCA HIS PSU County Patient care physicians, 1986 Physicians per 100,000 population 5th percentile. . . .............. 50 47 35 35 40 18 9 28th percentile. . vv os cbs ws sus ma 78 75 60 57 72 46 40 50th percentile. . . .............. 100 98 87 85 99 72 65 78th percentile: ws vw sasmms isis 137 133 124 122 138 114 108 95thpercentile. . .........co00eu. 231 218 224 205 209 204 214 Short-stay hospital beds, 1987 Hospital beds per 100,000 population 5th percentile. . . .............. 236 241 198 206 233 - - 25h PETCeIMIe: 2 sms 5 wm win ve mais 338 348 305 316 350 254 193 50th percentile. . . .............. 429 433 417 424 440 389 344 78th percentile: ; =o «mx vai dis 0 a ti 534 534 538 547 560 556 539 5th percentile... . . cco sms sa nine 836 834 933 1,042 933 1,131 1,090 Short-stay hospitals, 1987 Number of hospitals Sthpercentile. . « «cv insu viens 1 1 1 3 1 - - 25th parcentile. . «cc vu cmv viv ns 3 2 1 1 2 1 1 50th percentile. . . .............. 5 4 2 2 4 2 1 75th peroenlile, . «ws ow svn mus wis 8 8 4 4 7 3 2 OBth percentile... . «5 che swe os vem 20 22 12 13 25 8 5 Selected services, 1987 Percent of areas ICT SCaMNBE + « mv oie 2 wn 3 mnfow 86.3 82.5 69.8 65.2 78.3 53.0 45.5 Cardiac catheterization. . .......... 44.4 40.3 27.9 24.9 38.6 17.7 14.5 Magnetic resonance imaging. . . . . ... 26.1 24.3 16.5 14.8 24.3 10.5 8.4 Suffolk; (e) Putnam, Westchester, Duchess (Poughkeepsie MSA), and Ulster (a nonmetropolitan county). The Chicago MSA includes Cook, DuPage, and McHenry counties. In the 800-linked solution these coun- ties form a 6-county health service area with the Lake MSA (Lake county) and the Joliet MSA (Grundy and Will counties). In the 800-unlinked solution the Chicago MSA counties form a 4-county health service area with Lake county. Grundy and Will counties form their own 2-county service area in the 800-unlinked solution. The San Francisco MSA includes 3 counties (Marin, San Mateo, and San Francisco) that are linked with Sonoma county (the Santa Rosa-Petaluma MSA) to form a 4-county service area in the 800-linked solution. In the 800-unlinked solution these same 4 counties form 2 service areas, one including San Fran- cisco and San Mateo counties and a second including Marin and Sonoma counties. In addition to the District of Columbia, the Washing- ton MSA includes five counties in Maryland and six counties in Virginia. The 800-linked solution yields a 17-county service area for the Washington MSA that includes the entire Washington MSA as well as 5 adjacent nonmetropolitan counties. In contrast, the 800-unlinked solution splits the Washington MSA into 6 service areas that include a total of 23 counties. The six service areas range in size from two to five counties. One Washington MSA county (Calvert, Maryland) links with the Baltimore MSA, and another Washington MSA county (Frederick, Maryland) links with the York, Pennsylvania MSA. Two other Washington MSA counties (Stafford, Virginia and 10 Prince William, Virginia) link with only nonmetropolitan counties. The District of Columbia links with four Maryland counties within its MSA (Charles, Montgomery, Prince George's, and St. Mary’s) and the remaining Virginia coun- ties within the Washington MSA form their own service area (Arlington, Fairfax, Loudoun, and Alexandria). Health care resources Availability of health care resources by type of service area is shown in table C. The 1986 median physician to population ratio ranged from 85 per 100,000 in the 1400- linked solution to 100 per 100,000 in the 800-unlinked solution. The 1987 median hospital bed to population ratio ranged from 417 in the 1400-unlinked solution to 433 in the 800-linked solution. The median number of hospi- tals in 1987 was 2 for the 1400-area solutions and 4-5 in the 800-area solutions. The percent of areas with special- ized hospital services (magnetic resonance imaging (MRI), cardiac catheterization, and computerized tomo- graphic scanner (CT scanner)), was greater for the 800- area solutions than for the 1400-area solutions. All four solutions exhibited substantial variability across service areas in the supply of health care resources per area. For example, for the 800-unlinked solution the number of patient care physicians per 100,000 population in 1986 ranged from 50 at the Sth percentile to 231 at the 95th percentile and the number of short-stay hospital beds per 100,000 population in 1987 ranged from 236 at the 5th percentile to 836 at the 95th percentile. Data on health care resources are also shown for counties and HIS PSUs Table D. Availability of health care resources by type and metropolitan status of area: United States [Data are for the coterminous United States] Type and metropolitan status of area 800-unlinked 800-linked 1400-unlinked Type of resource Metro Nonmetro Metro Nonmetro Metro Nonmetro Metro Nonmetro Patient care physicians, 1986 Physicians per 100,000 population Shporcentile. , «sos nv ins aman 75 42 89 42 64 31 91 32 25th percentile. . . .............. 103 67 112 63 102 53 147 52 Both percentile. . .. .ccws vrs ws snsz 132 84 140 83 133 73 148 72 75th percentile. . . .............. 177 103 183 102 183 95 193 96 Ot percentlle. . «x vo vaie wu iwi oi 274 145 265 143 283 145 280 145 Short-stay hospital beds, 1987 Hospital beds per 100,000 population 5th percentile. . . .............. 235 241 242 239 187 204 237 195 25th percentiles w+ + wi xv w vo vw www a = 321 360 344 353 293 315 326 314 BOthpercentile. . ... 2s ves» aww 5a 393 462 4083 446 382 430 412 427 78thpercentile. . . cv +x vsuovuenas 495 575 509 561 499 567 512 565 OBth percentile. . + vo ss vos on a wsd 705 999 685 876 704 1,106 720 1,147 Short-stay hospitals, 1987 Number of hospitals Sthpercentile. . ............... 2 1 3 1 1 1 2 1 25thpercentile., ....c. vvws viens 5 2 6 2 3 1 5 1 S0thpercentile. . ............... 8 3 10 3 6 2 7 2 78thipercentife. . . cs ix cu vw cans 12 5 16 4 8 3 12 3 O5thpercentlle. .......co 0 susa 31 8 43 7 24 5 37 5 Selected services, 1987 Percent of areas CTscanner. . ................. 99.7 75.0 99.7 73.0 97.8 56.9 99.7 55.4 Cardiac catheterization. . . . ........ 80.3 14.2 89.4 13.0 72.1 75 88.1 6.9 Magnetic resonance imaging. . . . .. .. 48.4 7.3 56.0 6.7 42.0 4.7 52.6 4.0 Note: Metropolitan service areas include at least one metropolitan county. in table C for comparative purposes. There is more variability in health care resources at the county and PSU level because these geographic units are not service areas. For example, more than 5 percent of both counties and HIS PSUs do not have any short-stay hospitals and so residents of those areas must seek hospital care elsewhere. Table D compares the supply of health care resources in metropolitan and nonmetropolitan areas. As expected, physician supply was greater in the metro than the non- metro service areas. For example, in the 800-unlinked solu- tion the median number of physicians per 100,000 population in 1986 was 132 for the metro areas and 84 for the nonmetro areas. The median number of hospitals in 1987 in the nonmetro service areas was about a third of that in the metro service areas (3 compared with 8-10 for the 800-area solutions). In contrast, the median number of hospital beds per 100,000 population was lower for metro than nonmetro service areas (393 compared with 462 for the 800-unlinked solution), reflecting the much larger popula- tion in metro than nonmetro service areas. Nonmetro ser- vice areas were much less likely than metro service areas to have facilities for cardiac catheterization or nuclear mag- netic resonance. CT scanners were available in more than 70 percent of the nonmetro service areas for the 800-area solutions and nearly all metro service areas. Travel for health care An important criterion for a health service area is the extent to which it is self-contained with respect to the provision of hospital care (20). The degree of self- containment of the service areas has been evaluated using two data sources: 1988 Medicare data on routine short- stay hospital stays that were used to define the areas and 1989 National Health Interview Survey data on travel for short-stay hospital stays. Unlike the Medicare data set, which includes information about all counties in the United States, the HIS only includes data on a national sample of the civilian noninstitutionalized population. However, the HIS data provide information on all types of hospital stays (including specialized care) for the popula- tion under 65 years of age as well as those 65 years of age and over. Travel for hospital care has been measured in terms of both outflow (the proportion of stays by residents outside the area) and inflow (the proportion of stays provided in the area to nonresidents). In examining travel for hospital care it is important to consider metropolitan and nonmetropolitan areas separately. Although fewer hospital stays occur in nonmetropolitan areas, they are much more likely to involve substantial travel for care. Thus, counties have been classified according to metropol- itan status and population using categories developed by the Department of Agriculture (15). In addition, certain tables classify service areas according to whether or not they include a metropolitan county. Table E categorizes service areas according to the amount of travel outside the service area for routine hospital stays by Medicare beneficiaries. Service areas with less than 25 percent of residents’ stays occurring outside the service area are providing the great majority of the routine hospital care for their residents and can be 11 Table E. Percent distribution of areas and population according to travel for routine Medicare hospital stays by type of area: United States, 1988 [Data are for the coterminous United States] Type of area 800 800 1400 1400 Travel measure unlinked linked unlinked linked HCCA HIS PSU County Stays outside area by residents Percent distribution of areas Lessthan25 percent. . ............. 65.8 62.0 42.4 40.1 56.1 28.6 22.6 28-49 percent . oc: ume er huy sewn 33.8 37.2 50.5 50.5 37.9 39.6 38.0 S50percentormore. . .............: 0.4 0.8 i dy 9.4 59 31.8 39.4 Percent distribution of population living in areas Lessthan25 percent. . . ............ . 93.1 94.4 84.6 88.5 93.9 83.3 65.3 25-49 PBICBIL . . . vs suave cum vals 6.9 55 14.7 10.5 57 11.8 23.8 50 percentormore. . .............. 0.0 0.0 0.7 1.1 0.4 4.9 10.9 Stays inside area by nonresidents Percent distribution of areas Lessthan25 percent. . . ............ 93.5 93.0 83.9 81.6 82.1 68.6 64.5 25-40 percent i... : vw: sa ww wens wn 6.4 6.8 15.5 17.9 16.8 29.0 31.1 50percent or more. . . ...covvovvssso 0.1 0.1 0.6 0.5 1.2 24 4.4 Percent distribution of population living in areas lessthan25 percent. . ............. 95.9 97.5 90.1 93.2 96.6 84.7 66.0 D849 PErCENt vv 2 wus cw we win ni wy 4.0 25 9.6 6.6 3.2 14.4 30.4 0.2 0.2 0.2 0.9 3.6 BOPDErcent or more. . ... cvs sms nes 0.0 0.0 Table F. Percent distribution of areas and population according to travel for routine Medicare hospital stays by type and metropolitan status of area: United States, 1988 [Data are for the coterminous United States] Type and metropolitan status of area 800-unlinked 800-linked 1400-unlinked 1400-linked HCCA Travel measure Metro Nonmetro Metro Nonmetro Metro Nonmetro Metro Nonmetro Metro Nonmetro Stays outside area by residents Percent distribution of areas Lessthan25 percent. . ............. 90.2 45.4 98.2 41.9 80.1 24.9 95.2 24.5 94.9 35.2 25-49 POICOIY 4 www 5 waves wo 8 9.8 53.9 1.8 56.9 19.4 64.9 4.8 63.5 52 55.7 S0percemtormore. ............... - 0.7 - 1.2 0.5 10.1 - 12.0 - 9.2 Percent distribution of population living in areas Less than 28 parce. « « vu vo sx wv sms 97.0 66.0 99.4 62.4 92.5 46.8 98.1 44.9 99.1 56.0 2B.A0 POICOML + vi vp 3 Bio 5 0 Bb 3.0 33.9 0.6 37.3 75 49.3 19 49.2 0.9 40.7 S0PEICEntor MOB. « «. cov rurins se - 0.1 - 0.3 0.1 3.9 - 5.8 - 3.3 Stays inside area by nonresidents Percent distribution of areas Lessthan 25 Percent. « « «vs ws «os sn oo 94.8 92.4 95.7 91.5 85.5 83.1 87.7 79.8 92.7 76.3 25-49 percent . .. LL... 52 73 4.3 8.3 13.6 16.4 11.3 19.8 6.6 22.3 BO PerCont.Or MOIS. x v5 srw x 2 os # was - 0.2 - 0.2 0.9 0.4 1.0 0.4 0.7 1.4 Percent distribution of population living in areas Lessthan 25 percent. . . . ........... 96.1 95.1 98.2 93.2 90.8 86.7 95.7 81.8 98.1 85.7 28-4 POICEIN + vv pF HEE FEE 3.9 4.8 1.8 6.7 8.9 13.2 4.1 18.0 17 13.7 S50 percentormore. . .............. - 0.1 - 0.1 0.3 0.1 0.2 0.2 0.2 0.5 Note: Metropolitan service areas include at least one metropolitan county. considered adequate service areas. Areas with 25-49 per- cent of stays occurring outside the service area are not as well-defined, but still provide the majority of hospital care for their residents. Areas with 50 percent or more of stays outside the service area are problematic, as a majority of stays by residents are occurring outside the area. A similar classification for travel into the service area for hospital stays has also been made. The proportion of the popula- tion residing in areas with adequate, fair, and poor de- grees of self-containment are also shown in table E. In the 800-linked and 800-unlinked solutions, 62 and 66 percent of areas, respectively, had less than 25 percent 12 of routine Medicare stays occurring outside the areas in 1988 and less than 1 percent of the areas in these 2 solutions had a majority of stays outside the service area. In contrast, 56 percent of the HCCAs had less than 25 percent of stays outside the area and 6 percent of the HCCAs had a majority of stays outside the area. However, the proportion of the population residing in the most self-contained areas was similar for the 800-area solutions and the HCCAs (93-94 percent). Both 800-area solutions had few routine Medicare stays occurring in the areas by nonresidents. Although the pro- portion of HCCAs with fewer than 25 percent nonresident Table G. Percent of short-stay hospital stays outside area of residence by age, type of county, and type of area: United States, 1989 [Data are for the coterminous United States] Type of area Number of stays in 800 800 1400 1400 } HIS Age and type of county of residence sample unlinked linked unlinked linked HCCA PSU County Under 65 years of age Percent of hospital stays AlCOUNtBs . «cove vnnrwrn 8,770 18.6 156.3 22.3 18.2 14.8 19.5 33.4 fy 6,511 14.7 11.1 17.9 12.6 10.6 10.4 27.6 Large Core . .. «csv snr swe 2,329 9.2 6.6 11.2 8.0 6.7 45 16.6 Large finge. «sv wn wns as was 1,321 16.9 9.0 24.8 11.4 8.2 74 43.1 Medium. ............... 02,210 16.9 14.1 18.7 15.3 13.5 15.0 28.7 Small. ; cvs cmswrapams 651 20.8 20.6 23.9 21.5 18.6 20.9 28.5 Nonmetropolitan . . . ......... 2,259 30.6 28.1 35.8 35.4 27.9 48.0 51.3 UBER . chemin emamm nm nw 766 25.7 23.3 28.9 28.9 219 31.6 32.7 less urban. .. «vrwvmrns » 1,205 33.5 30.5 40.1 38.6 29.8 53.9 574 BUA: + 0 moon on simmons: wi 288 32.1 31.7 36.9 39.8 36.9 69.3 79.1 65 years of age and over Alcounties . : « vos cmsms ens 3,321 16.2 13.5 18.9 16.4 14.1 20.0 29.5 Metropolitan. . . ............ 2,281 10.1 7.6 12.2 8.5 7.6 7.9 20.1 LAIGB COMB + vv v minima ints 772 6.1 4.5 7.0 5.7 4.6 4.4 11.3 Largefringe. . ............ 477 12.6 5.6 16.7 7.4 6.7 5.1 32.3 MBI. + «vs ww mininies wine 808 10.9 9.3 13.6 9.6 9.6 10.5 20.5 SIAN. - ov vid Rn nm A 224 14.8 15.4 14.8 15.9 118 15.5 21.8 Nonmetropolitan . . . ......... 1,040 31.0 27.9 34.9 35.4 20.9 49.2 51.9 UMDaN ci vvvennwnn snes 305 22.3 20.0 24.8 26.8 21.1 28.8 31.0 Less Urbal ; vu v su vmm as was 621 35.2 31.2 40.1 38.6 32.9 56.4 58.4 Rural. ; vs conmrnmrme van 114 31.2 31.3 34.0 41.3 37.8 65.1 73.0 Source: National Health Interview Survey, 1989 Table H. Percent of routine Medicare hospital stays outside area of residence, by type of county and type of area: United States, 1988 [Data are for the coterminous United States] Type of area 800 800 1400 1400 HIS Type of county of residence unlinked linked unlinked linked HCCA PSU County Percent of hospital stays } ALCOUNHBS . , v:ussmrms swe 12.8 115 16.0 14.0 11.5 16.2 24.9 Metropolitan. . « «wx wis vis oes 9.3 7.3 11.5 8.3 7.1 7.1 18.4 Largecore ............. 7.0 5:4 8.8 5.9 54 4.1 © 129 Large Hinge. « « « suv mes 11.4 7.8 16.1 9.4 7.6 6.6 28.7 Medium. .............. 9.6 8.1 11.0 8.7 7.9 8.8 17.8 SMa, . vive rs nnamy ay 11.0 10.7 12.0 12.1 9.8 12.8 17.0 Nonmetropolitan . . . ........ 21.0 21.2 26.4 27.2 21.8 37.0 40.0 DB» vovsmiras sms Eins 15.3 15.5 18.3 19.7 15.0 22.2 23.2 LESS UTDAN . «ws sus vie wins 22.1 22.3 28.7 29.2 23.2 39.8 42.5 Rural, ; ., cores vmsmsmns 29.7 29.6 35.9 36.4 31.7 59.4 68.0 stays was somewhat lower than for the two 800-area solu- tions (82 percent compared with 93 percent), the percent of the population residing in such areas was similar for the HCCAs and the 800-area solutions (96-98 percent). The 1400-linked and 1400-unlinked solutions had 40-42 percent of areas with fewer than 25 percent of 1988 routine Medicare stays occurring outside service areas. In contrast, only 29 percent of HIS PSUs and 23 percent of counties had fewer than 25 percent of stays outside the area. The population residing in adequate service areas represented 89 percent of the population for the 1400- linked solution and 85 percent of the population for the 1400-unlinked solution. The comparable population pro- portions for HIS PSUs and counties were much lower, 69 percent and 65 percent, respectively. Table F compares travel patterns for routine hospital stays by Medicare beneficiaries for metro and nonmetro service areas. These results show that metropolitan areas exhibited less outside-area travel than nonmetropolitan areas. The vast majority of metropolitan service areas (80-98 percent across the 4 solutions) had less than 25 percent of stays outside the area of residence. In contrast, 13 the majority of the nonmetropolitan service areas (54-65 percent across the 4 solutions) had 25-49 percent of stays outside the area of residence. The HCCAs showed a similar pattern. Table G presents the percent of short-stay hospital stays outside the area of residence based on the 1989 National Health Interview Survey and compares travel patterns for persons under 65 years of age with those 65 years and over. Overall, the percent of hospital stays outside the area of residence was similar for both age groups for all four solutions, supporting the use of Medicare data to identify service areas that can also be used for those under 65 years of age. The greatest differences by age occurred for residents of metro counties where those under 65 years of age were somewhat more likely than older persons to travel outside the service areas (11-18 percent of stays for those under 65 years of age compared with 8-12 percent of stays for those over 65 years). Table H presents results comparable to those shown in table G, based on the 1988 Medicare data. The percent of routine hospital stays by Medicare beneficiaries outside the area of residence was slightly lower than that for all hospital stays by persons 65 years of age and over based on the HIS (table G). This is consistent with the restriction of the Medicare data base to routine stays that are less likely 14 to require travel. Overall, the percent of routine Medicare hospital stays outside the area of residence ranged from 11.5 percent for the 800-linked solution to 16 percent for the 1400-unlinked solution. The percent of stays outside the area of residence was 11.5 percent for HCCAs and 16 percent for HIS PSUs. In contrast, 25 percent of routine Medicare hospital stays occurred outside the county of residence. The linked solutions yielded slightly less travel outside service areas than the unlinked solutions for large and medium metro counties. Among nonmetro counties, the linked and unlinked solutions were almost identical with respect to outside area travel. The percent of stays outside service areas for residents of nonmetro counties was 21 percent for both 800-area solutions and 26-27 percent for the 1400-area solutions. The 1400-area solu- tions yielded relatively high percents of stays outside service areas for residents of less urban and rural counties (29 percent and 36 percent, respectively). The comparable figures for the 800-area solutions were 22 percent and 30 percent. Although there was relatively little travel outside the HIS PSUs for residents of metro counties, the PSUs did not function well as service areas for nonmetro counties. For residents of nonmetro counties 37 percent of routine Medicare hospital stays occurred outside the PSU of residence, about the same as for counties. Summary Health service areas were identified for the cotermi- nous United States on the basis of travel patterns between counties by Medicare beneficiaries for routine hospital care. We used agglomerative hierarchical cluster analysis and the average linkage method to group counties into service areas. Four alternative solutions were generated that differed with respect to the number of areas and whether or not counties within an MSA were linked within the same service area. The results showed that all four alternative solutions produced service areas that were adequate for the major- ity of the U.S. population. However, the 800-area solu- tions were more successful than the 1400-area solutions in the less populated areas of the country that are of partic- ular interest in studying availability of health care re- sources. Thus, the 800-area solutions are preferable to the 1400-area solutions. The 800-linked and 800-unlinked so- lutions differ in the following respects. The 800-linked areas do not split MSAs into more than one service area, are less likely to include more than one MSA, and more likely to have no MSAs. The 800-unlinked areas split 63 MSAs into 2 or more service areas, are more likely to link a nonmetro county with a metro county, and have a more uniform distribution of number of counties per area. Uniformity of size is a desirable characteristic in using the areas for small area analysis in health services research. Thus, although the 800-linked solution may be preferable for certain uses, the 800-unlinked solution appears to be the preferred solution for most applications. Comparison of the new health service areas to the HIS PSUs, geographic units that have been used to approximate service areas (3), showed that the HIS PSUs are just as self-contained as the new service areas in metropolitan areas. However, the HIS PSUs do not per- form well as service areas for nonmetropolitan counties. This is not surprising because the HIS PSUs consist of complete MSAs in metropolitan areas, but are often only single counties in nonmetropolitan areas. Despite the fact that the HCCAs are based on data that are now 20 years old, the percent of short-stay hospital stays outside the HCCAs was similar to that for the new 800-unlinked service areas. The HCCAs were less self-contained than the new areas for nonmetropolitan service areas with small populations. Although the service areas for specific counties, particularly in nonmetropolitan areas, may have changed, the picture for the nation as a whole is that service areas appear to be fairly stable over time. 15 References 10. 16 Paul-Shaheen P, Clark JD, Williams D. Small area analysis: A review and analysis of the North American literature. J Health Polit Policy Law 12(4):741-97. 1987. Transactions Systems Inc. Evaluation of alternative health area definition methods. DHEW Contract No. HRA 230-75-0080. 1976. Mitchell JB, Cromwell J. Variations in surgery rates and the supply of surgeons. In Rothberg DL ed. Regional variations in hospital use. Lexington: DC Heath and Company, 103-29. 1982. Makuc DM, Kleinman JC, Machlin SR. Effects of physician supply on use of physician services. In: American Statistical Association 1983 proceedings of the social statistics section. Toronto: American Statistical Association. 299-303. 1983. Shannon GW, Pyle GF. The definition and measurement of hospital market areas. Unpublished report. July 1989. Garnick DW, Luft HS, Robinson JC, Tetreault J. Appropri- ate measures of hospital market areas. Health Services Research 22(1):69-89. 1987. Codman Research Group, Inc. Relationship between de- clining use of rural hospitals and access to inpatient services for Medicare beneficiaries in rural areas. Lyme, New Hamp- shire. Technical report. January 1990. Stigler GJ, Sherwin RA. The extent of the market. J] Law Econ 28:555-85. 1985. Makuc DM, Kleinman JC, Pierre MB. Service areas for ambulatory medical care. Health Services Research 20(1): 1-18. 1985. Tolbert CM, Killian MS. Labor market areas for the United States. Agriculture and Rural Economic Research Service, U.S. Dept of Agriculture. Staff report no. AGES870721. August 1987. 12. 13. 14. 15. 16. 17. 18. 19. 20. . Humphrey AB, Scherzer GD, Marshall RJ, Buechner JS. A methodology for determining hospital service areas with marketing and planning applications. J Hosp Mkting 2(2):87-97. 1988. Tedeschi PJ, Wolfe RA, Griffith JR. Micro-area variation in hospital use. Health Services Research 24(6):729-39. 1990. Kleinman JC, Makuc D. Travel for ambulatory medical care. Medical Care 21:543-57. 1983. American Hospital Association. Hospital Statistics, 1987 edition, 220. Health Resources and Services Administration. Area Re- source File. March 1989. National Center for Health Statistics. Vital statistics of the United States, 1987, vol II, mortality, part A. Washington: National Center for Health Statistics. 1988. Hubert LJ. Hierarchical cluster analysis. In: Kotz S, Johnson NL eds. Encyclopedia of Statistical Sciences vol 3. New York: John Wiley and Sons. 1983. Anderberg MR. Cluster analysis for applications. New York: Academic Press. 1973. SAS Institute Inc. SAS User’s Guide: Statistics, Version 5 ed. Cary, North Carolina: SAS Institute Inc. 1985. Morrisey MA, Sloan FA, Valvona J. Defining geographic markets for hospital care. Law and Contemporary Problems 51(2):165-94. 1988. Appendixes Contents I. ARCTative diStATIEE MICHIE « «wv aos 54510 i om wae song in Bia) wseomis 4 nsw a 00010 670 30 dm imssns ios co #0 hen Sn a) 4 #0 wig 1.8) tons II. Maps of health service areas for the 800-area unlinked solution............. coon... ITI. Listings Of health SOrVICES ATCAG. + wasn 07 58mm sh vss 5g 5 EEHE BS 0208 90 580580 HE BF 50m 0m 5199 0 BBD 6908 pom ome ie 4019 5.3 IV: ODSICITIC SCIVICE UCAS. vb 5.0 w0 wis 00s 45 05508 5905 005 419.4 750 958 50 8 90 500 06 0 0k 6d 6.00 550 490038 05 07 8.0 0k 0 570 908 310 18 008 0 Bo 4 List of appendix figures LY Health service areas in Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut. . .. II. Health service areas in New York, New Jersey, Pennsylvania, Delaware, Maryland, Washington, DC, Virginia, nud WESE VATEIIIA. .o tnvev ws 500 1950 50 8uttbn ve wim 08 8 00 0080 mc 9 oom 960 90 5.8 im vo 0 or 8 ron mt ly 00 0 ITI. Health service areas in Ohio, Indiana, Illinois, Michigan, and Wisconsin ..............coviueiirnnnnen..n. IV. Health service areas in Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas. .... V. Health service areas in North Carolina, South Carolina, Georgia, and Florida ............................ V1. Health service areas in Kentucky, Tennessee, Alabama, and MISSISSIPPI. + «corn vrs nssnnnss vnsesansdvnsvns VII. Health service areas in Arkansas, Louisiana, Oklahoma, and TeXas. ..... couture inne VIII. Health service areas in Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, and Nevada ..... IX. Health service areas in Washington, Oregon, and California............... cc... List of appendix tables I Alphabetical list of State, county, and health service area numbers for four alternative solutions ............ II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside S00-unlinked Brea. .... a uirmrnvrtrmsivssmenrns smvesists seme nesss III. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents GuisiAe arSa8 IN A988 +. vy sus nus bios ov ws 0s Win vss o 310 2 00 85 H08 3 0800 54.018 5700858308 an 50 0092010 IV. Percent distribution of obstetric service areas and population according to travel for hospital births, by type Of aren: United States, 1084-80... covuiun smsissss vos moe msm meson esos ss ems we dash sss memos V. Percent of hospital births outside area of residence, by type of county and type of area: United States, JOBL-BB ov msmnimnmaman ns imodnn sso nassdssos diam sh orm 06 EE 80 A 4 010550 5500 004 on A060 5 0 5.08 00% 30 oh 04 0 sk 7 19 17 Appendix | Alternative distance measures The following five alternative distance measures were considered for possible use in defining health service areas: A 1-F;/G itp > Fp, 1-F,/G, otherwise. B. 1-(F; + F;)/(C, + C) C. 1-F;/(C-H) itp, > P, 1-F;/(C;-H) otherwise. E. 1-(Fy + F)I1C ite, > p, 1-(F; + F)16 otherwise. See page 4 for definitions of F, C, P, and H. All five measures contain a ratio of the flow of services between two counties and the consumption of services in one or both counties. The ratios are subtracted from 1 so that the distance measures are larger for pairs of counties with less interaction. All five measures range from 0 to 1. Three of the five measures take into account the relative size of the two counties, as measured by production of services. Distance A relates the flow of demand for services to the larger county from the smaller to the total consump- tion of services by residents in the smaller county, that is, the proportion of total services used by residents of the smaller county that are provided by hospitals in the larger county. Distance B relates the sum of flow between two counties to the total consumption of services by the residents of both counties. Distance measures C and D are 18 similar to A and B, respectively, except that the home consumption of services is excluded from the denomina- tor. Distance E relates the sum of flow between two counties to the consumption of services by the residents in the smaller county. Distance E is the measure used by the U.S. Department of Agriculture to define Labor Market Areas. Measures B, C, and D have some undesirable proper- ties. Measure B is often close to 1 because the total flow between two counties (the numerator) will often be quite small in relation to the total consumption of services by residents of both counties (the denominator). Distance measures C and D only take into account consumption which takes place outside the county of residence. Thus, the total demand for health care affects these measures much less than the other measures. As a result, a county that is practically self-supporting with an insignificant number of occurrences in other counties could be clustered with an adjacent county, rather than a smaller county that is depen- dent on services in neighboring counties. Of these five measures, A and E appear to make the most sense. It follows from the definitions that E assumes values less than or equal to A. If there is no patient flow from the larger to the smaller county, E will equal A, and the difference between A and E will increase with increas- ing patient flow from larger to smaller counties. Measure A has the advantage that it is a proportion and that it emphasizes the most important direction of flow, from a smaller to a larger county. However, measure A has the disadvantage that it ignores patient flow from a larger to a smaller county, even when two counties are similar in size. Appendix II Maps of health service areas for the 800-area unlinked solution Figure I. Health service areas in Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, and Connecticut ___ State boundary — Service area boundary 19 Figure II. Health service areas in New York, New Jersey, Pennsylvania, Delaware, Maryland, Washington, DC, Virginia, and West Virginia 20 80 Ld) 59 56 105 65 106 oF = 19 117 128 | 47 2 125 oe 34i5 78 12 26 2 84 59 42 110 as 139 57 28 100 “= \ 72 31 } 5 16 34 28. 92 69 50 109 51 107 63 3 7 135 60 3 vv, 23 13 3 14 70 i” 151 73 79 132 130 6 81 88 76 10 58 112 41 86 3 > 94 ’ 87% 36 J } 83 93 66 2 113 108 23 127 } 64 ___ State boundary — Service area boundary Figure Ill. Health service areas in Ohio, Indiana, lllinois, Michigan, and Wisconsin F — State boundary — Service area boundary 21 Figure IV. Health service areas in Minnesota, lowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas ____ State boundary — Service area boundary 22 Figure V. Health service areas in North Carolina, South Carolina, Georgia, and Florida 4 hse) 149 170 21a ¢ 243( 264 24 86 FO 29 dAR2 19 264 225 235 20 183 200 218 14d 258 167 208 262 Li 5 191 {,,., 242 145 173 182 157 hi 49 207 154 A 18 160 184 Y199 és 164 196 17 153 S i fh 15 176 24 72 : 26 250 212 193 222 166 _ 197 43 189 230 ven 144 23 24 174 = 228117 180 1 155 183 a 159 , 233 142 227 a7 b3 202 266 213 _ 165 200 P) — State boundary ss , LB — Service area boundary 23 Figue VI. Health service areas in Kentucky, Tennessee, Alabama, and Mississippi 188 146 482 484 412 » 175 1 171 455 259 (] 162 409 161 ___ State boundary — Service area boundary 24 Figure VII. Health service areas in Arkansas, Louisiana, Oklahoma, and Texas 587 535 - 429 =a i ys 472 co 446 “07 494 =i 81 445 4 405 oy 440 | 475 417 529 522 521 1 457 es 499 -i £04 478 442 432 43 483 430 tee ot “ B34 527 3 497 448 481 458 474 43 438 404 9 473 406 420 495 514 485 4% 480 4 519 428 479 53 485 52 Cs 434 441 439 443 503 4. 518 507 454 489 416 )530 415 492 “ws . on 4 415 444 447 450 426 452 501 449 aes 423 464 5 492 425 13 419 306 470 2 51 408 424 J $80 2) 403 467 536 410 505 2 513 433 531 om 437 427 520 ___ State boundary _ Service area boundary Figure VII. Health service areas in Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, and Nevada 734 784 729 716 26 733 728 7312 779 692 767 810 714 21 778 691 714 756 773 749 770 586 722 804 808 777 775 -_ 696 726 730 “ll 71 580 744 A 792 aT 763 708 760 703 809 754 pS — 745 707 rr SL 724 Bia 740 785 699 801 72% 769 772 732 508 700 ~~ heed. ___ State boundary ___ Service area boundary Figure IX. Health service areas in Washington, Oregon, and California ___ State boundary — Service area boundary 27 Appendix Ili Listings of health service areas Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked AL Autauga . ........ 171 274 138 231 AL Washington. . . .... 161 264 145 238 AL Balgwin .. uu Lh. 161 409 145 238 AL WIHCOX, 2 vu 2i5i3 56 5 175 278 158 251 AL Barbour ... ..00 0 162 411 147 374 AL Winston ......... 219 347 195 344 AL Bibb: sowssvwnns 150 395 148 241 AL BIOUY «v.05 06 wh wo 150 396 148 241 AR Arkansas. . . uv vs 527 910 411 859 AL BUNOCK., + «i ov 5s 171 274 138 231 AR Ashley . ......... 480 766 469 746 AL BUBBY. «van mains 259 458 253 445 AR Baxter . . ..xs cx ae 407 686 393 665 AL Calhoun . ........ 177 280 161 254 AR Benton: «v.05 05: 446 802 436 784 AL Chambers. ....... 172 406 155 392 AR BOONES i: votive 494 788 483 765 AL Cherokee . ....... 224 338 201 299 AR Bradley. . . «00 v5 + 473 757 461 736 AL CHRON, + vd 5 wives 150 252 148 241 AR Calhoun . ........ 486 777 474 755 AL Choctaw . . ; «+ sss » 412 691 396 377 AR Carroll vv v5 vanes 494 788 483 765 AL Clarke .......... 161 264 145 238 AR CHICO +504 + wonissiins wun 480 766 469 746 AL Clay. ou vmama sms 241 372 223 341 AR Cleat os sv view oma 448 729 438 711 AL Cleburne. . . ...... 177 280 161 254 AR OlaY: cos us vnnmis 571 982 557 958 AL Colles . uve snes 162 265 147 240 AR Cleburne. . ....... 457 738 444 717 AL Colbert. ......... 219 332 195 291 AR Cleveland . ....... 473 757 461 736 AL Coneguh. «civ vss 163 329 180 274 AR Columbia ........ 486 777 474 755 AL CO08a ....00 4am 241 372 223 342 AR Conway . ....u000s 432 849 411 682 AL Covington... .. .v. 171 360 216 320 AR Craighead. . ...... 571 723 557 704 AL Crenshaw . ....... 171 360 216 320 AR Crawford. vs = + 421 700 409 680 AL Cullman ......... 150 397 148 414 AR Crittenden. . ...... 499 794 399 670 AL Dae. . ss sus sass 162 265 147 240 AR Cross. .......... 499 794 399 775 AL Dallas. . ......... 175 278 158 251 AR Palas. .: we sms nny 496 791 485 767 AL DeKalb. ......... 224 366 201 331 AR Desha .......... 473 858 461 845 AL ENOL . a5 a 5 wi +4 179 282 138 231 AR Drew . ...q's sues» 473 757 461 736 AL Escambia . ....... 163 329 180 274 AR Faulkner. . . «55 432 711 411 682 AL Blowah: os coo moms 224 338 201 299 AR Franklin ......... 421 700 409 680 AL Fayette. . ........ 247 379 229 369 AR Fulton .... .»: w+ 574 977 575 954 AL Franklin ......... 219 347 195 291 AR Garland . ........ 448 729 438 711 AL OBNGVA, + 5 3 wsws ma 162 265 147 240 AR Grant , «wavy ow 432 71 411 682 AL Greene. ......... 156 258 142 235 AR Greens... .......: 571 982 557 958 AL HAG. . c snp svn 156 258 142 235 AR Hempstead . . ..... 404 740 427 699 AL Hemty. .......... 162 265 147 240 AR Hot Spring. . ....«: + 448 862 411 820 AL Houston « . oo vu iw. 162 265 147 240 AR Howard. . ........ 404 880 427 849 AL JACKSON .. . vv ov oi0 mb 210 445 237 429 AR Independence . . . .. 574 958 560 936 AL Jefferson. . ....... 150 252 148 241 AR lzard . .......... 574 958 560 936 AL Lamar. . c:smsmsee 461 743 450 361 AR JACKSON , yw 5i 0 mus 521 904 315 891 AL Lauderdale . ...... 219 332 195 291 AR Jefferson. . . ...... 473 757 461 736 AL Lawrence . cs vv wns 185 288 169 262 AR JORASON . vv www» 522 905 516 892 AL Led ...vuuuwivuns 179 282 138 304 AR Lafayette. . . ...... 404 683 427 699 AL Limestone. ....... 210 320 194 290 AR Lawrence ........ 571 723 557 704 AL Lowndes. ........ 171 274 138 231 AR 108, cvs wsnnms 499 794 399 775 AL Macon ....% .cens 179 282 138 231 AR LINCO. «vx vena 473 757 461 736 AL Madison... «as 0 on 210 320 194 290 AR Little River. . +.» 4 + 404 683 427 699 AL Marengo. ..« us « «ha 175 404 158 391 AR L008 ou hve 421 700 409 680 AL 7 247 379 229 370 AR Lonoke. .. cv evws 432 ™ 411 682 AL Marshall . ........ 224 366 201 332 AR Madison . ........ 446 727 436 709 AL MOBS 5 40 4 0 5 0» 161 264 145 238 AR Marion . ......... 407 686 393 665 AL Monroe. . . ....... 161 449 145 419 AR Miller. . .s:opswss 404 683 427 699 AL Montgomery . ..... 171 274 138 231 AR Mississippi. . . . .... 614 1010 598 983 AL MOIGan. ., . ca av 185 288 169 262 AR Monroe. . .: sess» 527 910 411 859 AL PITY « «von ve wens 175 278 158 251 AR Montgomery . ..... 448 729 438 711 AL Pickens ...:::s0: 156 391 142 235 AR Nevada, , « «a vom 404 740 427 699 AL PHBL. + on vives cls 4p 8 171 274 235 427 AR NeWION. iv» 20% 50 494 788 483 765 AL Randolph ........ 172 407 155 393 AR Ouachita. . + »vww vu 496 791 485 767 AL Russell. . ........ 166 269 158 246 AR POITY vo sian sma la 432 711 411 682 AL Shelby . css s5i0ins 150 252 148 241 AR Phillips . . . ....... 537 920 525 901 AL StCar .....00.5 150 252 148 241 AR Pike, .oomp smn 4 448 729 438 71 AL Sumter ....oo vw. 45 412 691 396 668 AR Poinsett . . ....... 571 723 557 704 AL Talladega ...:.-:.. 241 372 223 342 AR PO. +o oom spew 534 917 501 877 AL Tallapoosa. . . ..... 179 282 138 304 AR PORB & «ive nv alin 3 442 722 430 702 AL Tuscaloosa . . ..... 156 258 142 235 AR PHANG «vv v0 win ve 432 711 411 682 AL Walker . ......... 150 441 148 241 AR Pulaski. ... unis 432 711 411 682 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked AR Randolph + ccs 50 571 723 557 704 CA SantaClara . . ..... 751 1268 736 1248 AR Saline. .......... 432 711 411 682 CA SantaCruz ....... 802 1403 790 1389 AR Sool. uv rmawn sms 421 887 409 811 CA Shasta. ......... 710 1223 696 1204 AR Searcy . ......... 494 788 483 765 CA SISIB. wv 5% vw 50 753 1270 738 1250 AR Sebastian . . ...... 421 700 409 680 CA SISKIYOU . . «nes view 752 1373 737 1359 AR SBVIEr. = 2: cae naw 404 747 427 727 CA Solano. ......... 746 1262 750 1266 AR Sharp. .......... 574 958 560 936 CA Sonoma. . ww x wus 764 1284 754 1293 AR St. Francis. cx :05 +5 499 794 399 775 CA Stanislaus . . . ..... 737 1312 723 1302 AR Stone. .......... 574 854 560 841 CA Sutter... wavcmi nes 690 1202 731 1272 AR Bon. . oes smates 486 777 474 755 CA Tehama ......... 697 1350 684 1337 AR Van Buren. ....... 432 711 411 682 CA THORY. oi is munis 710 1223 696 1204 AR Washington. . .. ... 446 727 436 709 CA TAG. « ri 5a 789 1390 777 1376 AR White. .....conms 457 738 444 717 CA TUOIMPG vv vn 4 737 1368 735 1346 AR Woodiif. . «von s 457 738 444 717 CA Yerfura. « + sw ziwa a 790 1391 778 1377 AR Yell, os sawn wees 442 722 430 702 CA YOR: «one vin mows 0 = 709 1222 731 1242 CA Yuba . voir sv mome 8 i 690 1202 731 1272 AZ APBONER. w « vi. sms 3 765 1285 747 1263 AZ Cochise: sv cosws 700 1213 687 1195 co AGAMS sini vos ow 688 1200 691 1199 AZ Coconino . .. ...ux 699 1241 686 1224 co Alamosa . ........ 731 1245 718 1228 AZ Bla. sonws smims 699 1212 686 1194 co Arapahoe ........ 688 1200 691 1199 AZ Graham ......... 700 1327 687 1315 co Archuleta. .; ; :: z+ 740 1292 726 1278 AZ Greenlee. . . ...... 700 1327 687 1315 co BACH mis vo 20% am we 562 1386 547 1372 AZ - Maricopa. . . ...... 699 1212 686 1194 co Bent «vo sa mrwE es 745 1261 730 1241 AZ Mohave ... . +: 5x 803 1404 764 1339 co Boulder ......... 795 1396 784 1383 AZ Navajo. ......... 740 1255 726 1236 CO Chaffee. ....s mamas 786 1383 7H 1369 AZ Pima; ns cmamans 700 1213 687 1195 CO Cheyenne. . ...... 754 1381 691 1334 AZ Pinal ........... 699 1212 686 1194 co Clear Creek. ...... 688 1200 691 1199 AZ SantaCruz ....... 700 1213 687 1195 CO Conejos . ........ 731 1245 718 1228 AZ Yavapal «. «concn 699 1241 686 1224 CO Costilla, . s+ snows 731 1245 718 1228 AZ Yuma...» woven 00 787 1388 775 1374 CO Crowley ......... 745 1261 730 1241 CO Custer ...+ vs ies 812 1413 794 1393 CA Alameda. ........ 766 1286 750 1365 CO Dolla : cov wmsnc aa 761 1377 744 1363 CA ABING. - vvvvn vn 701 1265 688 1245 CO DeNVeE. «x v.v sim x wn 688 1200 691 1199 CA Amador . : cia. 750 1335 731 1242 CO Dolores. « : ux suze 740 1293 726 1279 CA BUS ... «var wms as 697 1210 684 1192 CO Douglas . ........ 688 1200 691 1199 CA Calaveras . ....... 750 1267 735 1247 co Eagle... vevws sind 711 1224 697 1205 CA Colusa .......... 690 1367 731 1273 co E) PESO ; 10 «0 wir wamn » 754 1271 691 1251 CA Contra Costa . . . . .. 766 1286 750 1365 co EIDBH: ov « «svn smu 688 1200 691 1199 CA Del Norte ........ 738 1252 724 1234 co Fremont . ..5 5. cov. 812 1413 794 1393 CA Bl Dorado ; « v5.4.5 + 709 1334 731 1242 co Garfield. . + + van ww 711 1224 697 1205 CA Fresno. ......... 718 1231 704 1212 co GUD. 500 20 dims 688 1200 691 1199 CA BIEN. «io ui emis nas 697 1210 684 1192 co Grant: v + «wv suas 688 1200 691 1199 CA Humboldt . . ...... 800 1401 789 1388 co Gunnison... ou 761 1280 744 1259 CA Imperial... voi ne 774 1309 756 1299 CO Hinsdale . . ....... 761 1280 744 1259 CA INYO. wvn wim sma sd 816 1419 803 1403 [efe] Huerfano. . . « ws «4 « 704 1217 690 1198 CA BIN «4 ui x 0 vores an 807 1408 799 1398 CO Jackson ......... 771 1295 752 1281 CA KINGS + » 5 vo 2s a3 718 1329 704 1317 co Jefferson. . . vv wus 688 1200 691 1199 CA Lake ........... 746 1362 750 1266 co Kiowa. .......... 745 1261 730 1241 CA Lassen. . yarn en 780 1342 763 1329 CO Kt Carson. ... us 754 1346 691 1404 CA Los Angeles. . . .... 723 1236 710 1219 co LaPlta. o vous ois ms 740 1292 726 1278 CA MECBIR, . 4 sis v5 5 a 718 1231 704 1212 CO Lake: 2 0 own wore ms 786 1384 772 1370 CA Manhi. . cz ca: 5000 764 1284 754 1292 co Larimer. . vies vis 796 1397 785 1384 CA Mariposa. . . ...... 737 1251 723 1233 co Las Animas . ...... 704 1308 690 1297 CA Mendocino . ...... 811 1412 770 1368 CO Lheoin. . wo: x50 754 1271 691 1251 CA Merced. « «.c. «ow mes 737 1251 723 1233 co LOGAN ©. 2.50 555 x wena 763 1283 746 1262 CA Modo.» vo0 93 wwe 710 1361 696 1351 CO Mesa........... 711 1224 697 1205 CA MBNO. . 5 5 508 3 0 ass 816 1419 803 1403 CO Mineral. . ........ 731 1338 718 1325 CA Moriterey. ... evs wa» 751 1385 773 1371 CO Moffat. . . ........ 735 1249 722 1232 CA Napa... cvcevviss 746 1262 750 1266 CO Montezuma . . . . ... 740 1293 726 1279 CA Nevada. ......... 753 1270 738 1250 co Montrose. . . ...... 761 1280 744 1259 CA Orange, «ui %s vais nw 723 1379 710 1366 CO Morgan. . . ous sus 760 1279 743 1258 CA PICBE. : 55 4 2h 2 709 1222 731 1242 co OIG, svn asm ms 745 1261 730 1241 CA PIUMES.. ... «is « « sr 2ius 780 1343 763 1330 CO OUIBY.. + +15 «woo vaim » 761 1280 744 1259 CA Riverside. . . ...... 768 1288 764 1338 co Park, .oopmws swan 688 1200 691 1199 CA Sacramento. . ..... 709 1222 731 1242 co PhDs «5 2505 4 wes 763 1283 746 1262 CA SanBenito ....... 751 1268 736 1248 co Pitkin. .......... 714 1339 697 1326 CA San Bernardino. . . . . 768 1288 764 1338 co Prowers. . « «« «vs ws 745 1378 730 1364 CA San Diego. . ...... 774 1310 756 1298 co Pueblo: . zw i ria 704 1217 690 1198 CA San Francisco . . ... 757 1275 754 1292 co Rio Blanco. . . ..... 711 1224 697 1205 CA San Joaquin . ..... 750 1267 735 1247 co Rio Grande . ...... 731 1338 718 1325 CA San Luis Obispo . . . . 781 1355 765 1345 CO ROMY . oo vom 735 1249 722 1232 CA San Mateo...» +» «uu 757 1275 754 1292 CO Saguache.. . + «: xs + 731 1338 718 1325 CA Santa Barbara . . . .. 781 1356 765 1344 co San Juan . .q nes 740 1292 726 1278 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions — Con. Service area number Service area number 800 1400 800 1400 800 © 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked co San Miguel . ...... 761 1280 744 1259 FL Pasco. .......... 227 343 240 432 co Sedgwick ...... 763 1328 746 1316 FL Pinellas. .......:. 227 413 240 432 co Summit. . ........ 688 1200 691 1199 FL POMC, 5 4) 5 cvs ww 202 312 188 284 co ToUBE. + ov vmews bs 754 1271 691 1251 FL PUlnam: « «vs sas 251 420 146 407 co Washington . . . . . .. 760 1294 743 1280 FL SantaRosa . ...... 163 266 180 274 co Weld ........... 760 1279 743 1258 FL Sarasota. ........ 213 323 196 327 co YUMA. i: nnininss 760 1349 743 1336 FL Seminole. . . ...... 142 244 215 318 FL St. Johns ........ 251 421 139 232 CT Fairfield ......... 121 223 116 209 FL RT 221 335 226 395 CT Hartford . ........ 4 5 62 161 FL Sumter. ......... 265 464 215 318 cr Litchfield. . ....... 85 97 62 65 FL Suwannee. . ...... 159 302 133 275 cr Middlesex . . . ..... 85 180 62 161 FL Taylor. .......... 183 286 166 259 CT New Haven . ...... 85 97 62 66 FL ION ow wei ao 9 159 261 133 226 CT New London ...... 20 104 95 188 Fl. VOIISIR. «vow own 142 263 146 239 CT Tolland. ......... 4 5 62 161 FL Wakulla ......... 183 286 166 259 CT Windham ........ 4 5 95 188 FL Wallory . sv viv in ve 163 295 177 270 FL Washington. . . .... 155 325 141 294 DC The District . . ..... 61 78 16 16 GA APOING: = 4.5 wk 5500 230 349 208 311 DE Kent. . vc cus mas 3 57 37 37 GA Atkinson . . ....... 236 363 217 322 DE New Castle . . ..... 75 168 101 194 GA Bacon .......... 174 277 157 371 DE SUSSEX... « «www vs 3 57 37 37 GA Baker........... 144 246 135 228 GA Baldwin ......... 201 311 187 283 FL Alachua ......... 159 261 133 226 GA Banks, sve an mses 157 259 143 236 FL Baker. ...: cosmos s 158 260 139 232 GA Barrow. ......... 164 267 134 227 FL Bay............ 155 257 141 234 GA Barlow , «sch ax on 154 387 140 380 FL. Bragiord , «zo 4 4 159 261 133 226 GA BenHill ......... 189 292 174 267 FL Brevard. . ........ 287 364 218 325 GA Boren. «ovis ove 180 344 164 305 FL Broward . . ....... 200 432 186 415 GA BIB: 4 5 0.x we chien 5 193 300 178 271 FL Calhoun . ........ 155 356 213 316 GA BleokIBY « + « + ve v « + 193 339 178 300 FL Charlotte. . ....... 213 323 196 292 GA Brantley ......... 158 305 139 277 FL CRIS. ons 5 6 monies 283 355 212 315 GA Brooks . . .... Lu. 180 283 164 257 FL Cay. ........... 158 260 139 232 GA Bryan, «+ ovina 143 245 150 243 FL Collier .......... 165 418 152 399 GA Bulloch. ......... 222 336 200 298 FL Colmbia «.. vss 159 302 133 275 GA Burke. . ......... 152 428 149 396 FL Dade .....omuimmi 200 310 186 282 GA Bulle ocov vias an 204 314 134 227 FL DeSoto . ...onvv 213 323 196 292 GA Calhoun ......... 144 342 135 303 FL Dig ....0ccms 159 261 133 226 GA Camden... vvv si 158 305 139 277 Fl. Duval, orsnls see ve 158 260 139 232 GA Candler ......... 222 336 200 298 FL Escambia . ....... 163 266 180 274 GA Carrol os von vn 177 299 161 272 FL Flagler . ......... 142 263 146 239 GA Catoosa . ........ 141 243 182 278 FL Franklin ......... 183 286 166 259 GA Charlton . ........ 158 383 139 232 FL Gadsden. ........ 183 286 166 259 GA Chatham. ........ 143 245 150 243 Fl. Gilchrist , wu suse 159 261 133 226 GA Chattahoochee. . . . . 166 269 153 246 FL Glades. ......... 165 268 152 245 GA Chattooga. . . ..... 154 256 140 233 Fl. Cul cnsmmameinms 155 257 141 234 GA Cherokee . ....... 190 293 134 227 FL Hamilton. ........ 255 454 243 435 GA Clarke .......... 164 267 163 256 Fl. Hardee. vou vw w+ 202 312 188 284 GA Clay. oo vv vn snes 144 342 135 303 FL Hendry.......... 165 268 152 245 GA Clayton. « +» + «5s + + 153 294 134 227 Fl. Hernando ........ 227 343 240 432 GA CHAN: . 32. + 5 roe wie 174 429 157 412 FL Highlands . . ...... 202 312 188 284 GA Cobb. ons nivmaims 190 297 134 227 Fl. Hillsborough . . .... 227 452 240 432 GA Coffee . ......... 236 363 217 322 FL Holmes. ......... 155 325 141 294 GA Colauitt, +i ¢viv sm 254 453 247 439 FL. Indian River. . ..... 237 364 218 326 GA Columbia ........ 152 254 149 242 Fl. JACKSON . . «vive 155 356 213 316 GA Cook «ooo 180 283 164 257 FL Jefferson. . ....... 183 286 166 259 GA Coweta. . ........ 172 275 134 227 FL Lafayette. . ....... 159 261 133 226 GA Crawford. . ....... 193 300 178 zr Fl. Lake os vivir wi 265 464 215 318 GA CHD 106 vim nas 197 307 184 280 FL LEO. ivy nm Hi 165 268 152 245 GA Dade . .......... 141 243 182 278 FL leon ......covvss 183 286 166 259 GA Dawson ......... 157 259 143 236 FL Levy. ........... 159 261 133 226 GA DeKalb ......... 153 255 134 227 FL Liberty .......... 183 286 166 259 GA Decatur... vo.ivin 228 345 203 306 Fl. MAgISON . wiv wn ow 183 419 166 403 GA Dodge .......... 206 384 190 378 FL Manatee . . ....... 266 465 239 431 GA Dol... 5 wuvie 5 aoe 197 307 184 280 FL Marion . . ........ 233 355 212 315 GA Dougherty . . ...... 144 246 135 228 FL Marti. von nnn wes 221 335 226 395 GA Douglas . ........ 190 297 134 227 FL Monroe. . ........ 200 310 186 282 GA ESV «inn wares ins 144 342 135 303 FL Nassau. . us ua sms » 158 260 139 232 GA Echols . ......... 180 283 164 257 FL Okaloosa ........ 163 295 177 270 GA Effingham . ....... 143 245 150 243 FL Okeechobee . . .... 221 376 226 356 GA Elbert. .......... 253 437 233 422 FL Orange. ......:.. 142 244 215 318 GA Emanuel. ........ 222 336 200 336 FL Osceola ......... 257 456 215 318 GA Evans. .......... 143 381 150 376 FL Palm Beach. ...... 221 376 226 355 GA Fannin .......... 173 276 156 249 Table |. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked unlinked unlinked linked linked GA Fayette... ....: en 153 294 134 227 197 346 205 308 GA Floyd s...u z sis bs 5 mas 154 256 140 233 | BGA Tabol ....svones 166 269 153 246 GA Forsyth. ......... 153 255 134 227 152 254 149 242 GA Franklin . «dws man 216 327 199 296 143 245 150 347 GA Fulton .......... 153 255 134 227 | GA TYlOr. . .: is sanas 193 300 178 271 GA yy CR 173 446 134 386 ; GA TeMalm....,...cun- 206 316 190 286 GA Glascock. . ....... 152 254 149 242 | GA Temell. ......wius 144 246 135 228 GA BlYNn. os wv nies 158 305 139 277 178 281 162 255 GA Gordon. . ........ 154 256 140 368 | BGA 2 THiasesswirmiws 189 292 174 267 GA Grady. «vi sw ry 178 281 162 255 206 436 190 421 GA Greens: . crave 164 422 163 256 157 333 143 297 GA Gwinnett. - .v.on uu 153 330 134 227 206 316 190 286 GA Habersham . ...... 157 259 143 330 [| GA TIOUP. us weemnas 172 275 1565 248 GA Hall, «5 wows mwas 157 259 143 236 189 292 174 267 GA Hancoek. cs vx + ov 201 311 187 283 193 300 178 271 GA Haralson. ........ 177 299 161 272 | BA 2 UNION. sscms eee 157 333 143 297 GA Hams. .covrmeme nw 166 269 153 246 269 468 257 449 GA Hart. ........... 216 327 199 296 141 243 182 278 GA Heard. «ows wus nas 172 275 155 248 153 330 134 227 GA Henry. .......... 153 294 134 227 | GA $2 WAG... anisms 174 277 157 250 GA Houston . «us cvs 193 300 178 271 152 254 149 242 GA win ........ Bar 4 13 189 292 174 267 250 415 249 441 GA JACKSON csv wh vs 164 267 163 256 230 349 208 311 GA JBSPBF : + vs maine on 220 334 134 227 166 269 153 246 GA Jeff Davis. . vx vv + 236 363 217 322 206 316 190 286 GA Jefferson. . ....... 250 414 149 384 | GA White......ueisus 157 259 143 236 GA Jenkins. . ........ 152 442 149 409 145 247 136 229 GA JOONSON = 55 & wit wa 206 316 190 286 197 307 184 280 GA Jones. .......... 193 300 178 271 253 438 233 423 GA Lamar. «evs mane n 269 468 134 227 193 300 178 271 GA Lanier. . ......... 180 344 164 35 | GA Worh........... 144 246 135 228 GA Laurens vc. suv 206 316 190 286 GA keds cv ouvamsnan 144 246 135 28 | A. AdaIr......0e0ws 546 1023 608 994 GA {EJ ERR 143 245 150 243 647 1054 631 1027 GA LINCOIY + + wv 515: 00 152 254 149 242 290 487 268 1131 GA LONG + 5m: arn we 5 10 ari 143 245 150 243 649 1056 633 1038 GA LOWNDES. « ws wisiv iw ve 180 283 164 257 596 1043 622 1014 GA Lumpkin... ... LL. 157 259 143 236 545 928 528 904 GA Magon vz s sev as 197 346 205 308 557 942 548 925 GA Madison . ........ 164 267 163 256 606 999 589 1044 GA MaHoN «owns 50 500s 166 269 153 246 557 942 548 925 GA McDuffie. . ....... 152 254 149 242 557 942 548 925 GA Molntosh. vv ois 158 305 139 277 560 1121 530 mM GA Meriwether . ...... 172 275 155 248 | A Buter........... 557 942 548 925 GA Miller . . ......... 228 345 203 364 633 1034 617 1005 GA Mitchell, , . : «x cvs « 178 281 162 255 633 1079 617 1070 GA Monroe. . ........ 193 398 178 271 | JA. Coss ...ciuiimimn 596 1043 622 1014 GA Montgomery . ..... 206 316 190 288 } JA ©. Cedaf ...ovewina 545 928 528 904 GA MOrgan. . « : svar « n 164 267 163 256 556 941 540 917 GA MUITBY . . + «ci 5 145 247 136 229 560 1122 530 1112 GA Muscogee. ....... 166 269 153 246 634 1035 618 1006 GA Newton. ......... 220 334 134 227 546 929 529 1036 GA OCONBE . «vv vin sms 164 267 163 258 | A. 2 Chay... eosiinl sie 625 1024 609 995 GA Oglethorpe . ...... 164 267 163 256 350 1009 327 982 GA Paulding. ........ 190 297 134 227 361 557 336 529 GA Peach ..cocnvwsn 193 300 178 271 681 1193 541 1107 GA Pickens ......... 190 293 134 2207 | A 2 Dallas. ....: imps 546 929 529 905 GA PIBIC® . «vic rmsn 174 277 167 250 | IA Davis........... 649 1056 633 1039 GA PKB: wv eam a vu vin 269 468 134 227 546 1088 529 1079 GA Polk. ........... 154 256 140 233 350 1009 327 982 GA Pulaski, ooo vmim an 193 339 178 300 340 998 588 971 GA Putnam. ......... 201 311 187 346 625 1024 609 995 GA QuUItMan .. « + » www 162 411 147 374 302 512 287 949 GA Rabun =: 500 wanes 268 467 254 446 626 1025 610 996 GA Randolph . ....... 144 410 135 397 557 1078 548 925 GA Richmond... «+ «w+ 152 254 149 242 | JA Foyd........... 634 1035 618 1006 GA Rockdale. . . ...... 153 255 134 227 556 941 540 917 GA SCBY + «v's wwivms in 166 269 153 246 644 1051 541 1021 GA Screven ......... 222 435 200 420 617 1013 601 986 GA Seminole. . ....... 228 345 203 306 557 942 548 925 GA Spalding. ........ 204 314 134 227 617 1013 601 986 GA Stephens ., ...... 216 386 199 379 606 999 589 972 GA Stewart. . ........ 166 269 153 246 556 941 540 917 31 Table |. Alphabetical list of State, county, and health service area numbers for four alternative solutions — Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked 1A Harding vc nannen 606 1173 653 1160 ID Carlbol ... + :xnsm5 696 1357 683 1347 1A Harrison . ........ 596 1090 541 918 ID Cassia .......... 730 1244 717 1227 IA Hemy. so vurwnse 340 998 588 971 ID Clark ou v5 www x ws 722 1235 708 1217 1A Howard. . .-. vocn ne 552 1126 535 1116 ID Clearwater. . . ..... 694 1206 681 1188 1A Humboldt « .. «ov bv = 585 971 570 947 ID OUSBE «+ suv www 722 1235 708 1217 1A OR. o ov gl 2 ls #0 560 945 530 906 ID Emore.......... 716 1302 702 1289 IA IOWR 5 evans sues 545 935 528 911 ID Franklin ; ces 5 2 wes 715 1228 701 1209 1A JECKSON © vi ans aa 302 512 287 949 ID Fremont......... 722 1235 708 1269 1A JABPBE ive ems ar 546 929 529 905 ID Gem ........... 716 1229 702 1210 IA Jefferson. . ....... 589 1059 635 1047 ID Gooding . ssw s nu 695 1207 682 1189 1A Johnson. ........ 545 935 528 911 ID Idaho. .......... 694 1278 681 1257 IA JONES. : ws wh sms ia 545 928 528 904 ID Jefferson. , :. +... 722 1235 708 1217 | IA Keokuk. ......... 589 975 573 952 ID Jerome. ......... 695 1207 682 1189 IA Kossuth cv vcovun 556 1118 540 1105 ID Kootenai. ........ 734 1248 721 1231 1A EBB cov on Hi hhh 367 568 343 537 ID Latah. . vc mnims ns 784 1371 768 1357 IA LI: v0 comm oe winin 545 928 528 904 ID Lem. 506 vaoawe oe 813 1414 800 1399 1A LOUISE ...0% dds 545 935 528 911 ID lewis. .......... 694 1206 681 1188 IA LAICAS, «5.05 ss 500 645 1052 629 1022 ID LINooN «vows 5 vn 695 1207 682 1189 1A LYON +n vm vm vin 544 927 527 903 ID Madisohn . .....: «54 722 1235 708 1269 IA VT 0 546 929 529 905 ID MInIGoKA. ou + « 20 730 1244 717 1227 IA Mahaska. ........ 589 975 573 952 ID Nez Perce. ....... 694 1206 681 1188 1A Maton . cis wun 546 1127 529 1115 ID Oneida. . iw voi ows 696 1392 683 1378 1A Marshall . ........ 679 1191 671 1178 ID Owyhee ......... 716 1365 702 1354 1A MIS: svn n uv ans 596 986 541 918 ID Payee. : «uv «nue ws 729 1243 716 1226 IA 1 665 1177 658 1165 ID POWBE ... i 445000: 696 1209 683 1191 1A MORONG. «vu sus 4a 560 945 530 906 ID Shoshone . . ...... 734 1248 721 1231 IA MONO, «ois wins 589 1099 573 1090 ID Teton: « oo arenas 722 1369 708 1355 1A Montgomery . ..... 647 1054 631 1028 ID TwinFalls . ....... 695 1207 682 1189 1A Muscatine . . ...... 641 1048 282 625 ID Valley. «ws so vwss 716 1229 702 1210 1A Obrien .......... 652 1064 637 1055 ID Washington. . . .... 729 1243 716 1226 1A 08CeoIa +. vv vn x 652 1065 637 1056 IA PBOB «oni oie vc 430 4 644 1051 541 1021 IL ACBINE vos mr om ss 353 548 329 521 IA PaloAlo.. ... ....= 625 1136 609 1124 IL Alexander . ....... 563 948 549 926 IA PlymouiY. . a5 ows me 560 1093 530 1085 IL Bond vows vuirws se 299 633 261 598 1A Pocahontas . . . .... 585 971 570 947 IL BOONG vinx va nim 3 291 488 270 463 1A POI: 5 +50 00 300 ims 546 929 529 905 IL Brown .......... 318 515 298 492 IA Pottawattamie. . . . . . 596 986 541 918 IL Bureal.. . vs vu: wns 336 533 317 510 IA Poweshiek. . ...... 686 1198 676 1183 IL Calhoun . ........ 299 496 261 453 1A Ringgold. . «+c 546 1089 529 1080 IL Caroll «ovr wnvws 361 557 336 529 IA Sac... vin unn 633 1034 617 1005 IL CABS. vine wml 318 515 298 492 IA Boot «cvs mans 641 1048 282 475 IL Champaign . ...... 279 477 262 454 1A Shelby. .....o0vn. 596 986 541 1035 IL Christian. . . «s+: 4 318 628 298 643 IA BIOUN 5 vw wos in 0 672 1184 664 1171 IL CBIR co: pve. won ceo 02 mi 311 508 308 501 IA SOY «oo vv tvs sins 606 999 589 972 IL Clay, . vows ms ewes 351 546 328 520 IA YAMA, wo vot uw nis 679 1191 671 1178 IL Clinton. ......... 325 663 261 453 IA Taylor: « . cxv enews 644 1051 541 1021 IL COS. 0ims s0sn 279 623 262 564 1A UMOAL vv sms smsims 546 1023 608 994 IL COOK. ws is wns al% 287 484 288 481 1A Van Buren. ....... 589 1059 635 1048 IL Crawiord. .. .» cv. 310 640 293 626 IA WEPBHO ov avis un 589 975 573 952 IL Cumberland. . ..... 279 623 262 564 IA Waren ss wee ws ms 546 929 529 905 IL DeKab.......... 373 577 324 517 IA Washington . . . . . .. 545 935 528 911 IL DeWitt i: cvvmaves 338 664 381 655 1A WEBYNG « viv vw svn 645 1052 629 1023 IL Douglas ......... 279 477 262 454 1A Webster . ........ 585 971 570 947 IL DuPage ..«uvswvs 287 484 288 481 IA Winnebago . . ..... 556 941 540 917 IL EOQat. «uvon vim os 40 333 668 375 650 1A Winneshiek . . ..... 395 1144 387 1132 IL Edwards. « vs wu 50 351 546 328 520 1A Woodbury. . ...... 560 945 530 906 IL Effingham ...c... vs +5 363 560 337 530 IA Worth: «2:05 sams 556 941 540 917 IL Fayette. . w+ cw «uw 363 560 337 530 1A Wright .. ccovmres 606 1138 589 1127 IL FOr iv sania sms 279 477 334 527 IL Franklin ......... 331 528 312 505 ID ABs: sncwa vmems 716 1229 702 1210 IL PURON cos nws mes 277 475 303 496 ID Adams .......... 729 1358 716 1348 IL Gallatin. . ........ 384 590 359 579 ID Bannock... «a «ov iw 696 1209 683 1191 IL CIN, + «sv vv wus 319 516 302 495 ID Bearlake........ 715 1228 701 1268 IL GIHNAY... - +05 05 2510 303 500 288 597 ID Benewah. ........ 798 1399 787 1386 IL Hamilton... ...... 383 589 358 575 ID Bingham. «vs ov vs 696 1209 683 1191 IL HanGock.. «ws sv www 367 568 343 537 ID Blaine. . ......... 808 1409 795 1394 IL Hardin .......... 384 590 359 578 ID Boise. vv vy swnns 716 1229 702 1210 IL Henderson ....... 340 998 588 971 ID BOAABY... oo sim ew 734 1281 721 1260 IL BIN. «ve ce rin om to A 307 504 282 475 ID Bonneville . . . «vows 722 1235 708 1217 IL OQUOIS © + vv va +e 343 539 321 514 ID Boundary ........ 734 1281 721 1260 IL JBOKSOM «vis inis 4s 331 585 312 565 ID BUHG: swap wp smn iiz 722 1235 708 1217 IL JASPBY 5 «iv sw wn 351 546 328 520 ID Camas vc con vmame 808 1409 795 1394 IL Jefferson. ........ 335 532 316 509 ID Canyon. ......... 716 1365 702 1354 IL Jersey .......... 299 496 261 453 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions — Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked IL Jo Daviess. . ...... 302 499 287 480 IN Crawiords o «uv ae 339 536 280 473 IL JOMNBON « o vive ovis 12 13 9 458 IN Daviess. . ...:.... 310 507 293 589 IL Kane . cz su gmavas 373 877 324 517 IN DeKalb.......... 304 638 279 472 IL Kankakee . ....... 343 539 321 514 IN DeArbomM;: +: v5 sx 4 321 559 266 459 IL Kendall. « vv ce vas 373 877 324 517 IN Decatur. ......... 313 639 274 618 IL BOOK : vo inn Zin ins 324 521 306 499 IN Delaware. . ....... 323 520 305 498 IL LaSalle. ......... 336 533 317 510 IN DUDOIS «5s v5. 505 394 676 380 654 IL LAKE in cnivmn e912 287 551 288 481 IN Elkhart. ......... 349 545 326 519 IL Lawrence ........ 310 507 293 486 IN Fayelfte. . .. cs. +s 385 592 360 584 IL LEB voip ws iwn mmuws 291 599 270 463 IN Floyd. « « cows vmsn 339 536 280 473 IL Livingston . . . ..... 336 641 334 615 IN Fountain. . ....:+. 333 530 314 507 IL LOGAN =. hvu ww ain 318 662 298 616 IN Franklin ......... 270 672 266 637 IL Macon .....u vue 305 502 289 482 IN Fullon vou cmeersz 312 591 295 580 IL MACOUPIA + vo 5 «0 se 334 531 315 508 IN GIOSON. + +0 sv 05 » 273 471 272 465 IL Madison . . 23 wuss 299 496 261 453 IN Grant........ 0. 389 621 363 613 IL Marion . ......... 358 554 332 524 IN GrBBNG. vv» swiss 341 537 320 513 IL Marshall , «so. 596 50 277 475 303 496 IN Hamilton... ...... 390 642 274 467 IL Mason .......... 277 475 303 496 IN Hancock, . + vv 00 4» 362 558 274 467 IL Massal. « «vo uvn 12 13 9 458 IN Harrison . ....0v.. 339 536 280 473 IL McDonough. . . . . .. 318 515 298 547 IN Hendricks + « «« vo = 275 473 274 467 IL McHenry. . ....... 287 551 288 481 IN HOMY. .. von vv nas 362 558 373 648 IL Mclean . cows 338 535 334 527 IN Howard. . ........ 364 562 338 531 IL Menard. . . ....... 318 515 298 492 IN Huntington ....... 332 529 313 506 IL MBICOr . 4 ue 20s mw 307 504 282 475 IN Jackson ......... 313 674 296 627 IL Momrce. ......... 325 522 261 453 IN 0 308 505 284 477 IL Montgomery . . .... 334 531 315 508 IN BY vous ins winds 332 616 313 611 IL MOrgar. «= ms ss ve 319 516 302 495 IN Jefferson. ... . «v0. + 321 518 304 497 IL Moultrie . ........ 305 502 289 482 IN JENNINGS.» 5 v5 5a 4» 313 510 296 488 IL OIE : owns viv www 291 488 270 463 IN JOPIIBON «cov mp iw & 275 473 274 467 IL PEOTA . v4 4 av vss 277 475 303 496 IN KNOX wie iwi vim ums 310 8507 293 486 IL PoItY . us conn sms 331 585 312 566 IN Kosciusko . . ...... 349 545 326 519 IL PIB. ili [elt seta co 305 502 289 482 IN LaPorte, : vw noises 378 582 352 556 IL PKB. sassvuwnoms 353 548 329 521 IN Lagrange ........ 304 501 279 549 iL POPB swiss 384 590 359 578 IN LARD ore 09 vi ms 308 505 284 477 IL Pulaski.......... 563 948 549 926 IN Lawrence ........ 329 526 311 504 IL PUDam. » cman vw 336 533 317 510 IN Madison . . . «vows» 390 643 372 647 IL Randolph . ....... 325 522 261 453 IN Marion «.. +. en 304 275 473 274 467 IL Richland. . .. .c sx» 351 546 328 520 IN Marshall . ........ 312 509 295 487 IL Rock Island . . . .. .. 307 504 282 475 IN Mardin. 3 cess nwa 310 507 293 589 IL Baling, vw vu vs sews 384 590 359 579 IN Miami. . ......... 389 622 363 614 IL Sangamon. . ....:. 318 515 298 492 IN MONIOB, + vs 5 vuman 341 537 320 513 IL Shuler; «va» + sows 318 515 298 492 IN Montgomery . ..... 402 681 385 659 IL Sool... cons name 319 516 302 495 IN Morgan. .. ue ce 275 473 274 467 IL Shelby. + + vs ws wns 305 604 289 601 IN INBWAONY.. ves 05 500 8 308 505 284 477 IL SLC .vvonnyns 325 522 261 453 IN Noble. .......... 304 501 279 472 IL Stark . .......... 307 504 282 475 IN OHIO «vv iv 30 simmers 321 559 266 459 IL Stephenson. . ..... 302 499 287 480 IN Orange. ......... 329 526 311 504 IL Tazewell . . ....... 277 475 303 496 IN Owen. ..cxsnsess 341 537 320 513 IL UID: co 5 so nmis wes 563 634 549 623 IN Parke... ons vnan 311 508 308 501 IL Vermilion. . .....«. 333 530 314 507 IN POMY vs c0 03 san 316 627 300 617 IL Wabash . vues vos 273 658 272 638 IN PG, oss mamras 310 507 380 654 IL Wanen. : «. ;vsses 324 521 306 499 IN Porler. ..; wuss 4 wo 308 505 284 477 IL Washington. . . .... 358 554 332 524 IN POSEY. + + vs ius wan 273 471 272 465 IL Wayne ...csmsnus 335 532 316 509 IN Pulaski. ......... 312 591 295 581 IL WHS. . ive 5505 5a 383 589 358 576 IN Panam. ween ses a 275 620 274 467 IL Whiteside . . ...... 361 602 336 599 IN Randolph . ....... 323 520 305 498 IL Wills vues nana 303 500 288 597 IN RIDIOY: os weno mss 321 559 304 602 IL Williamson. . . ..... 331 528 312 505 IN ROSA... + rns 505 5 0 275 644 274 467 IL Winnebago . ...... 291 488 270 463 IN SOCOM «sv vcs v vans 365 563 280 567 IL Woodford . ....... 338 535 303 496 IN i 275 645 274 467 IN Spencer . . ....... 273 471 272 465 IN AdBMS 1. oven vn x 332 615 313 610 IN St. Joseph. . ovo ue 312 509 295 487 IN Allen ........... 304 501 279 472 IN Starke .......... 378 582 352 557 IN Bartholomew . . . . .. 313 510 296 488 IN Steuben .. vu.» 304 618 279 472 IN Benton. ......... 300 497 285 478 IN Sullivan. . ........ 311 661 308 605 IN Blackford. ..... 5 328 520 305 588 IN Switzerland . . ..... 321 518 304 497 IN Boone ...c..000 401 680 274 467 IN Tippecanoe . . . .... 300 497 285 478 IN BIOWIY civ win i 3 on 4 341 537 320 513 IN Tipton .......... 364 562 338 531 IN Caroll ...on2 me sms 300 497 285 478 IN UNION. « vous smn es 385 502 360 585 IN CEES i vu tne 7 mt 364 632 338 531 IN Vanderburgh . . .... 273 471 272 465 IN Ol. wo smzmn sms 365 563 280 473 IN Vermithon . . we on 311 508 308 501 IN 7 EN 311 508 308 501 IN IGE: coe rei 0 we 311 508 308 501 IN OHAON «4 ov 5 5 00% 300 497 285 583 IN Wabash ......... 389 666 363 645 33 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked IN Warren. ......... 333 530 314 507 KS Nemaha ......... 554 1071 639 1062 IN WEEK. « v0» 508 3 2 0 273 471 272 465 KS Neosho. .. os visi 601 992 584 966 IN Washington . . . . . .. 339 625 280 473 KS Ness . coivme anes 682 1194 673 1180 IN Wayne : vo wus sues 385 592 360 585 KS NOON «osu 0 0 sais 637 1044 623 1015 IN Wal. oo cv vw 9 332 529 313 506 KS O8age +. .o rss 554 938 538 915 IN White... csmeusen 300 497 285 478 KS Osbome. +. «vw sn 622 1019 606 991 IN Whitley. . ........ 304 501 279 472 KS Otawa . . . ci von 569 954 555 932 KS Pawnee. , . +... « «us» 583 969 568 945 KS AIBN iv swiss ws 601 1164 644 1081 KS PHIDS . + 350 oe 9 5 637 1044 623 1015 KS Anderson ........ 624 1086 644 1082 KS Pottawatomie . . . . . . 554 993 538 967 KS Afchison . . .. ...:. 671 1183 663 1170 KS Pratt. cv wns vmmaw » 618 1015 603 988 KS Barber .......... 669 1181 661 1168 KS Rawlins. . ........ 636 1042 621 1013 KS Baron . u've ww 583 969 568 945 KS RONO sv vmrut mens 598 988 582 963 KS Bourbon. ........ 630 1030 614 1001 KS Republic. ........ 642 1049 627 1019 KS Brown .......... 554 1070 639 1061 KS RICE: «5 05 sors mia en 598 988 582 963 KS BOISE. » vis is ww an 576 961 561 937 KS RiIgY co vovnnses 554 993 538 967 KS Chase .......... 567 952 553 930 KS Rooks .......... 565 950 551 928 KS Chautauqua. ...... 451 732 442 1071 KS RUSH woova nv maen 583 969 568 945 KS Cherokee ........ 539 922 546 923 KS Russell. ......... 667 1179 660 1167 KS Cheyenne . . ...... 636 1042 621 1013 KS Sains. «vs visas 569 954 555 932 KS CIR cv savmrvmss 551 1040 534 1011 KS 0 NE 575 1058 562 1046 KS CIaY. ov: 0 on mere wie 662 1174 655 1162 KS SOAGWICK =. v « us vs 576 961 561 937 KS 0 og 642 1049 627 1019 KS Seward. « «vas vs un 587 973 572 950 KS Coffey .......... 567 952 553 930 KS Shawnee. . ....... 554 938 538 915 KS Comanche,» «vs + = 618 1015 603 988 KS Sheridan. . ....... 613 1008 597 981 KS Cowley. ......... 576 1124 652 1159 KS Sherman. ........ 659 1110 646 1097 KS Crawford, . . «sv 630 1030 614 1001 KS SM. coi ens visas 637 1142 623 1129 KS DEAN. + 27 vv 00 a 613 1008 597 981 KS Stafford. . ........ 598 1108 582 1096 KS DICKINSON « «ov 5 4 650 1104 634 1092 KS Santon, ...on0 vue 562 947 547 924 KS Doniphan Brown . oem vie 468 751 456 730 TN Lauderdale . ...... 248 394 230 385 TX Burleson. ... io. vs 464 746 453 726 TN Lawrence . .-. ..... 232 443 234 426 > Burnet .......... 425 704 403 674 TN Lewis’. ...os mrs + 208 318 192 288 TX Caldwell . ........ 510 809 403 831 TN Leon. . oo Jahn 210 320 194 363 TX CanoUn os ws niin 433 897 507 883 TN Loudon. : uw + & vies = 147 249 132 225 TX Callahan. ».. w+ ves 428 707 415 686 TN MACON oc «00.0% # 23 385 137 372 TX Cameron. . .....: 520 903 514 890 TN Madison . . » « ears on 188 291 172 265 TX Camp. .......... 465 890 454 874 TN MBHOH + co 5 nis 141 243 182 278 TX CASON. + 4.0 wo mvs 405 684 404 675 TN Marshall... nnn 208 318 192 365 TX CEBS vos nig 50 hi 485 776 441 754 TN MEY. vo naive + 208 318 192 288 > CaSO! 5.x oi to on wi 6 0 481 878 470 863 TN McMinn... LLL. 147 353 211 314 TX Chambers. . ...... 408 882 426 847 TN McNairy ;........ 245 377 227 358 TX Cherokee ........ 509 893 504 880 TN Meigs. . ......... 147 353 211 314 TX Childress. . . ...... 512 829 497 818 TN MONIOE. : s = six sv 5 147 353 211 314 TX ClaY 5 5005 250s win 420 699 402 673 TN Montgomery . ..... 234 361 168 261 TX Cochrah ... «s+ vse 406 885 392 867 TN MODIS. . vu «won 239 367 220 333 TX COR vv www ya mms 426 705 413 684 TN MOIGAN. « « «4 vw ms 226 341 132 225 > Coleman. .. «.. us: 468 751 456 730 TN OBION: is 4 wu ssvins wins 181 284 165 258 TX Collin. .......... 453 734 418 689 TN Overton ..: ssa 215 326 198 295 TX Collingsworth . . . . . . 524 907 518 894 TN POITY . ov vvmininnin 232 352 210 313 TX Colorado. . ....... 490 783 479 760 TN Picket . . ooo 0540 215 326 198 295 TX Comal ses coisinwse 517 900 395 667 TN POI. + vv 5 smn sm te 173 276 156 249 TX Comanche. . ...... 507 805 494 785 TN PUNBIY. 5 «+ sow 0s 215 326 198 295 TX CONONA0. v5 #0 300 0 426 705 413 684 TN RBA vor ov 0% 48 226 416 182 278 TX COOKE . iv sme 2m 495 790 418 868 TN ROBNG +: 4 un i 0 oes 226 341 132 400 TX ICOWYBH «soso aida inn 452 733 446 719 TN Robertson . . ...... 148 250 137 230 TX Colle: os nnsmas 512 829 497 818 TN Rutherford. . . ..... 211 321 137 230 TX Crane. . .... ve sas 444 725 434 707 TN SCO = vine msm 263 462 248 440 TX Crockett: vo. ce wesw 426 705 413 684 TN Sequatchie . ...... 141 243 182 278 TX Crosby. ......... 406 861 392 848 TN Sevier, . . wise smu 147 249 132 225 TX Culberson. «x... «x 415 694 398 669 TN Shelby... « +s 55 4% 146 248 399 670 TX Pallam «cc suo wsne 405 717 404 696 TN Smith. .......... 231 371 137 321 TX Dallas. . ......... 453 734 418 689 TN Stewart, . «env vans 234 361 168 261 TX DAWSON «4 + i065 ws 406 685 392 796 TN Sullivan... ....... 151 253 179 273 TX DeWitt. ......4.. 433 712 419 690 ™N Sumner .....s xxx 231 351 137 230 TX Deaf Smith: « «ove 2» 491 784 480 761 TN TOION, vi 5 5r0 sim 146 248 399 670 TX OBIE... vo «ings i 2 bw 438 718 425 697 TN Trousdale . . .....:. 231 351 137 321 TX Deon. wz sw nie =u 495 790 418 689 TN UNICO! © oss aman 169 272 179 273 TX Dickens ......:.: 406 861 392 848 TN Union. .......... 147 249 132 225 TX Dimmit. ......... 531 914 510 886 TN VanBuren........ 211 357 214 317 TX Donley: «vinens 405 684 404 675 TN Warren. ......... 211 357 214 317 TX Duval. .... 9000 437 716 424 695 TN Washington. . . .... 169 272 179 273 TX Eastland... wow ov 4 428 707 415 803 TN Wayne ......: «us 232 352 210 313 TX ECO wv x + 0k a wwe 444 725 434 707 TN Weakley . . ....... 181 412 165 398 TX Edwards. ........ 506 804 493 782 TN WHHB .. 5. «ii 5 i 5 v5 8 215 451 245 437 TX ElPasg0: sc + sms vw 415 694 398 669 TN Williamson. . . ..... 148 250 137 230 TX ENS... oon norvma on 453 734 418 689 TN WIIBON 5 40s vs pw es 231 371 137 230 TX Eratly. cu 05 cuinme 507 805 494 786 TX Falls, . .v ni vnvina 462 744 446 719 ™ Anderson . ....... 509 808 496 794 TX Fannin. ......... 436 715 423 694 TX ANCOIEWS . . ..c + cnvn 508 806 495 790 TX Fayelle... vo ow ams 490 783 479 760 TX Angeling . cx a ia 449 730 439 712 TX Fishel. . ....o os za 503 888 491 871 TX ATARSEE + vs 6 30 view 437 716 424 695 TX EIOY.: «rv min www 481 767 470 747 TX AICher ows evens 420 699 402 673 TX Foard, . «ova ms ai 458 739 445 718 TX Armstrong. . . ..... 405 684 404 675 TX FortBend . ....... 408 687 405 676 TX ALBSCOSR. &. . «+ vv ss 410 689 395 667 TX Franklin z sie wwii ws 465 748 454 728 1B PUSHIN cv cosmuis 408 687 405 676 TX Freestone . . ...... 489 781 477 758 TX Bailey. . ......... 406 685 392 664 TX BRO 5 vw Bw oie 410 832 395 822 > Bandera .. i «us ns 410 689 395 667 TX GAINES. co 5 ei 3 1 2 508 806 495 805 TX Bastrop. . . «ov bn 425 704 403 674 TX Galveston . . ...... 532 915 405 836 45 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions — Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked TX Garza. .......... 406 685 392 664 TX Matagorda. . . . .... 505 803 405 779 TX Gillespie . . ....... 506 804 493 781 TX Maverick. . . ...... 516 898 508 884 TX Glasscock . . . ..... 518 901 511 887 TX McCulloch. . . ..... 426 705 413 684 TX Goliad . .vunwvs en 433 712 419 690 > Mcltennan. : «oq +5. 462 744 451 724 TX Gonzales. . . ...... 410 779 395 875 TX McMullen . ....... 410 689 395 667 TX BIAY ow vawvn sens 460 742 449 722 TX Medina. . vue vrs 410 689 395 667 TX Grayson. . .. vw 5 wn 4 436 715 423 694 TX Menard... . ss 565 426 705 413 684 TX Gregg «ou 454 735 441 714 TX Midland . ........ 492 786 481 763 TX OUMBS x sss win noe 464 871 453 856 TX Milam, ... cocina cos 452 733 446 719 TX Guadalupe. . . ..... 410 779 395 667 TX Mills. . .......... 468 751 456 730 TX Hale. . «vi v5 vu ou 481 767 470 747 TX Mitchell. . ...cocune 503 800 491 776 TX Hao vs a0 50s 20 2 3 512 830 497 819 TX Montague . . . ..... 420 699 402 673 TX HAMUEON . « «v5 x vov 462 772 451 751 TX Montgomery . ..... 408 687 405 676 TX Hansford. .......: 405 841 404 675 TX Moore ...:e e504 405 717 404 696 TX Hardeman. ....... 458 739 445 718 TX MOIS. «cv im ean oe 465 748 454 793 TX Harel 0 cobs gins 413 692 426 698 TX Molley ..: ws va swu 406 685 392 664 TX Harris. . ......... 408 687 405 676 TX Nacogdoches. . . . . . 447 728 437 710 TX Harmison . . vous ans 485 840 441 714 TX Navamo . vu 0a 489 781 477 758 TX Hartley . . ........ 405 717 404 696 TX Newton. ......... 413 692 426 698 TX Haskell. . . «vc vu wy 428 769 415 749 TX NOIR: 500 woos os 503 800 491 776 TX HAYS: + + a wine 0S 510 809 403 674 TX NUSCBES.. . « 4 5 5% os 437 716 424 695 > Hemphill... ...... 460 742 449 722 > Ochiltree. . . . ..... 405 842 404 800 TX Henderson ....... 441 819 431 703 TX Oldham . cows vows 405 684 404 675 TX Hidalgo. . . ....... 427 706 414 685 TX Orange. ......... 413 692 426 698 TX Hill woos ihisms sme 462 850 451 778 TX Palo PIO: « vss » 479 764 467 744 TX Hockley ......... 406 685 392 664 TX Panola. ........- 454 892 523 899 TX Hood: voz ma nus ws 434 713 432 705 TX Parker « «ww + wy v0 434 713 432 705 TX Hopkins . ........ 453 834 418 837 TX PaIVIBE .......00 oi Wit 52 491 784 480 761 TX Houston . ...v ove 509 808 496 795 TX P8BOS: +5 4 + wysrww ws 492 891 500 876 TX Howard. . .. cs 009 518 901 511 887 TX Pol, vs 5 ow ww vo 449 730 439 712 TX Hudspeth . ....... 415 694 398 669 TX Potter. . ......... 405 684 404 675 TX Hont ... 10a sms 453 734 418 689 TX Presidio . vs vu vw 526 909 520 896 TX Hutchinson . . ..... 405 684 404 675 TX Rains. .......... 453 734 418 689 TX WOR. . «s swan aw ews 426 705 413 684 TX Randall. . .. vv uu: 405 684 404 675 TX JBBRL «wx remind ba 479 764 467 744 TX Reagan. . ..« «sm» 426 705 413 684 TX JEOKSON . ; 5 www wn 433 712 419 797 TX Real, ; «suns » wane 506 804 493 782 TX JASPBY + ¢ ns dE 413 692 426 698 TX Red River . . «us vs 438 718 425 697 TX JH DAVIS: vsok ve wi 526 909 520 896 TX Reeves. ......... 444 725 434 707 TX Jefferson. ........ 413 692 426 698 TX Refugio. «ws was awe» 433 852 419 839 TX JimHogg ........ 538 921 526 902 TX Roberts. . . ....... 460 742 449 722 TX JmWells ........ 437 716 424 695 TX Robertson. . «cu. « 464 746 453 726 TX Johnson. ........ 434 713 432 705 TX Rockwall. . ....... 453 734 418 689 TX JONBS., «bs sa wmamy 428 769 415 749 TX Rumels . .... vu 426 851 413 838 TX Karnes . ......... 513 870 498 855 TX Busk .. oc amevs 454 874 431 852 TX Kaufman... ...... 453 734 418 689 TX Sabine . . ........ 447 826 437 814 TX Kendall. . ..:0v00. 410 689 395 667 TX San Augustine . . . . . 447 728 437 710 TX Kenedy. ......... 437 716 424 695 TX San Jacinto . . ..... 408 687 405 676 TX Kon. . co sn sms nme 428 873 415 858 TX San Patricio. . . .... 437 716 424 695 TX Kerr. ......c0.n. 506 804 493 782 TX SanSaba ........ 468 827 456 815 TX Kimble... cvs voawsas 506 857 493 844 TX Schigicher. . . ... .. . 426 705 413 684 TX KANG. + vv vm wns 406 685 392 664 TX BCU iis withns ins 20% 519 902 512 888 TX Kinney .. «coos as 516 899 509 885 TX Shackelford . . . .... 428 707 415 686 TX Kleberg. . ........ 437 716 424 695 TX Shelly . ov 00344 447 728 437 710 TX ROOK us on vs nvm a 420 765 468 745 TX Sherman. . Jc... 405 717 404 696 TX laSalle. . ....i5a 410 832 395 822 TX Smith. v.45: 05 5509 441 721 431 703 TX Lamar. . ......... 438 718 425 697 TX Somervell . ....... 434 713 432 705 TX Lams... casepime 406 833 392 823 TX Sarr «os ws smamn 427 706 414 685 TX Lampasas. ....... 452 859 446 835 TX Stephens . ....... 428 707 415 686 TX LaVaCa . » 5 wim ns wa 433 883 419 865 TX Sterling: » «vs viva 426 896 506 882 TX LBB + mis tie 3 i 0 425 884 403 870 TX Stonewall . ....... 428 873 415 858 TX LEON. = 4. + to wom arosinian 501 797 489 772 TX Sutton . ......... 426 705 413 684 TX Liberty ...consam+ 408 687 405 676 TX Swisher .: : suv ass 405 684 404 675 TX Limestone. . ...... 462 744 451 724 TX Tarrant. ......... 434 713 432 705 TX LIPSCOMB os v5 wm» 469 752 457 731 TX TaAYIOr: + sv bai v0 428 707 415 686 TX Live Oak. ........ 437 716 424 695 TX Towel. ....onsws. 492 891 500 876 TX Llano. .......... 425 704 403 674 TX Terry ........... 514 875 499 860 TX LOVING =. «osama vs 415 694 398 669 TX Throckmorton. . . . .. 420 785 402 762 TX Lubbock . . ....... 406 685 392 664 TX THUS «nvr wns 465 748 454 728 TX RYAN oc onshim nso 406 685 392 664 TX Tom Green ....... 426 705 413 684 TX Madison . . ....... 501 797 489 772 TX THaVIS: vo sn 20s 5 425 704 403 674 TX Marion; od voi ns 485 776 441 754 TX THARYS io: ¢ 50s 0 000 000 449 730 439 712 TX Martin. . wooviis sms 492 786 481 763 TX Ter ois imum 413 863 426 789 TX MESON «x s5'v vn wins 506 804 493 781 TX Upshur. ......... 454 735 441 714 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked TX LIDDY, o.oo 3 vos tis 492 786 481 763 VA Camo . «o's win em 73 81 53 54 TX Uvalde o.oo ve vins 536 919 522 898 VA Charles City. . . .... 33 36 23 23 TX Val Verde ........ 516 899 509 885 VA Charlotte. . . ...... 33 205 97 190 TX VanZandt........ 441 721 418 689 VA Chesapeake City. . . . 6 7 45 45 TX VICSOL, os wn pins a 433 712 419 690 VA Chesterfield . . . .... 33 36 23 23 TX Walker . . ........ 449 825 405 676 VA Clarke « moms oni vw 25 28 19 19 TX Waller. «5 v5 sess 408 687 405 676 VA Craig... «oii nw ves 14 16 11 11 TX WAH. oon od 0 444 725 434 707 VA CUIPBPBT. vs 2 vx vss 99 121 76 99 TX Washington . . . . . .. 464 746 453 808 VA Cumberland. . . .... 33 205 97 190 TX Wobbly. ois «os ms mn 538 921 526 902 VA Dickenson. . ...... B5 61 40 40 TX Wharton . . ....... 505 803 405 780 VA Dinwiddie . ....... 77 87 23 23 TX Wheeler . ........ 460 742 449 809 VA ES8EY. . .msma 2 tw 33 36 23 23 TX Wichita. . ........ 420 699 402 673 VA Baldax ss us vs aman 69 76 16 16 TX Wilbarger ........ 458 739 445 718 VA Fauquier. . .%cvx ss 109 136 16 174 TX Willaey . .. ives oan 520 903 514 890 VA FIOYd = so6 15m aim a 14 16 11 11 TX Williamson. . . ..... 425 704 403 674 VA Fluvanna. . ....:.. 99 121 76 98 TX WISON . «ovis im name 410 689 395 667 VA Franklin . ........ 14 16 11 11 TX Winkler. . ........ 444 872 434 857 VA Frederick. « . io cms 25 28 19 19 TX WISE cos ialrios smn 495 881 432 828 VA GUBS .vvivn nmiawn 70 77 49 50 TX WOO. «+ cosms se 441 721 431 703 VA Gloucester. . . ..... 5 6 45 45 TX YORKUM «co «mvs 514 876 499 861 VA Goochland. . . ..... 33 36 23 23 TX YOUNG . - vx om sw a 3 420 785 402 762 VA Grays ou sw « »w #0 73 81 53 54 > BPR» sae ego 538 921 526 902 VA Greene. ......... 99 121 76 98 TX ZoVala .. .v atie ww BE 410 689 395 667 VA Greensville ....... 130 232 120 213 VA Hala% ... ov 500m 71 79 B1 52 uT Beaver. cui cman 755 1272 739 1252 VA Hampton City. . . . . . 5 6 45 45 ut Box Elder ........ 744 1317 694 1270 VA Hanover . ........ 33 36 23 23 ut Cache co vovivn vas 715 1228 701 1209 VA HeAMEO:. . ws 45 ew vs 33 36 23 23 uT Carbon... .. eu 05s 809 1410 797 1396 VA HEY: 5s wis so sms + 79 90 85 58 uT Daggett .wivuvmns 799 1400 788 1387 VA Highland. . ....... 97 117 73 22 ut DaviS.... . # wn mss 744 1260 694 1202 VA Isle Of Wight . . .. .. 6 135 45 45 ut Duchesne. . ...... 708 1259 694 1240 VA James City ....... 5 214 45 45 ut BMBLY: & adn 9s im ww 809 1410 797 1396 VA King And Queen. . . . 33 36 23 23 ut Garfield. . ........ 707 1290 693 1267 VA King George . ..... 135 237 16 16 ut Grand. « «cm. 5 0 50s 711 1297 697 1283 VA King William. . . .... 33 36 23 23 ut WON s vw vi amp ams 755 1272 739 1252 VA Lancaster ........ 33 158 23 147 uT JUD swe swine ems 703 1216 689 1197 VA LBS + smn wus wa 118 220 112 205 ut KENG v0 = bs prtin & 707 1290 693 1267 VA LOUCOUNY ,. + vo 8 4 5 69 211 16 16 uT Millard . ......... 703 1370 689 1356 VA LOWBA wii 000 99 121 76 98 uy Morgan, . «oss vv 744 1260 694 1202 VA Lunenburg. . ...... 71 79 51 52 ut Pils ovo wo Sn ee 707 1290 693 1267 VA Madison . . ....... 99 121 76 98 uT RICH. ace wiv joe in 8 744 1260 694 1202 VA Mathews. . ....... 5 6 45 45 ut Saltlake. .....+.. 708 1221 694 1202 VA Mecklenburg . . . . .. 71 79 51 52 uT SaVJUBIY fas swe 740 1255 726 1236 VA Middlesex . . . ..... 33 158 23 147 ut Sanpete . . ....... 703 1216 689 1197 VA Montgomery . ..... 70 77 49 50 uT Sever wi su sme sw 703 1216 689 1197 VA Nansemond. . . .... 6 135 45 45 uT Summit. . «s+ sens 708 1221 694 1202 VA Nelson... coe ssn 99 121 76 98 ut Too + ss wv waa 4 708 1221 694 1202 VA New Kent . ....... 33 36 23 23 ut Uintaly . .vooie ns us 708 1259 694 1240 VA Newport News . . . . . 5 6 45 45 ut BEAR nvm os mee 703 1216 689 1197 VA Norfolk/Portsmouth . . 6 7 45 45 (2) Wasalch o.oo sms 708 1325 694 1340 VA Northampton . . . . . . 137 239 129 222 ut Washington . . . .... 707 1220 693 1201 VA Northumberland . . . . 33 158 23 147 ut Wayne .. cov vipa 703 1216 689 1197 VA Nottoway. . . ...... 33 36 23 23 uT Weber . ......... 744 1260 694 1202 VA Ong... . oo vos 5 0 99 121 76 98 VA Page «» vv nn a cnn 63 70 46 47 VA Accomack . . . «vee. 187 239 129 222 VA Patrick . . ........ 79 90 55 58 VA Albemarle . . ...... 99 121 76 98 VA Pittsylvania . ...... 132 234 126 219 VA Alexandria City . . . . . 69 76 16 16 VA Powhatan ........ 33 36 23 23 VA Alleghany . ....... 30 33 24 24 VA Prince Edward . . . . . 33 205 97 190 VA Aelia . vn vite 33 36 23 23 VA Prince George . . . . . 77 87 23 23 VA AMNErst o.ov wuss 14 82 11 55 VA Prince William. . . . . . 109 136 16 16 VA Appomattox. . . .... 14 82 1 55 VA Pulaski. ......... 70 130 82 122 VA Arington. . ....... 69 76 16 16 VA Rappahannock . . . . . 99 121 76 99 VA Augusta . ......0: 97 117 73 9 VA Richmond . . ...... 33 36 23 23 VA Bah... sow bw vate 30 33 24 24 VA Roanoke. ........ 14 16 11 11 VA Bedford ... we + wens 14 82 Ia! 55 VA Rockbridge . . ..... 97 117 73 92 VA Bland. .......... 39 42 31 31 VA Rockingham. . . . . .. 63 70 46 47 VA BOASIoOurt ; © ..v ws we 14 16 1" 11 VA Russell, . : ous sus 151 174 179 124 VA Brunswick . . ...... 130 232 120 213 VA Scott ........... 151 253 179 273 VA Buchanan. ....... 39 173 31 162 VA Shenandoah . . .... 25 28 19 133 VA Buckingham. . . . ... 99 121 76 98 VA SY «ova ess 151 204 179 176 VA Campbell ........ 14 82 11 55 VA Southampton . . . . . . 6 135 45 45 VA CAINE «vv vv: + 135 237 16 16 VA Spotsylvania . . . ... 135 237 16 16 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked VA Slalldrd.. . «uv sv 0s 135 237 16 16 WI Burnett, . «vv «wan 314 511 297 489 VA BUY « 255 2.5 ws 2 2 77 87 23 23 WI Calumet . cov vans 355 550 Nn 490 VA Sussex. ......... 77 87 23 23 WI Chippewa . . . ..... 298 495 273 466 VA Tazewell . ........ 39 42 31 31 Wi CIA pov vw nv 8 ww 282 480 263 455 VA Virginia Beach . . . .. 6 7 45 45 WI Columbia ........ 301 498 286 544 VA Waren. «svi ries; 134 236 127 220 WI Crawford. «sv: «5.0 350 614 327 609 VA Washington. . . . . .. 151 253 179 273 WI Dane .. woos imi 301 498 286 479 VA Westmoreland . . . .. 33 36 16 16 WI DOAO. ws vis = 5 vw ew 306 503 354 561 VA WISE . 0 onusmen 85 61 40 40 WI DOOE 2 svi macnn 278 609 260 607 VA Wythe .......:v03 70 130 82 123 Wi Douglas . ........ 289 486 267 460 VA Yorks vauss som +4 5 6 45 45 WI DUNN «55 mews ms» 298 495 273 466 Wu Eau Claire. . . ..... 298 495 273 466 vT AQOISON . . .. csv 49 53 34 34 Wi Florence. . . vu «x + « 315 513 299 493 vT Bennington . . ..... 10 141 88 159 WI Fond Dulac ...... 306 503 291 484 vT Caledonia . . ...... 103 126 78 105 wi FOIBSE + vs wis vn v ws 284 482 265 457 VT Chittenden. . . ..... 49 53 34 34 wi Grant. : viv swiss 301 567 342 536 vT Essex........... 138 240 130 223 WI Green. .......... 326 523 309 502 VT Frankliny «os 0 0 52 49 53 34 34 WI Greenlake ....... 306 503 291 484 VT Grandlisle. ....... 49 53 34 34 Wi OWE «vu minor 80s 301 498 286 479 VT Lamole .. ows vin ws 103 126 78 106 wi HON .. io 2 5 ws 37% wows 284 659 366 641 VT OraREE: = « 7x 55» vu 15 17 14 14 Wi Jackson ......... 399 678 389 661 VT Orleans. +. «..+ 5% 103 192 78 181 WI Jefferson. . «vv .u vn 387 596 362 592 VT Rutland. . .. vv von 122 224 88 160 WI Juneau. ......... 391 653 364 636 VT Washington . . . . . .. 15 190 14 178 Wi Kenosha . «vo vv ru 382 587 356 570 VT windham . . «eux. 90 105 67 78 Wi Kewaunee. . ...... 278 476 260 452 VT Windsor . ........ 15 17 14 14 Wi LaCrosse ........ 290 487 268 461 WI Lafayette. .: . eo. v 0 326 523 309 502 WA ACES 5 sn anne sa 747 1382 761 1322 wi Langlade. . . ...... 379 583 353 558 WA Asotin. . ......... 694 1206 681 1188 Wi LINCOM + + smn vv mine 282 588 357 573 WA Boron ; «vos vive v3 702 1215 761 1321 WI Manitowoc. . . ..... 355 550 384 658 WA CHBIBN .. . 4 6 wp 4 0 747 1263 732 1243 WI Marathon «wow vn 282 588 357 572 WA Clallam. , vs oss » 785 1374 769 1360 Wi Marinette. . . ...... 315 513 299 493 WA Clark v cvia cmv 4 689 1333 695 1203 Wi Marquette . . ...... 301 498 286 544 WA Columbia ........ 717 1230 703 1211 WI Menominee . . . .... 379 583 353 559 WA COWIE, + vi vr 3o'miw 2 689 1253 695 1300 WI Milwaukee. . . ..... 280 478 354 560 WA Douglas . ....%u «.. 747 1263 732 1243 \ Monroe. . . ....... 290 487 268 461 WA Fomy « eume win ss 698 1274 685 1254 wi OconlD. ; suas « 29 278 476 260 452 WA FERRI 4 5 20 5.000 0 702 1215 761 1321 WI Oneida. ......... 284 482 265 457 WA Sadie. vo x nn 0% 694 1206 681 1188 WI Outagamie. . . ..... 344 540 291 490 WA Gray ov vw 0 wn 747 1263 732 1243 WI Ozaukee. ........ 280 478 354 560 WA Grays Harbor. . . . .. 758 1352 741 1342 Wi Pepin. .......... 298 637 273 586 WA Island. « cus vasa 736 1250 714 1223 WI Pierce .......... 370 573 350 552 WA Jefferson. . . ...... 785 1375 769 1361 WI Polk. . .......... 286 561 297 600 WA King. «sc va sasmas 736 1250 714 1223 wi Portage. ...: sa smns 400 679 390 662 WA Kitsap. . ......... 762 1282 745 1261 wi Price ........... 357 608 331 606 WA VIHIES: oovns wis mr 739 1254 725 1235 WI Racine . ......... 382 587 356 571 WA Kiokitat, wv v0 oon s 748 1264 733 1244 WI Richland. . ....... 301 567 342 536 WA Lewis. .......... 758 1276 741 1255 WI Rock . .......... 326 665 309 644 WA Lineal, suis wo bm 698 1211 685 1193 WI BOslc .. sosmenmsn 374 578 348 546 WA MBEOR «wt mams » 762 1282 745 1261 wi Sauk ........... 301 498 286 544 WA Okanogan. ....... 747 1324 732 1313 WI SAWYBY, vivian ton 357 553 331 523 WA Pachic: vio o0 vc wn inns 758 1319 741 1308 WI Shawano. . ....... 379 583 353 559 WA Pend Oreille. . . . . .. 698 1211 685 1193 WI Sheboygan . ...... 355 671 383 657 WA PIBICE 55 +anam is ) 794 1395 783 1382 Wi SL. Croix « + von al a 500s 370 573 275 468 WA SanJuan ........ 736 1250 714 1223 WI Taylor. . ......... 282 480 263 455 WA Skagit. . ......... 736 1250 714 1223 wi Trempealeau . . . . . . 290 629 268 619 WA Skamania . ....... 748 1264 733 1244 WI Vernon. ......... 290 487 268 461 WA Snohomish . . + yu 736 1250 714 1223 Wi Vilas ...... 284 482 265 457 WA Spokane. ........ 698 1211 685 1193 WI Walworth. . . ...... 387 597 362 593 WA Stevens ......... 698 1274 685 1254 WI Washburn. . ...... 314 511 297 489 WA Thurston . « ov «vv a + 758 1276 741 1255 Wi Washington . . . . . .. 280 478 354 560 WA Wahkiakum . . . .... 689 1253 695 1300 Wi Waukesha. . ...... 280 478 354 560 WA WallaWalla . . ..... 717 1230 703 1211 wi Waupaca. . ....... 344 540 291 490 WA Whatoom .. .... + 815 1416 802 1401 WI Waushara . . . «+. 306 565 291 574 WA Whitman. : . « «co 784 1372 768 1358 WI Winnebago . ...... 306 565 291 490 WA 1 739 1254 725 1235 WI Wood: . w vre ws ws sw 282 480 263 455 Wi Adams .......... 391 652 364 635 WV Barbour ......... 50 54 35 35 WI Ashland .... . «s+ 357 553 331 523 wv Berkeley... ...... 25 106 68 79 wi Bamon :..: ns uns 374 578 348 545 nv BOONE . . cows cvs 7 8 50 81 WI Bayfield ......... 357 553 331 523 wv Braxton, : uv ume 92 109 70 119 Wi Brown .: es ve5uv 278 476 260 452 WV Brooke. . icv 05 24 26 57 83 WI Butalo . co 5 ven 298 495 273 466 wv Cabell .......... 46 50 3 3 Table I. Alphabetical list of State, county, and health service area numbers for four alternative solutions —Con. Service area number Service area number 800 1400 800 1400 800 1400 800 1400 State County unlinked unlinked linked linked State County unlinked unlinked linked linked WV Calhoun. ........ 34 118 43 43 wv Roane .......... 34 118 43 131 wv Clay. 5c « 5.5 len ww 7 8 50 51 wv Summers . . vues 60 194 42 183 wv Doddridge. . . ..... 92 109 70 84 wv TaYIol. + 2.5 v5 sane + 31 197 94 187 wv Fayelle. . . + wan 60 66 42 42 wv TUCKEY «+ wv i amiss 50 54 35 35 wv GUBBE ...o0ev wi Fi dlin % 92 109 43 43 wv TYIBE 24 46 ite mem 24 46 32 32 wv Grant. soe m0 0 i 1 146 48 143 wv UPShUP. oo wows wy wa 92 196 93 186 wv Greenbrier. . . ..... 30 154 24 146 wv Wayne... : casino 46 50 3 3 wv Hampshire. .. ..... 25 28 19 19 wv Webster . ........ 7 134 50 142 wv Hancock. . . +... 380 584 57 83 wv Wetzel .......... 24 46 32 32 wv Hay. oo vwemes 4 146 48 143 wv WIR cae go dns wos 34 37 43 43 wv Hamson ... oo «wi ws 92 109 70 84 wv WOOD. « « vuivny wus 34 37 43 43 wv JEOKEOMN + «vo vei 5 34 156 43 154 wv Wyoming. . ....... 60 66 42 42 wv Jefferson. . ....... 25 106 68 80 Y Kanawha. . ....... 7 8 50 51 wy ADBY oi 1 ws miamms 771 1295 752 1281 wv Lowi: wu cums wiv ws 92 109 70 85 WY BIG HOI. oo vi ov vi + 749 1266 734 1246 wv neon. .... oe mens 46 50 3 3 wy Campbell ... vss 586 1348 571 1335 WV LOGAN ovine 4 oni » 13 165 10 153 WY Carbon. ......... 771 1296 752 1282 WV Maron . cco v svn s 31 34 25 in WY Converse . «+. s «us 726 1239 712 1221 WV Marshall . ........ 24 142 57 60 WY Crook. . ....v0v sus 586 972 571 948 WV MESON svn nus 283 481 264 456 WY Fremont . ...cc0.4 777 1351 759 1341 wv McDowell . ....... 39 42 31 31 WY Goshen ..+.:0s0. 580 1300 566 1287 wv Mercer .... «us was 39 42 31 31 WY Hot Springs. ...... 777 1322 759 1311 wv Mineral. . . ....... 1 1 48 49 WY JORNSON . 5 25 44 + 770 1291 751 1274 wv Mingo: icin ams 13 14 10 9 WY Laramie ......... 741 1256 727 1237 wv Monongalia . . . . . .. 31 34 25 25 WY Lincoln... ws sais aes 775 1314 757 1304 wv Monroe. . . ....... 30 154 24 146 wy Natrona ......... 726 1239 712 1221 wv Morgan... «c.a suis 25 188 19 175 WY Niobrara . « . +s sss 580 1301 566 1288 wv Nicholas . . ....... 7 134 50 51 WY PAIK: «wisi sm imu le 749 1266 734 1246 wv OND soso swsww 24 26 57 60 WY PIaBl: «cx nn vim n 797 1398 786 1385 wv Pendleton. ....... 63 70 46 47 WY Sheridan: . . «vv. :5 770 1291 751 1275 wv Pleasants ........ 34 37 43 43 WY Sublette . . ....... 775 1315 757 1305 wv Pocahontas. . .-. . .. 50 164 35 151 WY Sweetwater . . ..... 799 1400 788 1387 WV Preston. . . ...o + «+ 31 34 25 25 wy TBO x. 4. 0: fr 3% a ot 0 778 1315 757 1305 wv Putnam. ..: con anne 7 8 50 51 wy LE 792 1393 792 1391 wv Raleigh. . «cosas us 60 66 42 42 WY Washakie ........ 777 1323 759 1312 1% Randolph . ....... 50 54 35 35 wY Weston. . ........ 804 1405 791 1390 Wv BiHohIS «wns mamas 34 37 43 43 49 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 150 AL Jefferson. ....... 252 5.2 27 421 AR Sebastian ....... 700 73 3.9 150 AL Shelby ......... 252 8.2 4.6 421 OK LeFlore ........ 700 73 9.5 150 AL SL.Clalr ....:4 + 252 52 15.1 421 AR Crawford. ....... 700 7.3 3.4 150 AL Chiton......... 252 52 19.8 421 AR logan.......... 700 7.3 9.4 150 Al. Bib: uissmnmesss 395 52 29.0 421 AR Frankiiv ...:.005 700 7.3 6.2 150 AL Blount ......... 396 8.2 6.7 421 OK Sequoyah....... 845 7.3 10.8 150 AL. Cullman ..:: 4045 397 5.2 11.6 421 OK Haskell. ..:: 54495 866 73 20.4 150 AL Walker . ........ 441 52 4.2 421 AR Scott .......... 887 7.3 59 156 AL Tuscaloosa . ..... 258 11.4 6.7 432 AR Pulaski. .s: as 06% 71 8.2 41 156 AL Hale........... 258 11.4 20.8 432 AR Saline. ......... 713 8.2 44 156 AL Greene. ........ 258 11.4 19.5 432 AR Faulkner. ....... 711 8.2 4.0 156 Al. Pickens. . .u. us + 4 391 11.4 21.4 432 AR Lonoke. ........ 711 8.2 13.7 161 AL Mobile . ........ 264 7.0 41 bi AB Van Buren. , «. x. Hy 32 202 432 AR Grant.......... 71 8.2 45.3 161 AL Clarke ......... 264 7.0 10.7 i 161 AL Washi 432 AR Praltie v. uses 711 8.2 60.9 ashington . . .... 264 7.0 15.4 ; 432 AR Perry ....zriwsus 711 8.2 20.3 161 AL Baldwin. ........ 409 7.0 12.4 432 AR C 849 8.2 6 161 AL Monroe. . . ...... 449 7.0 11.3 ROINBY. oc ie te g 5 162 AL Houston ........ 265 9.9 5.7 post pi os Ha EA Te a 28 162 AL Dale........... 265 9.9 5.2 Bh: 130m mime : 162 AL Coffee ......... 265 9.9 15.0 446 AR Washington . . . . .. 727 11.2 6.8 162 AL Geneva. ........ 265 9.9 10.6 446 AR Madison ........ 727 11.2 8.3 162 AL Hey. .:...p:xx 265 9.9 4.6 446 AR Benton......... 802 11.2 8.9 162 AL Barbour ........ 411 9.9 17.4 446 OK Delaware. . ...... 802 11.2 26.9 162 GA Quitman ........ 411 9.9 39.3 448 AR Garland ........ 709 16.4 8.6 1m AL Montgomery. . .... 274 13.7 7.5 448 AR Clark .......... 729 16.4 27.1 171 AL Autauga ........ 274 13.7 18.8 448 AR Pig. . ivy ensws 729 16.4 27.8 71 AL PiKe.....s:00924 274 13.7 152 448 AR Montgomery. . .... 729 16.4 15.6 171 AL “Lowndes. , ...... 274 137 28.8 448 AR Hot Spring. ...... 862 16.4 27.3 7m AL Bullock Ni ae eA 274 13.7 19.3 457 AR White . . . ....... 738 205 13.0 7 AL Covington. ...... 360 13.7 20.6 171 AL. Crenseaw 360 13.7 19.8 457 AR Cleburne. ....... 738 20.5 35.6 Cs ’ : 457 AR Woodruff. ....... 738 20.5 25.4 175 AL Dallas. ......... 278 22.5 114 473 AR Jefferson. . . ..... 757 17.1 12.8 175 AL Penny .....caiu0 278 225 28.9 4 473 AR: DIBW i vv wsimw awe 757 17.1 17.2 175 AL Wilcox ......... 278 22.5 36.3 175 ALM & 404 205 357 473 AR Lincoln......... 757 17.3 15.4 BINGO’ coiries nt : : 473 AR Bradley. ........ 757 17.1 13.2 177 AL Calhoun ........ 280 18.3 15.7 473 AR Cleveland ....... 757 17.1 36.8 177 AL Cleburne. ....... 280 18.3 13.7 473 AR Desha ......... 858 7.3 30.0 i en Sand cascasnas 0 123 ore 480 AR Ashley ......... 766 37.8 436 Bie safe ’ 480 AR Chicot ......... 766 37.8 29.9 179 Al. LBB .issianrves 282 27.9 20.2 486 AR Union. ..ovonos 777 22.1 15.2 179 Al. EMO. ..:v++x 4 282 27.9 44.4 : 486 AR Columbia ....... 777 22.1 27.4 179 AL Tallapoosa. . . .... 282 27.9 16.1 486 AR Calhoun 777 22.1 62.1 179 AL. MEBBA . 29h 0p vw 282 27.9 38 | 0 TUE ’ ’ 185 AL Morgan. ........ 288 18.2 16.9 494 AR Boone ......... 788 22.8 17.6 185 AL Lowrends 288 18.2 205 494 AR Camroll . .oi vu ows 788 22.8 22.1 AWRACE =r 2tanica : 494 AR Searcy ......... 788 22.8 49.7 210 Al. Madison... vu 320 13.6 7.0 494 AR Newton. ........ 788 22.8 20.5 21) Al. mesons x. sx. 20 126 18.1 496 AR Ouachita. ....... 791 35.6 31.4 210 TN Lincoln... :se en 320 13.6 18.5 496 AR Dallas 791 35.6 47.2 210 AL Jackson ........ 445 13.6 218 { TC 7 HET eens : ’ 219 AL Lauderdale . . . ... 332 13.4 8.9 499 AR Coenen... 734 35.0 22 499 AR StFrancls..,...- 794 36.6 38.0 219 AL Colbert. ........ 332 13.4 8.9 ; 499 AR Cross. ......... 794 36.6 32.3 219 AL. Franklin .. vos ne 347 13.4 10.5 499 AR Lee 794 36.6 64.2 219 Al, WINSION nv wns 347 13.4 454 | 0 TT : ’ 224 AL Etowah......... 338 21.8 15.1 521 lb 904 269 265 224 AL Cherokee ....... 338 21.8 26.3 522 AR Johnson ..... +x 905 30.7 30.7 224 AL Von Cee 366 21.8 304 527 AR Arkansas. ....... 910 38.0 26.5 224 A. DeKalb : .u.ve un 366 21.8 22.7 507 AR Monroe. . ....... 910 38.0 62.0 241 AL Talladega ....... 372 33.0 31.2 4 AR PORK... ovximsns 91 46. 46. 241 AL Clay. wos ws nwen 372 33.0 28.4 LL ok 7 62 6.8 241 AL . COO8H mc ws sms 372 33.0 56.2 537 AR PhilipSis zs «ss wes is 920 30.4 30.4 247 AL Marion ......... 379 37.3 38.7 571 AR Craighead. ...... 723 17.9 6.9 247 AL Fayette. ........ 379 373 34.9 571 AR Poinset ........ 728 17.9 39.7 571 AR Lawrence ....... 723 17.9 14.3 259 AL Buller.......... 458 25.3 25% 571 AR Randolph . ...... 723 17.9 10.8 407 AR Bader .....osws 686 20.8 19.2 571 MO Dunklin. .....vx 982 17.8 38.2 407 AR Maron... + ow ds 4 686 20.8 27.2 50 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. Slate County area no. Area County area no. State County area no. Area County 571 AR Greene. ........ 982 17.9 6.4 768 CA San Bernardino. . . . 1288 13.2 13.2 571 ARC. .:covmvrnn 982 17.9 11.4 768 CA Riverside. . ...... 1288 13.2 13.2 574 AR Stone.......... 854 24.0 17.2 774 CA Imperial ........ 1309 6.3 18.3 574 AR Independence . . .. 958 24.0 25.4 774 CA San Diego. ...... 1310 6.3 5.6 574 AR Shap... ...usves 958 24.0 24.0 780 CA Lassen. ........ 1342 433 46.7 574 AR Zand . ..c. vue 00s 958 24.0 21.3 780 CA Plumas. .. oo 1343 433 40.2 574 MO Howell ......... 977 24.0 252 ( ' 0 rTHEEErmreeees ’ 574 AR FUROR. 5c on vine 977 24.0 22.1 781 CA San Luis Obispo. . . 1355 9.1 10.6 574 MO Oregon......... 977 24.0 19.9 781 CA Santa Barbara . . .. 1356 9.1 8.0 574 MO Shannofiu... aw. S77 24.0 43.1 789 CA Tulare. ......... 1390 143 143 614 AR Mississippi. . . .... 1010 35.9 33.3 7 A Vi 1 16.4 614 MO Pemiscot. . . . .... 1010 35.9 41.8 % CA. VEU. oxo ouvn 199 B, es CA H iat svn 7. s 699 AZ Maricopa. ....... 1212 7.5 6.5 300 YAORt 1401 5 5 699 AZ Pinal oii iemincue 1212 7.5 18.1 802 CA SantaCruz ...... 1403 13.7 13.7 699 AZ Gila........... 1212 75 10.1 807 CA Kem........... 1408 15.2 15.2 699 AZ Yavapai......... 1241 7.5 6.7 699 AZ Coconino ....... 1241 7.5 14.5 811 CA Mendocino ...... 1412 17.7 17.7 700 AZ Pima .......... 1213 6.8 6.3 816 CA NYG... .csvomsmas 1419 30.0 255 700 AZ Cochise ........ 1213 6.8 7.2 816 CA MONO. ....w «os 1419 30.0 63.5 700 AZ SantaCruz ...... 1213 6.8 5.2 688 CO DORVEr. «x vx is 1200 5.2 3.5 700 AZ Graham ........ 1327 6.8 16.9 688 CO Jefferson. . . . .... 1200 5.2 53 700 AZ Greenlee. ....... 1327 6.8 29.8 688 CO Amapahoe ....... 1200 5.2 6.9 787 AZ Yuma.......... 1388 24.3 24.3 688 CO Adams ......... 1200 5.2 5.9 688 CO Douglas «cvs» 1200 5.2 11.8 803 AZ Mohave ........ 1404 27.0 27.0 688 CO Summit. ........ 1200 52 24.4 690 CA Sutter. ......... 1202 18.8 177 688 CO Ebert. co: vvens 1200 5.2 40.6 690 CA: YUBd. ss n0sbs ns 1202 18.8 12.9 688 CO Grand... u% «bad 1200 52 24.8 690 CA Colusa......... 1367 18.8 36.9 688 CO Clear Creek. ..... 1200 8.2 8.3 697 CA Butte .......... 1210 1.2 11.2 oe oe pa. nhs agen 32% a ii 697 CA Glenn. ......... 1210 11.2 91 | ~~ TT Tre ’ ’ 697 CA Tehama ........ 1350 11.2 12.6 704 CO Pusblo....s us: 1217 10.6 7.4 709 CA Saziamorto. . . . . . 1222 78 6.8 704 co Huerfano 2 3B arma 1217 10.6 9.1 704 CO LasAnimas ...... 1308 10.6 13.5 709 CA Placer........»: 1222 7.8 10.6 704 NM Colfax 1376 10.6 25.5 709 CA Yolo. .: on ams sas 1222 7.8 7g 0 TE Meese ’ ’ 709 CA ElDorado....... 1334 7.8 11.7 711 CO Mesg..w:vuiwsp 1224 9.8 7.0 710 CA Shasta. ........ 1223 15.0 12.3 711 CO Garfield. ........ 1224 9.8 10.0 . 7M CO Eagle.» svuvs va 1224 9.8 30.1 710 CA Trinity. ....c.s:. 1223 15.0 23.6 : 710 CA Mod 1361 15.0 34.1 711 CO Rio Blanco. ...... 1224 9.8 17.3 DER 3 apie ts 44d 4 : 711 UT Grand. ......... 1297 9.8 24.8 718 CA FI88N0 .. conv sny 1231 7.7 6.0 71 CO PRIN. «cova 0s 1339 9.8 24.6 ns 2 Madge SEL TEEL Ji 2 He 731 CO Alamosa... ..... 1245 25.5 24.9 INES ; i vis 4 rw ins * / 731 CO Conejos ........ 1245 25.5 222 723 CA Los Angeles. . .... 1236 4.1 3.9 731 CO Costa. ..: uo. vss 1245 25.5 32.1 723 CA Orange. ........ 1379 4.1 49 731 CO Rio Grande . ..... 1338 255 17.6 737 CA Merced. ........ 1251 9.1 8.4 = 22 Sonus ie gassuay bi a a 737 CA Mariposa. . ...... 1251 9.1 202 § TY 020M HEE aewees ’ : 737 CA Stanislaus . . ..... 1312 2.1 7.2 735 CO Routt. ..:xious- 1249 30.4 26.9 737 CA Tuolumne ....... 1368 9.1 17.1 735 CO Moffat, .:vvvv +s 1249 30.4 335 746 CA Solano......... 1262 19.9 19.7 745 CO ler. conv 1261 30.3 23.1 746 CA Napa. «oman 1262 19.9 17.4 745 CO Bent .onsvnnmsrn 1261 30.3 275 746 CA Lake ....ecisuns 1362 19.9 24.0 745 CO Crowley ........ 1261 30.3 39.1 750 CA San Joaquin. . . ... 1267 145 11.6 Te go Jo, reba uray i Ms 27 750 CA Calaveras ....... 1267 14.5 S14 | TC 0200 THe ess ’ ’ 750 CA Amador. ........ 1335 14.5 30.0 754 CO EIPasO....::s%- Lg 11.4 7.8 751 CA SantaClara. ..... 1268 9.0 95 ¥oe OO Toor + mums wis ian 114 m7 : 754 CO Linco. « vrwiv mas 1271 11.4 42.9 751 CA San Benito. . ..... 1268 9.0 13.4 % 751 CA Moriere 1385 20 70° 754 CO KitCarson....... 1346 11.4 44.9 ees sinus : 754 CO Cheyenne. ...... 1381 11.4 41.9 2 a iy mama a Ss iy 760 CO Weld .......... 1279 19.3 20.1 EE ’ ’ 760 CO Morgan. ........ 1279 19.3 13.6 757 CA San Francisco . . . . 1275 10.1 6.4 760 CO Washington. . . . .. 1294 19.3 23.9 757 CA SanMateo. ...... 1275 10.1 15.9 760 CO Yuma.......... 1349 19.3 21.2 764 CA Sonoma ........ 1284 11.5 9.0 761 CO Montrose. ....... 1280 20.4 18.2 764 CA Matin. .... 544+ 1284 11.5 16.8 761 CO Gunnison ....... 1280 20.4 36.7 766 CA Alameda... ..... 1286 7.8 7.3 el £9 gor Miguel . ev u ue > 2s 333 766 CA Contra Costa . . . . . 1286 7.8 8.6 IRs ne we Rei oe . | 51 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 52 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 761 CO Hinsdale. ....... 1280 20.4 50.0 165 FL Hendry......... 268 15.6 20.5 761 CO Delta ..., sn + 1377 20.4 19.4 165 FlL.- Glades’, . ; +04 +» 268 15.6 42.1 763 CO Logan. ......... 1283 27.4 26.8 165 FL Collier ......... 418 15.6 18.9 763 CO Phillips. ........ 1283 27.4 25.8 183 BL Leoh :vvin: wuss 286 15.6 7.6 763 CO . ‘Sedgwick :.'«« . . . 1328 27.4 31.5 183 FL Gadsden. ....... 286 15.6 12.1 786 CO Chaffee. . ....... 1383 29.3 34.1 183 Fl. : Taylor: s os sews 286 15.6 15.4 786 CO Lake .... oo” 1384 203 21.4 183 FL Wakulla. . ....... 286 15.6 10.9 183 FL Jefferson. ....... 286 15.6 50.4 795 CO Boulder. ........ 1396 23.0 23.0 183 FL Franklin ........ 286 15.6 28.0 ; 183 FL“ Liberdy « :o3sw san 286 15.6 47.7 796 CO Larimer... :..««:. 7 14.1 14.1 GO ‘Larimer 138 183 FL Madison. ....... 419 15.6 25.4 812 CO Fremont ........ 1413 27.0 26.4 200 FL. Dade...» e029 310 8.5 4.6 Or wn vw swan 47. dh CO" Custer HS 270 75 200 FL Monroe. ........ 310 8.5 14.0 4 CT Hartford ........ 5 8.7 75 200 FL Broward ........ 432 8.5 12.6 4 CT: Tolland: v sous vas 5 8.7 7.7 : 202 BL. ‘Polk... .5osniw 312 14.3 13.4 4 T Windham ....... 5 8.7 18.4 } C dna 202 FL Highlands . ...... 312 14.3 16.9 85 CT New Haven ...... 97 10.6 8.0 202 Pl. Hordes. ..=: wus 312 14.3 17.4 85 CT “Lichfield . + » oz =» + 97 10.6 18.3 = or en. oe ro a6 213 FL Sarasota........ 323 19.7 20.8 Shasta ’ ’ 213 FL. Charoite......:. 323 19.7 173 121 CT Fairfield. . ....... 223 11.4 11.4 213 FL DeSoto ........ 323 19.7 16.8 3 MD Wicomico « . « «+ 4 8.3 5.8 221 FL St Lucie s+: x5 uss 335 17.4 17.3 3 MD Worcester . . ..... 4 8.3 12.0 221 FL Martin. ......... 335 17.4 17.0 3 MD Somerset ....... 4 8.3 5.5 221 FL.. PaimBeach....:. 376 17.4 17.6 3 DE "“SUSsoX. ..:s52 50 57 8.3 8.5 221 FL Okeechobee . . ... 376 17.4 13.0 DE Kent........... 57 5 s 32 9:3 227 FL Pasco.......... 343 8.8 10.7 75 MD Harford. «..o 5.» 85 13.6 26.8 227 FL Hernando ....... 343 8.8 10.1 75 MD Cecil .......... 85 13.6 11.1 227 FL Pinellas. ........ 413 8.8 8.1 75 DE New Castle . ..... 168 13.6 8.9 227 FL Hillsborough . . . .. 452 8.8 8.2 142 FL Orange......... 244 10.8 8.3 233 FL Manon ... uo vv vw 355 19.1 19.1 142 FL Seminole. ....... 244 10.8 7.1 233 FL CHIUS. . ius snes 355 19.1 19.1 Fi Volusia. ........ 10. 3 143 ~aVilusa ges g 58 237 FL Brevard. ........ 364 12.6 11.9 142 FL. Flagler ; . xs «xm 263 10.8 21.5 Indi A 142 FL Lake .......... 464 10.8 15.8 237 FL Indian River. . .... 364 128 14.2 142 FL Sumter......... 464 10.8 48.3 251 FL Putnam. ........ 420 31.8 33.5 155 FL Bay........... 257 20.6 95 251 FL StJotns ..ovsuie 421 31.8 30.2 155 Fl GU. ov rio iin 257 20.6 7.8 257 FL Osceola ........ 456 15.3 15.3 155 FL Washington... ... 325 20.6 20.0 155 FL Holmes. ........ 305 20.6 433 266 FL Manatee ........ 465 17.0 17.0 155 Fl, +.:JBCKSON ov gem oo 356 20.6 30.7 143 GA Chatham. ....... 245 9.4 45 155 Fl. .CRNOUN 5.56 00 va 356 20.6 29.8 143 GA Liberty .cowvvunn 245 9.4 9.9 158 LOWE 260 73 5.3 143 GA Effingham ....... 245 9.4 3.2 143 GA Tattnall. ........ 245 9.4 24.0 158 FL "Olay: iuvvurvs 260 73 17.0 143 GA Bryan. ..:us uv en 245 9.4 18.3 158 FL Nassau......... 260 7.3 6.0 143 GA Long . .. cv vra 245 9.4 44.6 158 Fl. Baker... .yoews 260 7.3 13.1 143 SC Beaufort ........ 354 9.4 15.6 158 GA GlyPN, . . «usps 305 73 71 143 SC - Jasper . is «isu 354 9.4 9.1 158 GA Camden ........ 305 7.3 52 143 Gh Evers 381 94 26.0 158 GA ‘Brantley ........ 305 7.3 389 | TTT FETs ’ 158 GA Mcintosh. . ...... 305 7.3 23.7 144 GA Dougherty. ...... 246 14.8 7.2 158 GA ‘Charlton. ...5.. « « «.» 383 7.3 9.9 144 GA Worth. ......... 246 14.8 19.2 159 FL Alachua ........ 261 127 8.1 144 OA 188 senrnnenan 240 18 20 144 GA Terrell. . ........ 246 14.8 11.0 159 FL Lewy. .......... 261 12.7 25.9 144 GA Baker. . «sas sain 246 14.8 19.1 159 FL. Bradford . «daa ws 261 127 19.8 144 GA. Eally .« vou 0 x mms 342 14.8 34.4 159 FL. UBIO: & + ocvnmin ais 261 12.7 6.3 144 GA Calhoun ........ 342 14.8 4.1 159 Fl, "DM 5 sv ssp ws au 261 12.7 21.2 i 144 GA Clay... +15: 5535 342 14.8 39.8 159 FL Gilchrist ........ 261 12.7 8.4 144 GA 'Randoloh 410 148 19.2 159 FL Lafayette. ....... 261 12.7 33.0 : BAC ie oi erinie : 159 FL . Columbia . .. «+4 302 12.7 7.2 145 GA Whitfield. ....... 247 12.6 13.0 159 FL Suwannee. ...... 302 12.7 14.4 145 GA Murray ......... 247 12.6 11.2 159 FL. “Hamilton. ; uz + 454 12.7 21.7 152 GA Richmond . . . .... 254 8.6 28 163 FL Escambia ....... 266 8.4 4.6 152 SC Aken.......... 254 8.6 6.5 163 FL SantaRosa ...... 266 8.4 4.9 152 GA Columbia ....... 254 8.6 4.5 163 FL Okaloosa. ....... 295 8.4 8.4 152 GA McDuffie. ....... 254 8.6 5.0 163 Fl. - Walton . . oi vss ss 295 8.4 171 152 SC Edgefield. . ...... 254 8.6 12.9 163 AL Escambia ....... 329 8.4 11.3 152 GA Linon, « :vsivims 254 8.6 36.5 163 AL Conecuh........ 329 8.4 31.8 152 GA Warren. ........ 254 8.6 14.1 152 GA Glascock. . +. wun 254 8.6 24.4 5 LA al 26% 156 Bs 152 GA Taliaferro. . ...... 254 8.6 61.0 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 152 SC Bamnwell ........ 382 8.6 29.6 190 GA Cobb.......... 297 25.7 25.7 162 GA Burke.......... 428 8.6 14.3 190 GA Douglas ........ 297 25.7 40.1 152 GA Jenkins. ........ 442 8.6 19.2 190 GA Paulding........ 297 257 17.4 153 GA Fulton. ......... 255 7.5 6.5 193 GA Bibb covuibnens 300 12.7 6.1 153 GA DeKalb. ....v:us 255 75 6.0 193 GA Houston ........ 300 127 8.9 153 GA Rockdale. . ...... 255 75 7.8 193 GA Peach. ..... +a 300 12.7 12.8 153 GA Forsyth. .......- 255 7.5 10.7 193 GA Jones.......... 300 12.7 26.8 153 GA Clayton. ....+.u: 294 7.5 59 193 GA Wilkinson ....... 300 12.7 52.8 153 GA Fayette. ........ 294 75 23.3 193 GA Twiggs. ........ 300 12.7 32.4 153 GA Henry. ......ius 294 7.5 8.5 193 GA Taylor... usmws 300 12.7 58.8 153 GA Gwinnett. ....... 330 75 71 193 GA Crawford. . ...... 300 12.7 25.0 153 GA Wallon... «.0.6 330 7.5 23.1 193 GA Bleckley . ....... 339 12.7 11.3 154 GA Floyd. ......... 256 13.0 5.0 a 2 Pia SERA i id £3 154 GA Polk. .......... 256 13.0 10.7 CRIOBY «ue x : : 154 GA Gordon. ........ 256 13.0 30.6 197 GA CHD co vv ivasvs 307 28.2 17.5 154 GA Chattooga. ...... 256 13.0 15.2 197 GA Dooly.......... 307 28.2 28.6 154 GA Bartow ......... 387 13.0 20.1 197 GA WICOX «i vs vos ns 307 28.2 66.1 157 GA Hall ........... 259 16.3 105 2 = Sy on ton AG 8 4 pe a a 157 GA Habersham . . . . .. 259 16.3 21.1 BOON vv 02 i» xiv ‘ A 157 GA Lumpkin........ 259 16.3 9.7 201 GA Baldwin. ........ 311 229 20.4 157 GA White... ....... 259 16.3 10.4 201 GA Putnam. ........ 311 22.9 26.8 157 GA Banks. ....cuous 259 16.3 52.5 201 GA Hancock. ....... 311 229 25.4 ls7 OA DaWSOR +0502 24s 259 123 Se2 204 GA Spalding. ....... 314 30.6 23.9 157 GA Union. . .. «ei vm 333 16.3 14.7 204 GA Butts 314 30.6 491 157 GA Towns .:s sunuss 333 16.3 84 {| = 2 Eh IRAE ’ : 164 GA Clarke ......... 267 15.4 8.0 206 GA LOWES vin runs 310 207 Na 206 GA Telfair. ......... 316 20.7 24.4 164 GA Jacksom .: .: sap = 267 15.4 22.0 206 GA Johnsom........ 316 20.7 40.0 164 GA Barrow. ........ 267 15.4 18.8 . 206 GA Montgomery. . . ... 316 20.7 27.2 164 GA Madison. ......: 267 15.4 17.7 206 GA Treutlen ........ 316 20.7 26.1 164 GA Oconee ........ 267 15.4 10.6 206 GA Wheeler ........ 316 20.7 16.4 164 GA Morgan. . ...uwss» 267 15.4 20.8 206 BAR DOBGB + rs wuncms 384 20.7 18.9 164 GA Oglethorpe . ..... 267 15.4 9.1 206 GA Toombs 436 207 26.8 164 GA Greene. ........ 422 15.4 280 | 7 20M EHErreraree ’ : 166 GA Muscogee. . . .... 269 10.7 5.9 215 GA, Wot. oo omgnpame Sar 242 315 216 GA Franklin ........ 327 24.2 24.6 166 Al. Bussell. ..iswsms 269 10.7 9.1 216 GA Stephens 386 24.2 18.6 166 GA Chattahoochee. . . . 269 10.7 14.6 POBBE ww tsps ’ 166 GA Harris. ......... 269 10.7 35.9 220 GA Newlon: ..::: «+ 334 45.0 42.6 166 GA Talbol ..« venus 269 10.7 373 220 GA Jasper ......... 334 45.0 57.6 1% SA SiBWAt a: anus 229 107 184 222 GA Bulloch. ........ 336 26.9 20.9 166 GA Manon « .. vs we ws 269 10.7 8.5 222 GA Emanuel. ....... 336 26.9 29.8 166 GA Schley ......... 269 10.7 54.4 166 GA Webst 269 10.7 26.4 222 GA Candler. ........ 336 26.9 23.1 BOBIEF + wieiv ew ee 7 : 222 GA Screven ........ 435 26.9 34.5 172 GA THOU: ws durin as 275 17.5 92 228 GA Decatur. ........ 345 32.2 32.2 172 GA Coweta. ........ 275 17.5 12.9 : : 228 GA Seminole. ....... 345 32.2 33.8 172 GA Meriwether. . . . ... 275 17.5 22.9 208 GA Miller 345 300 29.8 172 GA Heard. ......... 275 17.5 30 | 0 TTT orn ’ i 172 AL Chambers. ...... 406 17.5 21.7 230 GA Wayne ......... 349 28.7 21.0 172 AL Randolph ....... 407 17.5 31.7 230 GA Appling. ........ 349 28.7 41.1 174, GA Ware .......... 277 20.0 17.2 236 GA Coffee ......... 363 35.2 29.4 Joy 2 Pass bain aE Ye x 22 20 236 GA Jeff Davis . ...... 363 35.2 325 HEY meee : : Atki 19 0 Ged % i 174 GA CHACH «uv von mw 429 20.0 24.8 46 Ga thinson 909 a £25 250 GA Jefferson. ....... 414 38.2 43.3 178 GA Thomas ........ 281 21.4 13.0 250 GA Washington . . . . . . 415 38.2 32.9 178 GA Grady. ......... 281 21.4 11.1 178 GA Mitchell. . . ...... 281 21.4 51.8 253 GA Elbert. ......... 437 33.2 32.8 253 GA Wikes ......... 438 33.2 % 180 GA Lowndes. ....... 283 15.6 9.0 ] 8 180 GA Brooks ......... 283 15.6 33.0 254 GA Colquitt. ........ 453 32.5 32.5 180 OA COOK», ux sme 283 156 169 268 GA Rabun ......... 467 28.1 28.1 180 GA: EchOiS ... «ovis ms 283 15.6 10.3 180 GA Berrien. ........ 344 15.6 21.3 269 GA Upson ......... 468 42.1 238.5 180 GA Lanier.......... 344 15.6 9.1 269 GA Lamar.......... 468 42.1 65.4 ) 269 GA Pike........... 468 42.1 61.5 189 GA TH. ..ovevansns 292 20.5 20.4 189 GA BenHil. ........ 292 20.5 16.0 302 IL Stephenson. ..... 499 15.7 23.0 189 GA Turner ......... 292 20.5 28.7 302 IL Jo Daviess. ...... 499 15.7 20.1 189 BA WR. ov ses vs 292 20.5 23.1 302 IA Dubuque. ....... 512 15.7 6.5 302 IA. Jackson ........ 512 15.7 21.4 190 GA Cherokee ....... 293 25.7 19.1 190 GA Pickens. ........ 293 257 18.0 340 IA. Des Moines . . .... 998 26.9 15.5 53 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 340 IA Henry.......... 998 26.9 35.2 606 IA. Hamilton. ....... 999 27.1 14.3 340 IL Henderson ...... 998 269 59.3 606 IA Wright « ..ormian 1138 27.1 54.9 350 WI Crawford. . ...... 614 405 43.7 606 IA. Hardin ......... 1173 27.1 37.0 350 IA Clayton: « «+s vs.» 1009 40.5 34.5 617 IA Guthrie. ........ 1013 42.4 43.7 350 IA Delaware. . ...... 1009 40.5 46.7 617 IA Greene......... 1013 42.4 40.8 367 IA LBB wy wuwisima «ois 568 33.9 245 625 A Cla. .000 cima 2a 1024 29.7 25.9 367 IL Hancock. . ...... 568 33.9 41.3 625 IA Dickinson ....... 1024 29.7 30.1 367 MO Clark . coviwy oman 568 33.9 50.9 625 JA PalcAlfto. ....... 1136 29.7 34.4 395 IA Winneshiek . . .... 1144 35.6 35.6 633 A S30. ...00000 0 1034 32.6 42.9 : 633 A Cahoun........ 1034 32.6 37.2 545 IA bans ess vss van 928 10.5 5.8 633 A Caroll... 1079 206 210 545 IA Benton. ...... 928 10.5 8.7 545 IA. JONBB, «wna 2 emis th 928 10.5 14.6 634 A Floyd... comin aus 1035 41.2 44.9 545 JA Cedar. ci vuvanien 928 10.5 31.6 634 IA Chickasaw. . ..... 1035 41.2 35.6 545 IA. Johnson ........ 935 10.5 8.3 641 IA Scott .......... 1048 135 9.9 oe A Washagion Co > 102 3 641 IA Muscatine. . . .... 1048 135 25.7 545 IA LOWS wuiy way visi 935 10.5 59.8 644 IA. Page .......... 1051 48.0 39.4 644 IA Fremont ........ 1051 48.0 43.7 $8 A POR. cnimr ini 928 24 8.3 644 IA Taylor. ......... 1051 48.0 72.8 546 IA Wammen......... 929 8.4 9.2 546 IA Jasper......... 929 8.4 13.8 645 IA LICE, wo vais mis os 1052 40.8 41.2 546 IA. Dallas. ,..vu vss 929 8.4 11.0 645 IA Wayne ......... 1052 40.8 40.4 346 IA. Madison .. une ees 929 £4 5.9 647 IA Montgomery. . . . .. 1054 31.8 28.5 348 IA. DBE visens wut 928 24 105 647 IA Adams. ........ 1054 31.8 42.9 546 IA UNION. 5 cvs me men 1023 8.4 9.5 546 IA Adair .......... 1023 8.4 21.3 649 IA Appanoose . ..... 1056 27.4 23.3 546 IA Decatur. ........ 1088 8.4 9.1 649 NG RE, 1056 27.4 33.7 50 A BAS... siirne laos 34 35 652 A OUHBE Loins 1064 38.2 39.1 546 A MBER ww ven ny 34 27 652 IA. Osceola ........ 1065 38.2 36.0 556 IA Cerro Gordo... . . . 941 20.9 11.0 665 IA Mitchell. ........ 1177 4358 4338 556 IA Hancock. ....... 941 20.9 17.4 556 IA Winnebago . . .... 941 20.9 32.1 672 A Sioux.......... 1184 40.4 40.4 556 1A Franklin ........ 941 20.9 17.9 679 1A Marshall ........ 1191 39.2 26.8 556 A Worth.......... 941 20.9 19.3 679 A Tama. .....:... 1191 39.2 61.8 556 IA Kossuth ........ 1118 20.9 41.5 681 1A Crawford. . ...... 1193 48.7 48.7 557 IA Black Hawk . . .... 942 17.2 8.7 557 IA Bremer. ........ 942 17.2 11.6 686 IA Poweshiek. ...... 1198 30.4 30.4 557 IA Buchanan ....... 942 17.2 21.1 694 ID NezPerce....... 1206 14.4 12.1 557 IA Butler.......... 942 17.2 39.5 694 WA ASOHN. Jvcvsi san 1206 14.4 13.6 557 IA Grundy. ........ 942 17.2 35.8 694 ID Clearwater. . . . . .. 1206 14.4 7.6 557 IA Fayette. ........ 1078 17.2 303 694 ID Lewis.......... 1206 14.4 15.2 560 IA. Woodbury . . ..... 945 9.6 45 694 WA Garfield. ........ 1206 14.4 30.6 560 NE Dakota ......... 945 9.6 5.9 694 ID Idaho.......... 1278 14.4 21.0 560 IA- Monona . ....... 945 9.6 14.6 695 ID TwinFalls....... 1207 14.5 13.9 560 IA da... 945 9.6 8.8 695 ID Jerome. ........ 1207 14.5 12.3 560 NE Thurston. ....... 945 9.6 8.6 695 ID Gooding. ....... 1207 14.5 19.3 560 NE Dixon. ......... 945 9.6 18.7 695 ID Lincoln. ........ 1207 14.5 14.7 560 IA Plymouth. . .. . va» 1093 9.6 13.0 560 IA BuenaVista. ..... 1121 9.6 25.5 696 ID Bannock. ....... 1209 19.4 9.2 560 IA Cherokee . ...... 1122 9.6 8.1 696 ID Bingham Wi a Ne ew 1209 19.4 36.4 696 ID Power. ......::¢ 1209 19.4 20.7 585 IA Webster ........ 971 31.1 29.5 696 ID Carboy ..vuv.n-- 1357 19.4 25.4 585 IA° Humboldt ....... 971 31.1 255 696 ID Oneida......... 1392 19.4 25.3 585 IA Pocahontas . . .... 971 31.1 42.6 716 ID AG3...:.000500s 1229 6.9 7.8 589 IA Wapello ........ 975 229 13.0 716 D Gem .......... 1229 6.9 7.2 589 IA- Mahaska. ....... 975 229 28.9 716 ID . Valley. .cuumins 1229 6.9 7.3 589 IA Keokuk. ........ 975 22.9 41.9 716 ID Boise.......... 1229 6.9 20.6 589 IA Jefferson. ....... 1059 22.9 20.8 716 ID Emore......... 1302 6.9 11.8 589 IA. VanBuren....... 10569 22.9 32.8 716 ID Canyon. ........ 1365 6.9 4.7 589 IA. Monroe. ........ 1099 29 24.2 716 ID Owyhee ........ 1365 6.9 3.2 596 IA Pottawattamie. . . . . 986 222 17.3 722 ID Bonneville . . . .. .. 1235 16.2 14.6 596 IA Shelby ......... 986 22.2 18.5 722 ID Madison ........ 1235 16.2 8.6 596 IA Mills... 986 22.2 29.9 722 ID Jefferson. ....... 1235 16.2 9.0 596 IA. Cass .......... 1043 22.2 26.7 722 ID Fremont ........ 1235 16.2 18.3 596 IA Audubon, ....... 1043 222 20.9 722 ID Custer ......... 1235 16.2 60.7 596 IA Harrison ........ 1090 22.2 34.0 722 ID Butte.......... 1235 16.2 33.3 606 BA Got cag ve mes 999 27.1 20.7 722 ID Clk vse nvive in 1235 16.2 20.0 606 1A Boone . ........ 999 27.1 24.2 722 ID Teoh. oo 205 wes 1369 16.2 14.6 54 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked ——————————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 730 ID Cassia ......... 1244 29.4 27.2 333 IL Vermilion. ....... 530 21.1 14.8 730 ID Minidoka. ....... 1244 29.4 323 333 IN “Fountalty, . s ws ous 530 213 38.1 734 ID Kootenai. ....... 1248 20.7 17.2 on iN ig monn win in Si o y Sad 734 ID Shoshone . ...... 1248 20.7 14.2 GBP xw x 00 wise + 2 d 734 ID Bonner... ...... 1281 20.7 31.1 334 IL Macoupin ....... 531 37.4 43.3 734 ID Boundary ....... 1281 20.7 24.9 334 IL Montgomery. . .... 531 37.4 29.1 734 ID Benewah. ....... 1399 20.7 26.3 355 IL JelierSon. « . . _ Ro 18% 808 ID Blaine. ......... 1409 29.9 28.6 335 IL Wayne ......... 532 23.1 32.9 808 ID Camas......... 1409 29.9 375 356 LaSalle oe ns - 158 ie 813 ID Lemhi.......... 1414 29.8 29.8 336 IL Bureau......... 533 18.9 20.3 277 IL Peoria ......... 475 12.9 5.1 - y Pura Eh ow win Se a 5 277 IL Tazewell ........ 475 12.9 7.8 MOFSIOR . -acp 1 0 2 : : 277 IL Fulton. ......... 475 12.9 26.2 338 IL Mclean ........ 535 25.1 15.5 277 IL Mason ......... 475 12.9 34.6 338 IL Woodford . ...... 535 25.1 51.4 277 IL Marshall ........ 475 12.9 54.0 338 IL DeWitt ......... 664 25.1 34.1 279 IL Champaign . . . ... 477 11.9 7.6 343 IL Kankakee ....... 539 15.0 10.6 279 IL Douglas ........ 477 1.9 14.0 343 IL Iroquois ........ 539 15.0 26.3 278 LL POreunsnnvnss C4 ne 54 351 IL Richland. ....... 546 43.4 20.5 279 IL Coles. ......... 623 11.9 11.9 : LC 279 IL Cumberland 623 11.9 35.3 3s BY twosome 345 434 432 be PRL / : 351 IL Jasper ......... 546 43.4 72.9 287 he COOK... cuvmans 484 6.1 5.4 351 IL Edwards. ....... 546 43.4 48.7 267 IL. DuPage... sales 452 8. 78 353 IL Adams. ........ 548 12.9 6.7 287 IL Lake .......... 551 6.1 7.5 ‘ pod Ah Be oy 037 353 IL Pke........... 548 12.9 17.8 CAOAY wr 4 : 353 MO Lewis. ......... 548 12.9 175 291 IL Winnebago ...... 488 13.9 10.2 353 MO Marion «ci. .% vines 1036 12.9 11.8 291 IL Ogle .......... 488 13.9 24.9 353 MO Pike. .......... 1036 12.9 21.4 291 IL Boone ......... 488 13.9 72 353 MO Ralls .......... 1036 12.9 37.8 29 Lh. L88.uunemens ns 58 139 288 358 IL Marion ......... 554 2256 18.9 299 I. Madison........ 496 24.9 245 358 IL Washington. . .... 554 22.6 36.6 299 IL Jersey ......... 496 24.9 18.8 381 A CURR. eee oo es 28 ais 299 IL Cahoun........ 496 24.9 31.9 299 IL Bond 633 24.9 309 361 i . Conoll nv wi wus 557 25.3 48.7 OO i028 ¥ 0 2 : 361 IL Whiteside . ...... 602 25.3 18.4 hi k il ses mamas 3% : 22 363 IL Effingham... .... 560 25.4 18.4 Weir a fy 2 2 : : 363 IL Fayette. ........ 560 25.4 32.3 305 Le MOO cau o wens 502 54 7.0 373 IL Kane .......... 577 15.4 12.6 305 IL Piatt. .......... 502 15.4 35.7 ; 373 IL DeKalb......... 577 15.4 24.1 305 IL Moultrie ........ 502 15.4 20.5 or Lae hd bg nd 305 IL Shelby ......... 604 15.4 gem | Be Tm NONRLL sient es ” : 307 IL Rocklsland . . .... 504 14.5 8.4 i ES a snes Eo Sa 2 307 IL Henry. ......... 504 14.5 gap | = 2 = TERR: : ’ 307 IL Mercer. ........ 504 14.5 23.9 384 IL Saline. ......... 590 23.3 18.1 307 IL Stark .......... 504 14.5 55.6 384 IL Gallatin. ........ 590 23.3 34.0 318 IL Sangamon. ...... 515 11.4 57 Sos k Hari ne mule os 2 Ro z 2 318 IL McDonough. . . ... 515 11.4 11.6 PR wie mnyining 318 IL. ©Ca88..«h zens 515 11.4 35.8 273 IN Vanderburgh . .... 471 8.7 4.0 318 IL Menard. ........ 515 11.4 7.2 273 IN Warrick. . ....... 471 8.7 49 318 I. Schuyler. .. ..... 515 11.4 9.4 273 IN GIBBON. «vues ws 471 8.7 7.2 318 IL Brown ......... 515 11.4 416 273 IN Posey. ......... 471 8.7 6.2 318 IL Christian. ....... 628 11.4 14.2 273 IN Spencer ........ 471 8.7 59.8 318 I. Logan.......:.. 662 11.4 18.8 273 IL Wabash ........ 658 8.7 17.0 319 I Morean. ..; exe 516 21.8 177 275 IN Maron ......... 473 6.2 4.7 319 IL Greene. ........ 516 21.8 30.8 275 IN Johnson........ 473 6.2 8.6 319 IL Scott .......... 516 21.8 22.5 275 IN Hendricks . ...... 473 6.2 8.8 324 IL KNOX «vvvvvn 521 10.3 9.3 278 IN Morgen. oven sms ii 22 55 Le ll ey be on 275 IN Putnam. ........ 620 6.2 11.7 pF LB Al, 2 : 275 IN Rush .......... 644 6.2 18.4 325 IL St.Clair ........ 522 16.6 16.2 275 IN Shelby ......... 645 6.2 10.0 sec fH BERCOIDN «uwsisue 522 162 121 300 IN Tippecanoe . . . . .. 497 15.0 9.1 325 I. MODE. ...c0ems 522 16.6 20.7 2 305 IL Clinton 663 16.6 219 300 IN CHMON sic sv wasims 497 15.0 23.7 nc malate Ab 2 5 300 IN White. ......... 497 15.0 15.8 331 IL Williamson. . ..... 528 14.7 10.3 300 IN Carroll ......... 497 15.0 24.1 331 IL Franklin ........ 528 14.7 15.9 300 IN Benton......... 497 15.0 14.1 ik y i 2HrEe mi Se §od a 304 IN Allen .......... 501 9.3 87 BW mnsmimy ans : 304 IN Noble. ......... 501 9.3 18.6 55 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 304 IN Lagrange ....... 501 9.3 21.0 385 IN Fayette. ........ 592 16.4 17.3 304 IN Whitley. . ....... 501 9.3 7.4 385 IN Union. ......... 592 16.4 46.6 is vg BE bn nw i 2 dv 389 IN Grant. ......... 621 24.4 17.4 BRD ov: 3h 3 20 8 : 389 IN Miami. ......... 622 24.4 33.2 308 IN Lake .......... 505 8.8 6.8 389 IN Wabash ........ 666 24.4 30.9 $08 IN Boh... wens $03 £8 hae 390 IN Hamilton. ....... 642 22.6 45.1 308 IN“ 'JB8per &.: vs swiss 505 8.8 20.7 390 IN Madison 643 22.6 13.7 308 IN Newton. ........ 505 8.8 Srp | SH WW SiedsfN..cesicens : - _ NH «on 507 0a 7 394 IN Dubois. ........ 676 10.9 10.9 310 IN Daviess. ........ 507 20.8 16.6 401 IN Boone ......... 680 34.3 34.3 310 IL Lawrence ....... 507 20.8 14.8 402 IN Montgomery. . . . .. 1 3 ” 310 IN Pike. .......... 507 20.8 65.2 ) BRigome:y. 83 g27 27 310 IN Martin, ......... 507 20.8 63.5 551 KS. «. FOI. vil eve iid 934 22.3 17.5 310 IL Crawford. ....... 640 20.8 18.9 551 KS Gray .......... 934 22.3 31.0 aN Vigo... 875 101 | SS ks meade... toi 25 203 31 IN Clay. .vuuiinrms 08 17s 15.9 551 KS Clark .......... 1040 223 32.0 nM IN Vermillion ....... 508 17.5 23.7 311 IL Clark ...oovn... 508 17.5 53.8 554 KS Shawnee. ....... 938 8.9 5.3 311 IN Parke.......... 508 17.5 21.4 554 KS Jefferson. ....... 938 8.9 27.3 311 IN Sliven. . cue 661 17.5 15.3 554 KS Osage ......... 938 8.9 24.4 312 IN St Joseph. . ..... 509 12.4 95 554 KS. JACKSON . .v uv as 938 8.9 4.8 554 KS Wabaunsee. ..... 938 8.9 29.9 312 IN Marshall ........ 509 12.4 13.4 Sat KS Riley a by cig 312 IN in wie issn 591 1s 24.0 554 KS Pottawatomie . . . . . 993 8.9 3.6 312 IN Pash; is enenn 81 2. 86.4 554 KS Brown ......... 1070 8.9 9.4 313 IN Bartholomew . . . . . 510 18.4 11.9 554 KS Nemaha ........ 1071 8.9 5.2 313 IN Jennings. ....... 510 184 157 562 KS Grant.......... 947 24.0 27.3 313 IN Decatur. ........ 639 18.4 28.5 562 KS Stanton. } 947 24.0 22.0 313 IN Jackson ........ 674 18.4 20.4 562 CO Baca .......... 1386 24.0 205 321 IN Jefferson. ....... 518 23.0 21.3 565 KS Ellis... ........ 950 10.0 10.2 321 KY Carroll ......... 518 23.0 23.7 wi oo Bers bine or one 321 IN Switzerland . . . . .. 518 23.0 10.0 Sob Ke Crotany orn a0 oe ig 321 KY Trimble. ........ 518 23.0 28.0 ir KE Oye [mamas ou ip bigs 321 IN Dearborn. ....... 559 23.0 23.7 565 KS Logan. ......... 1039 12.9 19.5 321 IN Ripley.......... 559 23.0 27.7 565 KS Trego. ......... 1131 12.2 12.1 321 IN OhiO .......... 559 23.0 18.1 323 IN Delaware. . . . .... 520 17.8 10.6 567 BS LY0F owmpms 950 ios 1 567 KS Coffey ......... 952 19.6 30.1 323 IN Randolph ....... 520 17.8 37.2 20s KS Ome Ser ig os 323 IN Blackford... ...... 520 78 215 567 KS Greenwood . . . . .. 1128 19.6 28.4 329 i cwrence vse on pis 2 A wz 569 KS Saline. ......... 954 14.4 14.2 320 range. ........ 526 A 33. 569 KS Oftawa......... 954 14.4 6.6 332 IN Huntington. . . .... 529 26.4 31.5 569 KS Lincoln. ........ 954 14.4 8.3 332 IN Wells. ......... 529 26.4 15.5 569 KS Ellsworth. : . ..... 1094 14.4 23.8 332 N ARIE viv wien ue 224 Ses 575 KS Finney ......... 960 16.9 15.5 332 JBY «viv singers 6 6.4 A 575 KS Kearny ......... 960 16.9 8.3 339 IN Floyd. ......... 536 21.6 19.3 575 KS Haskell. ........ 960 16.9 20.9 339 IN Harrison . ....... 536 21.6 17.3 575 KS Scott .......... 1058 16.9 17.3 339 IN Crawford. ....... 536 21.6 42.8 575 KS Lane .......... 1058 16.9 22.8 339 IN Washington . . . . . . 625 21.6 23.5 575 KS Hamilton. ....... 1082 16.9 21.2 341 IN Monroe. ........ 537 21.7 1.3 576 KS Sedgwick . ...... 961 5.7 4.9 341 IN Greene. ........ 537 21.7 27.6 576 KS Butler.......... 961 5.7 6.2 341 IN Owen.......... 537 21.7 335 576 KS Sumner......... 961 57 85.2 341 IN Brown ......... 537 21.7 58.8 576 KS Harvey ......... : 981 5.7 7.6 349 IN Ekhart......... 545 19.7 12.9 322 ke om FRE AE 4) ja 22 349 IN Kosciusko . . . .... 545 19.7 36.0 Yet teh : k Sin HE resins #58 08 oh 583 KS Barton ......... 969 21.2 17.2 362 IN H £ 558 303 32.0 583 KS Pawnee ........ 969 21.2 20.6 ANODE v's vm xin 3 : 4 583 KS Rush .......... 969 21.2 35.1 364 IN Howard. ........ 562 14.8 12.3 583 KS Edwards. ....... 969 21.2 24.2 So i Inn mets wes 2% Jae 182 587 KS Seward......... 973 32.1 24.9 BIS ruetmemoe 9, 587 OK Texas. ......... 973 32.1 41.2 365 IN Clark .......... 563 30.6 31.9 587 OK Cimarron. . . ..... 1038 32.1 35.3 365 IN Scott .......... 563 30.6 26.3 587 KS Morton. ........ 1038 32.1 15.1 378 IN LaPorte ........ 582 17.6 13.8 598 KS Remo .......... 988 22.8 19.1 378 IN Starke ......... 582 17.6 33.0 598 KS Rice........... 988 22.8 32.5 40s IN Waye......... i 164 54 598 KS Stafford. ........ 1108 22.8 31.4 56 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinkey —————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 601 KS Labelte. ..: 20500 992 23.6 22.3 12 IL Johnson........ 13 10.6 50.4 601 KS Neosho. ........ 992 23.6 19.0 12 KY: Ballard ; . ou 004 13 10.6 4.4 601 KS. Allen ....simwams 1164 23.6 22.3 12 IY = ILYOH oo vi nie aig 13 10.6 40.4 601 KS Woodson ....... 1164 23.6 45.2 12 KY Hickman. .....:. 13 10.6 29.6 618 KS Pratt. .......... 1015 31.1 27.2 2 RY. Caigle. ..1...00 3 195 59 ; 12 KY Livingston. ...... 15 10.6 55 618 KS Kiowa. ......... 1015 31.1 31.9 12 KY Crittenden 15 10.6 6.2 618 KS Comanche. ...... 1015 31.1 422 { r= 2 eneeartaen ’ ’ 621 KS Wyandotte. . . .... 1018 21.2 22.7 » RY PlSerrvswviesn i“ zs 25.0 621 KS Leavenworth 1018 21.2 15.4 3 WV" Mingo... «vein vs Ei 233 209 Tr : ’ 13 WV Logan... «sues » 165 23.3 23.0 622 KS Mitchell. . ....... 1019 37.3 19.8 622 KS Osbome........ 1019 37.3 49.0 5 BY FOYE... os vo ns 20 24 0s 622 KS Jewell 1019 37.3 60.8 Is KY JOSSAINGs. meh 20 24 42 KR 18 KY Woodford ....... 20 13.4 6.3 624 KS Johnson ......«. 1021 373 47.3 18 KY Breathitt . ....... 20 13.4 12.8 624 KS “Miami. o. 0 aa4 9. 1021 373 20.5 18 KY LBB. csvssmwimms 20 13.4 28.6 624 KS "Franklin... ..c.us 1021 37.3 17.8 18 KY Wolfe. ......... 20 13.4 45.9 624 KS Anderson ....... 1086 37.3 33.1 18 KY Owsley......... 20 13.4 47.3 624 KS Douglas ........ 1140 37.3 13.0 18 KY . PBIY ... cc. crimp 27 13.4 11.4 630 KS Crawford... ..... 1030 216 16.5 > Ny fre SEIELATEE > 4 50% 630 KS Bourbon. ....... 1030 21.6 16.6 Ds ; ? 630 KS Linn. .......... 1030 21.6 56.8 18 KY Scoft.......... le9 2 0s 18 KY Clark... .suss030 203 13.4 73 635 KS Marshall ........ 1041 43.5 43.9 18 KY Powell ......... 203 13.4 21.2 635 KS Washington... ... 1041 43.5 43.0 27 KY Boyle.......... 30 205 16.2 636 KS Rawlins. ........ 1042 43.6 43.0 a7 KY Lincon......... 30 20.5 17.4 636 KS Cheyenne....... 1042 43.6 44.4 27 KY Tosgyisss:%sons 30 20.5 32.7 637 KS Phillips. ........ 1044 34.8 35.2 27 KY Garard......... 172 20.5 17.8 637 KS Norton ......... 1044 34.8 45.8 27 KY Mercer. .. «uw: 212 20.5 225 637 KS 8mith...c40:m3 1142 34.8 23.4 29 KY Warren. ........ 32 14.1 11.4 642 KS Cloud. ......... 1049 28.6 33.6 2 Ny fen et a RIE 8 > 4 i 64 KS Republic. ....... 1049 28. 21.3 TOS oh 0 ill : : 2 Pipi 8 29 KY Butler. ......... 32 14.1 13.8 643 KS Wichita, . «i ws wus 1050 46.4 49.3 29 KY Edmonson. ...... 32 14.1 17.2 643 KS Greeley. .. «+: 1050 46.4 41.5 29 KY Logan. ......... 207 14.1 16.7 650 KS Geary.......... 1057 32.0 277 37 KY Pulaski......... 40 28.0 21.9 650 KE. .MOWIB. «ovo vows 1057 32.0 32.8 37 KY Wayne ......... 40 28.0 21.9 650 KS Dickinson . ...... 1104 32.0 34.0 37 KY McCreary ....... 40 28.0 51.7 659 KS Sherman........ 1110 34.3 30.1 40 KY Henderson ...... 43 20.7 15.7 659 KS Wallace. ........ 1110 34.3 73.4 40 KY Union.......... 43 20.7 32.1 659 KS THOMAS . wv wviv mis 1111 34.3 30.3 CL 45 KY Christian. ....... 49 19.6 15.4 661 KS Harper......... 1129 34.5 26.6 45 KY Todd .......... 49 19.6 51.2 661 KS Kingman. ....... 1130 34.5 445 45 KY Tigg «:csvvemeo 49 19.6 17.5 662 KS Clay. .......... 1174 38.3 38.3 45 KY Hopkins ........ 107 19.6 7.8 45 KY Webster ........ 107 19.6 51.5 664 KS Stevens ........ 1176 50.0 50.0 45 KY Caldwell ........ 107 19.6 19.2 667 KS Russell......... 1179 40.7 40.7 45 KY Muhlenberg... . . . 178 19.6 17.0 669 KS Barber ......... 1181 26.4 26.4 51 KY Whitley. ........ 55 23.5 30.2 } 51 KY KNOX .vcme 2 oem 55 23.5 23.6 671 KS Atchison. ....... 1183 25.1 25.1 51 KY LAaWe. « ovo iun a 143 23.5 16.7 674 KS McPherson . ..... 1186 28.0 28.0 51 Ky Clay........... 217 23.5 19.4 675 KS Wilson ......... 1187 33.1 33.1 53 KY Barren ......... 59 19.9 14.0 53 KY Hat........... 59 19.9 40.6 682 KS Ness.......... 1194 45.8 458 53 KY Metcalfe ........ 59 19.9 16.1 2 KY Kenton......... 2 14.1 12.0 53 KY Monroe. ........ 179 19.9 15.8 2 KY Boone ......... 2 14.1 125 62 KY Madison ........ 68 27.9 224 2 KY Grant.......... 2 14.1 15.5 62 KY Estill .......... 68 27.9 25.4 2 KY Gallatin. ........ 2 14.1 42.0 62 KY Jackson ........ 68 27.9 50.9 2 KY Campbell ....... 193 14.1 125 62 KY Rockcastle. . . . . .. 137 27.9 25.2 2 KY Pendleton....... 193 14.1 38.0 67 KY Rowan ......... 74 19.9 125 1 KY Floyd. 000000 12 23.2 20.2 67 KY Morgan. ........ 74 19.9 15.4 11 KY Johnson elk Arte 12 23.2 22.1 67 KY Eliott. ......... 74 19.9 20.8 1 KY Magoffin. ....... 12 23.2 39.8 67 KY Menifee ........ 74 19.9 50.5 12 KY McCracken ...... 13 10.6 3.2 82 KY Taylor.......... 93 30.1 26.9 12 KY Graves , ..j «us sus 13 10.6 12.8 82 KY Green. ......... 93 30.1 42.3 12 KY Marshall ........ 13 106 145 82 KY Adair.......... 138 30.1 295 12 IL Massac......... 13 10.8 27 82 KY Russell......... 209 30.1 24.9 57 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine Percent of routine stays outside 800-unlinked stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked UNHINKEY estimation unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 89 KY Hard o.uiowess 103 221 16.6 423 LA St Helena....... 702 17.0 17.2 89 KY Nelson .....oon. 108 22.1 36.4 424 LA Lafayette... ..... 703 6.7 47 89 KY Meade .:....... 103 22.1 55.8 : 424 LA Iberig: cows van 703 6.7 5.8 89 KY LaB i ueiwe ins 103 22.1 16.1 : 424 LA St.Martin ....... 703 6.7 11.1 89 KY Grayson... « «us 189 22.1 125 x 89 KY Breckinridae 201 25 1 204 424 LA Vermilion. . .: « «a» 820 6.7 1.9 POSNER: vices 2 : 424 LA Acadia ......... 844 6.7 13.4 28 J Haigh on ces 15 157 228 439 LA Lincoln. ........ 719 33.1 17.2 96 KY Bell:.siwivvsos 116 19.7 13.7 ) 96 IN Claiborn 424 19.7 03.4 439 LA Jackson ........ 719 33.1 24.0 BIDOISE ee fee sit : : 439 LA Bienvile . ....... 719 33.1 58.3 102 KY BOWOON : suis 54 125 32.5 37.4 439 LA: UAIOR. , coun os 886 33.1 42.5 102 KY Nicholas... u. 4s 125 825 52 443 LA Ouachita. ....... 724 4.9 3.2 104 RY Clinton ....sssss 127 28.8 35.8 443 LA Franklin ........ 724 4.9 6.7 104 KY Cumberland. . . ... 127 28.8 21.6 443 LA Richland ........ 724 4.9 4.0 114 KY Franklin ........ 170 35.4 20.8 43 LA ‘Wom Catal. «x » wx 724 4.9 14 443 LA Caldwell . ....... 724 4.9 6.1 114 KY Anderson ....... 170 35.4 53.6 114 1 Owen 171 35.4 41.8 443 LA Morehouse ...... 811 4.9 6.6 REA EE ’ ’ 443 LA EastCarroll . ..... 1418 4.9 20.5 115 KY Calloway. ....... 176 24.2 15.5 450 LA Fepides ........ 731 12.0 6.7 115 IN Caroll viv ssw 370 21.2 27.8 450 LA Avoyelles. . ...... 731 12.0 10.3 115 TN Benton......... 370 21.2 24.2 115 TN Henry 403 212 16.4 450 LA Allen ...vvu: vas 731 12.0 19.6 EEL sure : 450 IA Gant.......... 731 12.0 1.9 116 KY Fleming ........ 182 26.7 27.2 450 LA LaSalle ........ 759 12.0 9.9 116 KY Mason .......a 183 26.7 19.0 450 LA Catahoula ....... 759 12.0 56.8 ne KY Bracken . sue viens 183 267 429 463 LA Stlandry....... 745 15.1 18.5 119 KY HamBon ..csvova 221 24.8 20.2 463 LA Evangeline. ...... 745 15.1 8.9 119 KY Robertson. ...... 221 24.8 58.3 467 TH, mm 750 125 85 123 KY Lefcher......... 225 23.7 28.7 467 LA Lafourche ....... 750 12.5 8.9 131 KY Montgomery. . . . . . 233 34.6 206 467 LA Assumption. ..... 750 125 39.1 131 KY Bath... vice ome 233 34.6 53.8 470 LA Calcasieu ....... 754 11.6 6.2 272 BY Jeon: ovens 3 5.1 3.7 470 LA Jefferson Davis. . . . 754 11.6 19.2 . 470 LA Cameron. ....... 754 11.6 6.0 272 KY Bulli... ... na 3 5.4 7.4 270 KY Oldham ........ 3 5.1 5.8 470 LA Vermon......... 798 11.6 29.6 270 KY Shelby ......... 3 5.1 9.5 470 LA Beauregard... ... 798 11.6 8.3 272 KY Hemty. .:covosna 3 5.1 12.0 500 LA St. Tammany ..... 795 18.9 16.4 272 KY Spencer........ 3 5.1 11.2 500 — LA Washington... ... 795 18.9 18.2 272 KY Maron... wes 218 5.1 245 500 MS Pearl River. . ..... 879 18.9 25.3 272 KY - ‘Washington's z wus + 22 51 75 523 LA Clabome . ...... 906 335 335 316 KY Daviess. :..... x» 69 11.1 6.5 316 KY Ohio .......... 69 11.1 11.4 528 LA StMary........ 911 29.0 29.0 316 KY Mclean ........ 69 11.9 24.0 530 EA WIN o. snow 5 sown ow 913 35.9 35.9 3s Praca #@enm swe oo on iy 22 MA Middlesex . . . . . .. 24 5.3 46 BY oe sirint als fae : 4 22 MA Suffolk . ........ 24 5.3 4.1 403 LA Orleans. ........ 682 4.0 33 22 MA Norfolk . . ....... 24 53 6.8 403 LA Jefferson. ....... 682 4.0 3.4 22 MA Plymouth. . ...... 24 5.3 5.2 403 LA St.Bernard ...... 682 4.0 2.2 22 MA Barnstable. . ..... 160 5.3 8.1 403 LA Plaquemines . .... 682 4.0 25 32 MA Hampden ....... 35 6.5 6.3 403 LA SL.James ..: uv» 682 4.0 47.6 30 MA Hampshire 35 65 73 403 LA St. Charles ...... 771 4.0 13 [| 2 2 TF TEEEReessms ’ 403 LA St. John The Baptist. 77 4.0 5.6 68 MA Bristol. . ........ 75 16.8 17.3 416 LA Caddo ......... 695 52 37 68 Bl Nowpom. vin sau 1s a8 Be 416 LA Bossier. ........ 695 5.2 2.8 74 MA E8SBX....csvven 83 18.2 13.7 416 LA DeSoto......... 695 52 3.6 74 NH Rockingham. . . . .. 83 15.2 22.3 416 LA Sabine... sven 695 8.2 10.7 101 MA Worcester. ...... 123 10.3 9.9 416 LA Natchitoches . .... 753 5.2 14.7 , 101 MA Franklin ........ 123 10.3 14.7 416 LA RedRiver ....... 753 5.2 1.9 416 LA Webster ........ 813 5.2 53 111 MA Nantucket . ...... 161 28.4 28.4 419 LA East Baton Rouge . . 698 8.2 4.9 112 MA Berkshire ....... 162 13.7 6.2 419 LA Livingston ....... 698 8.2 28.2 112 NY Columbia ....... 163 13.7 16.9 419 LA Ascension. ...... 698 8.2 7.6 112 NY Greens......... 163 13.7 39.4 419 LA Iberville. ........ 698 8.2 4.5 120 MA Dukes ......... 200 25.7 25.7 419 LA West Baton Rouge. . 698 8.2 55 419 LA East Feliciana. . . . . 698 8.2 7.8 1 MD Allegany ........ 1 9.0 3.8 419 LA West Feliciana . . . . 698 8.2 8.7 1 WV Mineral. ........ 1 9.0 7.6 419 LA Pointe Coupee. . . . 824 8.2 7.2 1 MD Garett .......:- 145 9.0 11.8 ) 1 WV Grant.......... 146 9.0 18.8 423 LA Tangipahoa. . .. .. 702 17.0 17.0 1 WV Hardy. ......... 146 9.0 52.9 58 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 16 MD Baltimore. . ...... 18 75 6.2 309 MI Kent.....:vovvon 506 13.0 8.5 16 MD Howard......... 18 75 17.7 309 Ml lonia .......... 506 13.0 28.3 16 MD Caroll... 51 18 75 18.9 309 MI Montcalm ....... 506 13.0 14.3 16 MD Anne Arundel. . . .. 148 75 8.1 309 MI Mecosta ........ 617 13.0 21.3 16 MD Calvert... ...v ss: 149 75 16.1 309 M Gratiot .....000 655 13.0 20.4 61 MD Prince George . . . . 67 7.2 7.8 317 MI Wexford ........ 514 38.5 34.9 61 MD Charles. ....«. +s. 67 7.2 8.3 317 MI Osceola .....s:x 514 38.5 34.5 61 MD Montgomery. . . . .. 78 72 9.5 317 MI Missaukee. . ..... 514 38.5 28.1 61 DC The District . . .... 78 72 4.8 317 ME Lake cue uavimn on 514 38.5 61.6 61 MD St.Marys. ....... 213 7.2 10.3 320 MI Ottawa ......... 517 31.9 30.2 107 MD Queen Annes. . . .. 132 22.0 45.1 320 MI Allegan......... 517 31.9 34.7 107 YD Osehesier snus is 20 138 | Wm MI Saginaw ........ 519 19.2 11.8 107 MD Talbot. ....c:vcvx- 132 22.0 11.2 107 MD Caroli 132 25.0 21.7 322 MI Tuscola. ........ 519 19.2 36.4 IOUS «=. lz il o * 322 MI Huron. ......... 648 19.2 27.7 24 MD Kent. vs covme wns 220 205 206 327 MI Berrien. ........ 524 27.9 21.8 9 ME Cumberland. . . . .. 10 8.9 7.8 327 MI VanBuren....... 524 27.9 33.9 9 ME Sagadahoc ...... 10 8.9 12.9 327 Mii, Cass i uitaww 524 27.9 43.0 2 Ne i) oawaine ime a 22 ry 328 MI Genesee. ....... 525 10.9 9.2 EDI it 4m : 328 MI Lapeer. ........ 525 10.9 26.0 i7 ME Ponobsotl...:s:a 19 10.9 22 330 MI Midland ........ 527 22.0 15.8 17 ME Piscataquis . ..... 19 10.0 8.6 . 330 MI Gladwin ........ 527 22.0 25.2 17 ME Hancock... ««+: x 124 10.0 7.0 7 ME Washingt 124 100 94 330 MI Isabella. ........ 575 22.0 28.5 BEAN. +10 k ne : 330 Ml Clare. ......... 575 22.0 223 38 ME Kennebec ....... 41 14.8 13.0 342 Ml Bay........... 538 196 123 38 ME Somerset ....... 41 14.8 10.7 a8 ME Waldo 147 148 27.0 342 MI Ogemaw. ....... 538 19.6 23.0 Tory ' ’ 342 MI Arenac......... 538 19.6 25.2 95 ME Androscoggin. . . . . 115 1.7 6.8 342 MI Oscoda. ........ 538 19.6 55.0 95 ME Franklin .....«.. 115 11.7 17.4 342 MEL I0SCD vu soins ve 605 19.6 29.2 9s ME OXORE .vvimssines 157 7 67 348 MI Kalamazoo. . . . ... 544 13.9 12.8 98 ME YOK. :.oviissmsn 119 23.2 26.8 348 MI St. Joseph. ...... 544 13.9 16.4 98 NH Strafford. ....... 119 23.2 15.9 359 MI Houghton . ...... 555 20.8 213 136 ME Aroostook . ...... 238 9.1 9.1 359 Ml Baraga. ......u 555 20.8 20.1 271 MI Grand Traverse. . . . 470 18.7 9.6 53 ML Kewoohaw. .v vov + 555 20.8 180 271 Ml ADEM . voier was 470 15.7 36.1 371 MI Alpena... ..ccs en 574 38.2 18.7 271 Ml Leelanau. ....... 470 15.7 12.6 371 MI Presque Isle. . . . . . 574 38.2 37.5 271 MI Kalkaska........ 470 18.7 19.5 37 MI Alcona ......... 574 38.2 52.9 an Ml Benzie ......... 470 15.7 1341 371 MI Montmorency. . . . . 574 38.2 70.9 271 YI Mansien. me omen oe 87 154 ars MI Hilsdale . . ...... 579 24.3 24.1 274 M Wayne......... 472 4.4 3.2 375 Ml Branch......:s5 579 24.3 30.2 274 MI Oakland ........ 472 4.4 6.6 375 MI Jackson ........ 649 24.3 17.9 274 MI Macomb. .. «ass 472 4.4 3.9 375 MI Lenawee........ 650 24.3 32.2 on ul van “hums piv 34 No 381 MI Calhoun ........ 586 16.9 13.1 GOR «=v 2x : 381 Ml Barry .......... 586 16.9 33.3 253 Ni BORA. vss ns 423 150 12d 392 MI Roscommon. . . . .. 656 47.4 58.5 285 MI Ealon...:ueemss 483 16.0 179 : 392 MI Crawford. ....... 656 47.4 25.6 285 MI Clinton. ........ 483 16.0 19.9 392 MI Otsego 657 47.4 36.0 285 MI Shiawassee. . .. .. 626 16.0 24.1 80: tw EPC re ! ’ 293 MI Marquette . . . .... 490 9.0 73 286 WI Polk Ep 561 16.0 16.8 286 MN Chisago ........ 561 16.0 25.1 293 MIL Alger scien: ws on 490 9.2 9.3 286 MN Ramsey ........- 940 16.0 14.5 293 Ml Dela... osmsu0 571 2.2 10.9 286 MN Dakota......... 940 16.0 22.4 293 MI Schoolcraft . ..... 57 9.2 7.9 286 MN Washington 940 16.0 119 293 Mo Luce :......... 598 9.2 17.8 GION: wre 3 : 294 MI Emmet. ........ 491 15.9 12.3 259 UN SLiods..esnvns 450 107 89 289 WI Douglas ........ 486 10.7 8.6 294 MI Cheboygan ...... 491 159 1.7 : 289 MN Cadion. ....o«uu 486 10.7 6.1 294 Ml Charlevoix. . . .... 491 15.9 129 4 ; 289 MN PRB. . ou oss 4.00 486 10.7 56.3 294 MI Mackinac ....... 491 15.9 37.7 294 Ml Chibpewa 667 15.9 14.1 289 MN Lake! ou. vwinms 603 10.7 6.8 POW, oc v2 : 289 MN Cook .......... 1085 10.7 19.3 296 MI Muskegon. ...... 493 16.7 10.3 289 MN Itasca.......... 1153 10.7 13.0 296 MI Newaygo. ....... 493 16.7 37.8 289 MN Koochiching. . . . .. 1154 10.7 17.3 296 Ml Oceana. ........ 493 16.7 16.9 i 296 Ml Magen «oo 251 167 27 396 MN Rice. .......... 1168 25.4 25.4 297 MI St.Clair ........ 494 2256 20.8 30 MN. HeAnepIn «vss 923 liz 75 297 Ml Sanilac 494 2.6 27.9 540 MN Anoka ......... 923 71.2 16.4 tro ! : 540 MN Scott c.vvvinnan 923 11.2 20.6 59 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 540 MN Carver ......... 923 11.2 7.4 539 MO Barton ......... 1109 13.1 19.7 540 MN Sherburne. . ..... 923 112 53.5 541 MO St Louis... ..... 924 45 48 540 MN Le Sueur. ....... 923 11.2 44.0 oH 541 MO St Louis City . . . .. 924 4.5 3.0 540 MN Mcleod ........ 1037 11.2 5.5 : 541 MO Jefferson. ....... 924 4.5 4.4 540 MN Sibley. ......... 1037 11.2 23.1 540 MN Wright 1083 11.2 205 541 MO Franklin ........ 924 4.5 9.0 OL emer : : 541 MO St. Francois. . . . . . 965 45 3.8 552 MN Olmsted . ....... 936 11.7 5.0 541 MO Washington . . . . .. 965 45 3.2 552 MN Fillmore ........ 936 11.7 27.0 541 MO lron........... 1072 4.5 12.5 552 MN Wabasha. . ...... 936 n.7 8.5 541 MO St. Genevieve. . . . . 1087 4.5 16.8 Sen MN “DOLIS + 5200 wns 936 hr 7.9 548 MO Jackson . ....... 931 9.1 6.5 552 MN Winona. ........ 1060 11.7 16.3 548 MO: Ca88 ... + wea sn 931 9.1 13.3 552 MN Mower ......... 1061 11.7 53 548 MO Johnson. ....... 931 9.1 19.8 552 A Howard. .. , «x «ws 1126 11.7 27.1 550 MN Steele 1156 117 14.2 548 MO Lafayette. ....... 931 9.1 17.7 REEL Wee ’ : 548 MO Clay. «oma vw saan 959 9.1 6.3 573 MN Blue Earth. . . .. .. 957 29.1 22.3 548 MO Platte. ......... 959 9.1 21.6 573 MN Nicollet. . ....... 957 29.1 30.7 548 MO Ray ....::vv5:+ 959 9.1 3.7 573 MN Waseca ........ 1084 29.1 44.7 548 MO Clinton. ........ 959 9.1 21.5 573 MN Watonwan. ...... 1158 29.1 30.0 548 MO Daviess. ........ 959 9.1 41.5 582 MN Otter Tail. . ...... 1066 31.8 317 548 MO Coldwell» «us wie $59 81 39 582 MN ‘Grant. . «9s + 55% 1066 31.8 32.3 549 MO Greene. ........ 932 11.6 4.0 588 MN Steams. . . ...... 974 19.8 12.7 549 MO Christian. ....... 932 11.6 55 549 MO Webster . ....... 932 11.6 3.9 588 MN Benton... :«'v ss 974 19.8 16.4 588 MN Meeker 974 19.8 34.8 549 MO Poliv.wswsnwss 932 11.6 11.6 tn ’ : 549 MO Stone.......... 932 11.6 a1 592 MN Kandiyohi . ...... 979 16.4 14.9 549 MO Wight ......... 932 11.6 4.9 592 MN Chippewa . ...... 979 16.4 12.4 549 MO Dallas. ......... 932 11.6 8.1 592 MN Yellow Medicine . . . 979 16.4 20.2 549 MO: Douglas . . «« si. « 932 11.6 3.9 592 MN. SWilt .....v5:95 1148 16.4 15.4 549 MO Ozark. . .:.va:29¢ 932 11.6 59.8 592 MN Lac Qui Parle. . . .. 1166 16.4 20.2 549 MO Dade .......... 932 11.6 19.9 597 MN Beltrami ........ 987 184 16.5 i bo itn. sereens hi 31% 352 oy 4 f Bos ed a ; ; B97 UN. Clearwater 987 3 25.8 549 MO Barry .......... 1006 11.6 16.6 602 MN Brown ...cs..u. 994 34.7 25.6 549 MO Laclede. . ....: +5 1091 11.6 6.2 602 MN Redwood ....... 994 34.7 35.0 549 MO Texas. ......... 1141 11.6 19.2 lle. . ....... 4. 47.2 502 MN Reavis 1977 547 553 MO Boone ......... 937 17.8 15.1 603 MN: sani vc .ooxvims 995 22.0 19.0 553 MO Cooper. ........ 937 17.8 17:4 603 MN Kanabec. ....... 995 22.0 23.0 553 MO Howard. ........ 937 17.8 6.0 603 MN Mile Lacs . ...... "nn 22.0 23.7 553 MO Randolph ....... 1119 17.8 5.8 604 MN Pennington . . . . .. 996 30.8 16.2 553 MO Chariton . ....... 1119 17.8 33.2 604 MN Marshall . ....... 996 30.8 37.2 558 MO Macon ......... 1137 17.8 29.8 604 MN Red Lake ....... 996 30.8 58.5 553 MO Shelby ......... 1137 17.8 41.2 604 MN Ritson «co on v0 ww 1151 30.8 24.3 563 IL oN. «ov sams es 634 14.9 17.5 608 MN Douglas . ....... 1001 15.8 14.9 563 MO Cape Girardeau . . . 948 14.9 7.6 608 MN Pope .......... 1001 15.8 13.6 563 VD Sool. ses ssunan $40 145 72 608 MN - Stvens oo 1182 15.8 18.8 563 MO New Madrid. . . ... 948 14.9 276 563 MD Penty ....uw ums 948 14.9 8.9 609 MN Lyon Car ged 1003 31.6 34.1 563 MO Mississippi. . . . . . . 948 14.9 8.9 609 MN Lincoln. ........ 1003 31.6 27.3 563 IL Alexander . ...... 948 14.9 24.2 612 MN Crow Wing ...... 1007 25.7 16.1 568 MO Bolinger. .....s. Das 149 9.2 612 MN Case .......o ux 1007 25.7 43.1 563 IL Pulaski. ........ 948 14.9 42.5 612 MN Hubbard... ..... 1007 25.7 32.4 563 MO. Butler, + wu pnsmnie 990 149 155 612 MN Wadena ........ 1007 25.7 207 563 MO Stoddard wT 990 14.9 10.1 612 MN Aitkin... 1112 25.7 30.9 563 MO" BIDISY. vv vsvnrua 990 ii oe 563 MO Wayne ......... 990 14.9 20.9 619 MN Nobles......... 1016 28.3 29.4 563 MO Carter. ......... 990 14.9 33.5 619 MN Jackson ........ 1016 28.3 21.6 . 619 MN Cottonwood. . . . . . 1107 28.3 33.1 566 MO Adair .......... 951 234 12.7 619 MN Murray EE 1135 28.3 28.6 566 MO Sullivan. .....«:. 951 23.4 24.8 566 MO Putnam. ........ 951 23.4 27.1 626 MN. Matin, ...:on:0: 1025 30.0 27.9 566 MO Knox .......... 951 23.4 35.4 626 IA Emmet......... 1025 30.0 338 566 MO Schuyler. ....... 951 23.4 25.7 631 MN Roseau. ........ 1031 33.2 37.7 566 MO Scotland. ....... 1098 23.4 30.0 631 MN Lake Of The Woods . 1031 33.2 23.2 581 MO Cole. .......... 967 34.3 9.1 646 MN Freeborn. ....... 1053 27.5 26.1 581 MO Miller... ........ 967 34.3 43.3 646 MN Faribault. ....... 1053 27.5 29.2 581 MO Crawford. ....... 967 34.3 99.7 - 581 MO Gasconade . . .... 967 34.3 43.2 539 MO Jasper ......... 922 13.1 7.5 581 MO Moniteau. . . . . ... 967 34.3 26.3 539 MO Newton. ........ 922 13.1 104 581 MO Osage ......... 967 34.3 © 107 539 KS Cherokee ....... 922 1341 16.6 581 MO Maries . .. ...... 967 34.3 32.4 539 MO McDonald . . ..... 922 13.1 41.2 60 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of— unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 581 MO Callaway. ....... 1134 34.3 35.8 422 MS Coahoma ....... 701 22.1 14.4 501 MO Buchanan. ...... 978 17.1 11.6 i i Lashes “ms 7 Sow i 591 MO Nodaway. ....... 978 17.1 133 | TY nea ’ : 591 MO Andrew. ........ 978 17.1 11.6 431 MS Warren. ........ 710 15.2 11.3 591 KS Doniphan ....... 978 17.1 34.2 431 MS Claiborne ....... 710 15.2 24.2 591 MO DeKalb......... 978 17.1 48.6 431 LA Madison ........ 812 15.2 19.1 ik MO GOY vvcny sais 978 171 Jus 435 MS Adams ......... 714 19.7 10.7 591 MO ‘Worth. . im: sas o 978 17.1 12.4 5 591 MO Harri 1145 17.1 342 435 LA Concordia. ...... 714 19.7 18.0 BUBO cs wos ott: 4 : 435 LA Tensas......... 714 19.7 30.9 599 MO St. Charles ...... 989 28.4 29.2 435 MS Franklin ........ 838 19.7 29.6 599 MO Lincoln. cues sons 989 28.4 20.4 435 MS Jefferson. ....... 839 19.7 30.7 599 MO ‘Warren. aces swe $29 284 73 455 MS Jones. ......... 736 25.3 19.7 607 MO Henry. ......... 1000 27.4 23.9 455 MS Jasper ......... 736 25.3 28.3 607 MO St. Clair ...: 0:5 1000 27.4 19.0 455 MS SID . vu cw smi au 736 25.3 81.9 607 MO Bates.......... 1159 27.4 36.4 455 MS Wayne ......... 837 25.3 24.2 627 MO Pulaski.......-:. 1027 38.2 61.9 456 MS Harrison .......: 737 13.4 8.7 627 MO Phelps ......... 1027 38.2 25.0 456 MS Hancock. ...::.. 737 13.4 22.5 627 MO Denb. ..ovvsnsss 1027 38.2 28.8 456 MS Stone.......... 843 13.4 35.0 627 MO Reynolds. ....... 1101 38.2 52.6 459 MS Lafayette. ....... 741 24.3 13.1 639 MO Vernon... «+: 1046 46.5 38.7 459 MS Calhoun ........ 741 24.3 24.6 639 MO Cedar. ......... 1046 46.5 55.2 459 MS Yalobusha....... 741 24.3 36.3 656 MO Petts.......... 1080 32.6 10.6 461 MS Lowndes. ....... 743 19.7 10.1 656 MO Morgan. ........ 1080 32.6 68.0 461 AL lamar ........- 743 19.7 43.3 656 MO Benton. ..:..... 1080 32.6 58.1 461 MS Noxubee. ....... 743 19.7 21.6 656 MO Saline. ......... 1081 32.6 26.6 461 MS. Clay... + « 6onin sas 889 19.7 15.7 657 MO - Carroll. ov0 mw 000 1102 36.4 40.6 471 MS Jackson .. «+... 755 16.9 13.6 657 MO Livingston ....... 1103 36.4 33.2 471 MS George. ........ 755 16.9 31.8 663 MO Madison. ....... 1175 41.5 41.5 477 MS Pike. ....s.u: 3 762 27.1 16.1 666 MO Audrain. ........ 1178 38.7 16.3 477 MS Walthall Bg Amamn din 762 27.1 21.4 477 MS Amite. ......... 762 27. 57.4 666 MO Montgomery. . . . .. 1178 38.7 59.0 477 MS Mari 846 271 a3 1 666 MO Monroe. . ....... 1178 38.7 72.6 BION 5 2 2s wi ie - - i 482 MS UNION. . ova 768 33.1 22.6 873 MO Lio. «= wmnwe wo 1135 $52 552 482 MS Pontotoc. . . . . ... 768 33.1 39.0 676 MO Taney.......... 1188 31.7 31.7 482 MS TiDpah .. . ws vss 5 856 33.1 39.5 678 MO Atchison. ....... 1190 40.2 27.8 484 MS Oktibbeha. ...... 775 28.7 26.3 678 MO Holt........... 1190 40.2 53.2 484 MS Webster ........ 775 28.7 26.1 683 MO Grundy. ........ 1195 415 35.4 484 MS Choctaw........ 775 28.7 30.4 683 MO Mercer... ...... 1195 415 64.5 4% VS WISER «vs nme 805 27 $29 487 MS Bolivar ......-.- 778 22.2 28.7 M MAB cos 2005s 1197 7. 47.2 id 0 Ca ze 487 MS Sunflower . . . . ... 778 22.2 18.5 409 MS Forrest. .... uns 688 12.4 6.2 487 MS Leflore ......... 796 22.2 14.5 409 MS Lamar. ......... 688 12.4 11.1 487 MS Humphreys . . .... 796 22.2 87.7 409 MS Perry .......... 688 12.4 11.7 ) 409 MS Greene. ........ 688 12.4 46.2 Sa us ry cee 7% fas $22 409 MS Covington. ...... 823 12.4 DFT bE ey Ne mmeishirs 5 : Ningian 488 MS Issaquena....... 780 16.7 52.2 411 MS Hinds... ciou40 690 8.1 6.0 411 MS Rankin ......... 690 8.1 4.2 498 MS Alcom ......... 793 29.6 20.8 411 MS Madison . . ...... 690 8.1 3.9 498 MS Tishomingo... 793 288 23 411 MS Sool :i:usvainsee 690 8.1 16.1 502 MS Grenada. ....... 799 30.5 245 411 MS Leake. ......... 814 8.1 17.9 502 MS Montgomery. . . . .. 799 30.5 277 411 MS Simpson........ 860 8.1 4.1 502 MS Caroll ... ...vx v0 799 30.5 570 411 MS Attala.......... 3 . . a pa, 7 2 ne 504 MS Jefferson Davis. . . . 801 45.2 55.9 Coe | ’ 504 MS Lawrence ....... 801 45.2 38.2 412 MS Lauderdale... 9 108 go 511 MS Copiah......... 817 36.9 43.4 412 MS Neshoba........ 691 10.6 15.0 511 MS Lincol 818 36.9 312 412 MS Newton. ........ 691 10.6 11.2 EOlY er ve: mrimn 5 : 412 AL Choctaw. ....... 691 10.6 18.6 525 MS Wilkinson ....... 908 375 37.5 412 A. Sumter... .«cs:5 691 10.6 28.2 412 MS Kemper. ........ 691 10.6 7.8 £83 MS Ya200. 40 ux meus 916 332 382 412 MS Clarke ......... 855 10.6 12.1 691 MT Rosebud. ....... 1203 6.4 23.2 418 WE EBs he a tm tn 697 23.2 9.0 691 MT BigHOm. «ovum» 1203 6.4 7.7 691 MT Treasure. ....... 1203 6.4 122 418 MS Monroe. . ....... 697 28.2 59.1 b 691 MT Yellowstone. . . . .. 1208 6.4 55 418 MS Prentiss ........ 697 28.2 11.4 691 MT Cabon...szvsms 1208 6.4 4.5 418 MS Itawamba ....... 697 23.2 11.9 691 MT Stillwater 1208 6.4 74 418 MS Chickasaw. . ..... 877 23.2 Eig [| ER EE Seiimeth ete mie : ’ 61 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of— unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 691 MT Golden Valley. . . . . 1208 6.4 12.0 170 NC Person......... 273 11.0 5.9 691 MT Musselshell . . .. .. 1298 6.4 35 170 NC Vange. ; usw ies. 417 11.0 15.8 692 MT Cascade. ....... 1204 65 44 170 NG Waren......... 9n7 110 35.9 692 MT Pondera........ 1204 6.5 5.3 186 NC Guilford. . . ...... 289 13.3 8.8 692 MT Teton.......... 1204 6.5 5.7 186 NC Randolph ....... 289 13.3 13.8 692 MT Chouteau ....... 1204 6.5 17.5 186 NC Rockingham. . . . .. 289 13.3 11.5 692 MT Judith Basin. . . . .. 1204 6.5 37.8 186 NC Davidson. ....... 444 13.3 26.6 o Tos. re tie on 2 7 192 NC Iredell. ......... 298 135 12.1 UAE a tee : : 192 NC Alexander ....... 298 13.5 18.5 712 ME Hh wuoncnanes 1225 25.4 20.5 194 NC Watauga. ....... 301 25.4 23.3 712 MT Blaine. ......... 1225 25.4 25.8 ) 194 TN Johnson. ....... 301 25.4 35.6 712 MT Liberty ......... 1225 25.4 29.7 ot NC Adhe 0% od 08 712 MT Phillips. ........ 1316 25.4 355 f|- "BY AL ca.rie. : 2 713 MT OWSSOME «meres os 1008 95 50 195 NC Pitt ........... 303 13.8 14.5 . 195 NC Beaufort ........ 303 13.8 9.2 713 MT Mineral. ........ 1226 9.3 55 : v 195 NC Martin. ......... 303 13.8 8.9 713 MT Granite. ........ 1226 9.3 31.5 : 195 NC Washington. . . ... 303 13.8 16.2 713 MT Lake .......... 1320 9.3 15.7 195 NC Hyde .......... 303 13.8 19.8 713 MT Sanders ........ 1321 9.3 15.2 165 NC Teel 200 128 56.7 713 MT Ravalli ......... 1326 9.3 5.7 YHOU. ce sweats : : 714 MT Custer ......... 1227 25.1 15.6 i> NO Woke ..qiusues 308 1a 122 198 NC Franklin +o» wo 50 + 308 14.4 23.5 714 MT Fallon. ......... 1227 25.1 32.6 ; 198 NC Johnston. ....... 350 14.4 19.7 714 MT Powder River . . . . . 1227 25.1 51.0 190 NC Wiser 250 kp ou 714 MT Carter. ......... 1227 25.1 237 | JS NE Shera 2 8 714 MT Praitie ......... 1227 25.1 32.7 203 NC Orange......... 313 28.2 37.6 714 MT Garfield. . ....... 1227 25.1 44.6 203 NC Chatham. ....... 313 28.2 18.8 721 MT Lewis And Clark . . . 1234 19.2 15.7 2% ne Po EEE REEL ee ee 20a 721 MT Jefferson. ....... 1234 19.2 538 | ~~~ WY EYP melee ’ : 721 MT Broadwater . . .... 1234 19.2 8.0 205 NC Moore ......... 315 17.0 15.3 728 MT Flathead . . ...... 1242 12.8 11.4 205 NG Bichmond .}.. eq Sus 170 152 708 MT Lincoln 1042 12.8 17.9 205 NC Montgomery. . . . .. 315 17.0 1558 AAA : : 205 NC HOKE. .o% anos os 315 17.0 43.1 Pe We Shanon rules sn Hi a 28 207 NC Robeson. ....... 317 28.6 23.3 BOIS 2 v5 dria le me " 6 207 NC Scotland. ....... 317 28.6 29.0 742 MT SilverBow. ...... 1257 12.6 10.7 207 NC Columbus. ...... 359 28.6 36.3 742 MT Deer Lodge . . . . .. 1257 12.6 18.4 207 NC Bladen ......... 359 28.6 26.6 742 MT Beaverhead. ..... 1344 12.6 10.0 209 NC Jackson . . ...... 319 20.1 19.5 756 MT Park. .......... 1273 29.8 26.4 209 NC Macon ......... 319 20.1 19.1 756 MT Sweet Grass . . . .. 1273 29.8 43.7 209 NC Swain. ......... 319 20.1 16.7 787 ME BBWS wo Fe oes s 1087 270 30.2 209 NC Graham ........ 319 20.1 32.7 767 MT McCone . ....... 1287 27.0 29.5 214 NC Nash .......... 324 215 27.8 767 MT Wibaux. ........ 1287 27.0 427 214 NC Edgecombe. . . . .. 324 215 17.4 767 MT Richland. ....... 1347 27.0 21.8 214 NC Halifax ......... 324 21.5 13.8 773 MT Gallatin. ........ 1305 19.7 15.1 214 NG" Norhanptoni«.. + «+ 324 25 50 773 MT Madison . ....... 1306 19.7 35.4 218 NC Gaston. ........ 331 19.1 23.2 775 MT Meagher... .. _ 454 36.0 218 NC Lincoln. ........ 331 19.1 34.2 778 MT Wheatland 1331 43.4 51.7 218 NC Cleveland . ...... 368 19.1 10.1 Je k 218 NC Rutherford. . ..... 368 19.1 14.2 7 MN Bonsausic nEwe «i iii ee a 225 NC Buncombe. . . . ... 340 7.0 5.3 oeriv mins sins ’ ’ 225 NC Henderson ...... 340 7.0 9.1 788 MT Powell «.u:uwims 1389 48.3 48.3 225 NC Transylvania. . . . .. 340 7.0 9.5 810 MT Fergus. ........ 1411 26.2 24.6 2s Ny Masison wins 4 mie 49 ae 7; 810 MT Petroleum . . . . . .. 1411 26.2 64.3 5 RYN: co seis % 149 NC Forsyth. ........ 251 71 6.6 229 NC Caiawba........ 343 ni 14.5 145 NG Sumy... 251 7.1 a3 229 NC Burke.......... 348 11.7 12.0 bi NC SIOhSE aoe ot 71 HE 229 NC Caldwell . ....... 348 11.7 8.1 149 NC Yadkin ......... 251 74 6.9 235 NC Rowan ......... © 362 16.1 16.3 149 NC Davie.......... 251 7.1 18.2 235 NC Cabarrus. ....... 362 16.1 14.8 167 NC Mecklenburg . . . . . 270 8.7 87 235 NC Stanly.......... 423 184 34 167 NG: UPON. +c 55x 0n on 270 8.7 5.7 238 NO ESRSiE.... crsnnms 365 22.6 11.7 167 NC Anson ......... 270 8.7 18.7 238 NC Duplin ......... 365 22.6 45.8 168 NC New Hanover. . ... 271 11.0 73 = N ss = Sw 22 2s CL 168 NC Brunswick . .. .... 271 11.0 19.8 BYNS: covey iow i 2 : 3 168 NC Pender......... 271 11.0 9.2 240 NC Yancey......... 369 28.7 45.0 170 NC Durham ........ 273 11.0 6.8 2% NS A at oe ey io 170 NC Granville. ....... 273 11.0 11.2 HEARN: = vis ia 400 - : 62 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked ——————————————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 242 NC Craven... suv as 373 15.4 12.1 638 ND Rolette . ........ 1045 29.3 28.9 242 NC Carteret ........ 373 15.4 10.6 638 ND Towner. .s....:s 1045 29.3 32.7 242 NC Pamlico .. v0. c5 ; 373 15.4 11.3 638 ND Pierce. ......... 1146 29.3 28.0 24 ie fons owe mn ins gs 154 az 655 ND Eddy .......... 1075 38.4 35.7 BOW, vss 2 mxims + ’ : 655 ND Ramsey ........ 1076 38.4 26.2 243 NC Chowan ........ 374 30.2 22.3 655 ND Benson......... 1076 38.4 62.8 243 NC Hertford ....o00:. 439 30.2 26.0 3 243 NG Bettie. . 439 30.2 27.9 670 ND Grant... sews es 1182 31.7 31.7 243 NC Gates. ...wvswe sn 439 30.2 64.0 680 ND Dickey ......... 1192 45.8 26.7 256 NC Wilkes... ...... 455 303 303 680 ND LaMoure .....:= 1192 45.8 75.1 258: NC Cherokee . ...... 457 46.4 39.6 2 rs da nEmy anes AL Ss pes 258 NC Clay. .......... 457 46.4 e37 | ~~ 0 TET : 542 NE Douglas . ....... 925 8.3 4.5 261 NC McDowell . ...... 460 26.5 26.5 542 NE Sampy.......... 925 8.3 4.4 262 NC Cumberland. . . . .. 461 19.2 14.2 542 NE Cass ..:vvsosus 925 8.3 31.8 262 NC - Sampson ..... 1417 19.2 28.6 542 NE Washington... ... 925 8.3 3.8 264 NC Pasquotank . ..... 463 327 17.5 S42 NE Bult,uuvvuvanas $25 83 12. 542 NE Dodge ......... 1014 8.3 49 264 NC Dare ...o vies 463 32.7 56.7 ; 542 NE Saunders ....... 1014 8.3 38.0 264 NC Perquimans. ..... 463 32.7 44.4 ; 264 NC Camden 463 30.7 24.5 542 NE Cuming......... 1115 8.3 28.0 Troy ’ 542 NE Colla®. .... es: 1172 8.3 18.4 267 NC Alamance ....... 466 26.4 26.4 555 NE Hall ooo. 939 15.3 11.8 543 ND Burleigh ........ 926 10.7 6.5 555 NE Merrick. :.v usu 939 15.3 21.0 543 ND Morton. ........ 926 10.7 8.0 555 NE Howard. ........ 939 15.3 12.5 543 ND Mercer ...«+wsms 926 10.7 57 555 NE ‘Gresley. . ..ccvs 4 939 15.3 39.6 543 ND Sioux.sso 50560 926 10.7 14.9 555 NE. . Hamilton « «+055 1120 15.3 15.0 Bs ND: WOE. wosrers gis 926 lay 19.4 559 NE Holt........... 944 30.7 316 543 ND: Oliver. . coum sas 926 10.7 7.0 559 NE Brown ....:.::5 944 30.7 26.9 543 ND Mclean ........ 1095 10.7 22.6 559 NE Rock .......... 944 30.7 22.7 543 ND Sheridan. ....... 1095 10.7 20.8 559 NE K Pah 944 307 477 543 ND Emmons. ....... 1139 10.7 8.4 ya Paha. xx mains : 547 ND Cass .......... 930 9.1 6.6 561 NE Lancaster ....... 946 8.0 7.5 561 NE Seward. .....: qs 946 8.0 7.0 547 MN Clay... ives es 930 9.1 4.0 a. 561 NE Saline. ......... 946 8.0 6.3 547 MN Norman ........ 930 9.1 10.3 561 NE Gage ....u:umss 1026 8.0 6.7 547 ND Ransom ........ 930 9.1 4.3 561 NE Jefferson. ....... 1026 8.0 10.2 547 ND Sargent. ........ 930 9.1 19.9 561 NE Buti 1133 8.0 16.7 547 ND Steele. ......... 930 9.1 19.0 MDOF: x vx or ele : 547 MN Becker......... 1017 9.1 12.9 564 NE Adams ......... 949 16.7 13.2 547 MN Mahnomen . ..... 1017 9.1 10.6 564 NE Clay. ...vovuwvns 949 16.7 36.5 547 ND Bames......... 1100 9.1 6.2 564 NE Nuckolls ........ 949 16.7 12.6 547 ND Griggs . «vu: wnss 1147 9.1 20.2 564 NE Webster ........ 949 16.7 12.0 $47 NG HO «ones nnn nw 3 7 570 NE Buffalo ......... 955 19.3 14.1 547 ND Richland ....,.: 1189 9.1 52 K 547 MN Wilkin 1189 9.1 20.8 570 NE BAMBY «+. views 955 19.3 27.2 Troe : : 570 NE Sherman........ 955 19.3 40.2 550 ND Ward .......... 933 14.3 73 570 NE Dawson ........ 1132 19.3 9.3 550 ND Bottineau ....... 933 14.3 11.9 570 NE.“ Frontier... « «ss + uy 1132 19.3 62.2 550 ND Mountrail. . ...... 933 14.3 19.9 570 NE Gosper......... 1132 19.3 23.3 550 ND McHenry. ....... 933 14.3 28.1 570 NE Franklin ......«» 1165 19.3 24.0 5% lL —— re ne J 577 NE Madison . ....... 962 19.4 20.2 Bs waar nr : : 577 NE Antelope. ....... 962 19.4 17.7 579 ND Wiliams ........ 964 17.4 15.0 577 NE PiBICO. . +» vu wuss 962 19.4 18.5 579 ND McKenzie ....... 964 17.4 25.1 577 NE Stanton. ........ 962 19.4 17.8 579 ND Divide. . vy um sms 964 17.4 18.0 877 NE Wayne ......... 1157 19.4 20.7 584 ND Grand Forks. . . . .. 970 14.9 12.0 580 NE Scott Bluff. ...... 966 14.2 87 584 ND Walsh. ...o «5 970 14.9 10.5 580 NE BoxButte ....... 966 14.2 14.8 584 ND Nelson......... 970 14.9 27.9 580 NE Momil. ... e010. 966 14.2 8.7 584 ND Pembina. ......: 984 14.9 9.4 580 NE Garden. ......«: 966 14.2 14.7 584 ND Cavalier ........ 984 14.9 18.0 580 NE Banner......... 966 14.2 11a 584 MN POI. «. « «i stvmnle ans 1116 14.9 17.8 580 NE Grant.......... 966 14.2 28.3 605 ND Stark . ......... 997 19.2 15.4 580 MN Morrison. ....... 1067 19.8 20.2 580 MN Todd .......... 1068 19.8 39.2 605 ND Dunn.......... 997 19.2 27.0 580 WY Goshen. ........ 1300 14.2 17.9 605 ND Golden Valley. . . . . 997 19.2 30.0 580 WY Niob 1301 142 418 605 ND Billings. ........ 997 19.2 50.0 WODIND : vs vwitws : / 632 ND Stutsman. . . . . ... 1033 30.7 25.6 593 NE Lincoln 0B oie #6 8 ew 980 21.0 18.4 593 NE Keith ....sconmns 980 21.0 33.0 632 ND Foster ......... 1033 30.7 27.9 + 632 ND Wells . 1160 30.7 42.6 593 NE Perkins. ........ 980 g 21.0 12.9 593 NE Hooker......uxs 980 21.0 32.6 63 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 593 NE McPherson ...... 980 21.0 33.3 93 NJ Warren. ........ 113 28.0 27.3 593 NE AUR: vo nenmsie 330 210 167 108 NJ Monmouth. . . . ... 133 14.9 13.4 594 NE valley. ......... 983 34.6 43.4 108 NJ Ocean ......... 133 14.9 16.1 594 NE Garfield. «. cv «vos 983 34.6 31.8 594 NE Wheeler . 983 346 50.5 126 NJ Mercer......... 228 10.6 10.6 594 NE Custer.......«= 1092 34.6 30.4 127 NJ Cumberland. . .... 229 17.6 17.6 594 NE Logan ..:xeu0 we 1092 34.6 61.8 508 NM Lea ........... 806 29.2 238 594 NE Thomas ........ 1092 34.6 57.9 508 TX Andrews. ....... 806 29.2 22.9 594 NE LOUD covnnvsnns 1082 840 180 508 TX Gaines......... 806 29.2 55.7 594 NE Blaine. ......... 1092 34.6 54.8 lilo. ....... s ; on NE Yow on wo me | SEM Seme 2 - 4 3 sy NE Polew.voesnsnsn 005 9 34.5 693 NM Sandoval... ..... 1205 6.8 14.1 613 KS Decatur. ....... 1008 36.3 25.1 693 NM S0COI0 .o.ounms 1205 6.8 5.0 613 KS Sheridan. ....... 1008 36.3 42.0 693 NM Torrance. ....... 1205 6.8 8.4 613 Ne Red WHIOW ov 502 1161 36.3 34.0 704 NM SantaFe........ 1237 12.1 12.4 613 N Hitchcock . .... .. 1161 36.3 57.9 724 NM Rio Amba . ...... 1237 12.1 9.0 616 NE Phelps .....«e:5 1012 22.0 12.4 724 NM Los Alamos . . .. .. 1237 124 16.7 616 NE Furnas ......... 1012 122.0 28.3 724 NM T8O8 .uswmewsis 1353 12.4 13.6 616 NE Harlan ... esx 1012 22.0 35.0 705 NM Curmy.......... 1038 20.5 16.1 628 NE Richardson ...... 1028 32.2 31.2 725 NM Roosevelt . ...... 1238 20.5 22.8 628 NE Pawnee ........ 1028 32.2 343 725 NM DeBaca........ 1238 20.5 24.6 651 NE Boone ......... 1062 36.4 47.7 725 NM Quay. ..osvvans 1380 20.5 27.6 651 NE Platte, .. cu:is9 25 1063 36.4 32.1 732 NM DonaAna ....... 1246 19.3 19.9 651 NE Nance ......... 1096 36.4 36.5 732 NM Lung oo we vmama 1246 19.3 15.8 654 NE Chase ......... 1073 34.4 24.7 732 NM Sierra. ......... 1307 19.3 21.2 654 NE Hayes. ... veri «cs 1073 34.4 63.9 740 AZ Navajo . vx evi 1255 24.6 49.3 654 NE Dundy .....:q:4 1074 34.4 43.9 740 UT SandJduan ....... 1255 24.6 23.0 658 NE Fillmore ........ 1105 40.8 33.4 740 C0 LaPiad vuvunns I= 240 ns 658 NE Th 1106 408 486 740 CO Archuleta. ....... 1292 24.6 21.9 YB #452 tir mee : 740 CO Sanduan ....... 1292 246 a7 660 NE Johnson........ 1113 39.6 32.4 740 CO Montezuma . . . ... 1293 24.6 6.5 660 NE Otoe .......... 1114 39.6 42.6 740 CO Dolores. . ....... 1293 24.6 25.3 668 NE Cherry ......... 1180 59.0 59.0 740 NM SandJuan ....... 1354 24.6 13.6 765 NM McKinley. ....... 1285 21.7 13.8 687 NE Nemaha ........ 1199 32.2 32.2 765 AZ Apache. ........ 1285 21.7 30.0 i BH SEEN. cveannas » 158 2s 769 NM Otero. ......... 1289 19.8 15.3 5 VI WIndSOR aus or oo w 19.8 ne 769 NM Lincoln 1289 19.8 31.3 15 YT Orange. .....«:ssmn 17 13.3 go | 7 Try ’ : 15 VT Washington... ... 190 13.3 114 772 NM Chaves. ........ 1303 8.6 75 15 NH Sullivan. .......s 215 13.3 13.8 772 NM Eddy . wun 1304 8.6 9.7 90 NH Cheshire. ....... 105 25.4 23.0 793 NM Grant. ...n:imeos 1394 28.2 22.1 90 VT Windham ....... 105 25.4 29.6 793 NM Hidalgo. . ....... 1394 28.2 44.9 91 NH Merrimack . . . . . .. 108 14.0 11.4 793 NM Catron ac sv vw ns 1394 28.2 59.5 91 NH Balknag .... + con 108 14.0 14.9 801 NM San Miguel . ..... 1402 26.3 25.5 91 NH Hillsborough . . . .. 169 14.0 11.6 801 NM Morag .ooossuens 1402 26.3 14.3 91 NH Carroll ......... 210 14.0 31.6 801 NM Guadalupe. . . .... 1402 26.3 28.3 138 NH CoOS . ......... 240 17.2 12.9 801 NM Harding ........ 1402 26.3 76.7 138 YT ES8BX. ...w: nu 240 17.2 49.8 814 NM Union. ......... 1415 23.6 23.6 23 NJ Camden........ 25 13.2 11.8 701 NV Washoe ........ 1214 11.8 11.2 23 NJ Gloucester. . . .... 25 13.2 123 701 NV Churchill... ..... 1214 11.8 9.0 23 NJ Salem. ....v..uss 25 13.2 14.8 701 NV Humboldt . ...... 1214 11.8 20.8 23 NJ Burlington... .... 184 13.2 15.5 701 NV Pershing... .:+:x 1214 11.8 6.7 36 NJ Bergen. ........ 39 14.0 13.0 701 NV Storey ......... 1214 11.8 8.8 i 701 NV Omsby «sowie 1265 11.8 1.2 36 NJ Passaic......... 39 14.0 20.3 36 NJ Hudson 208 14.0 10.7 701 NV Douglas ........ 1265 11.8 29.8 EE ’ ' 701 NY Lyon .ssmeww sme 1265 11.8 6.7 64 NJ Aflantic. ......:.s 71 18.9 18.3 701 CA Alpine. ......... 1265 11.8 43.7 64 NJ CapeMay....... 71 18.9 18.9 701 NV Mineral. . ....... 1332 11.8 2.6 66 NJ Middlesex ....... 73 11.5 18.0 707 NV Car. ou 2 vss on 1220 10.1 8.7 66 NJ Somerset ....... 73 11.5 22.7 707 UT Washington... . .. 1220 10.1 12.5 66 NJ ESSeX. . vou vk 55s 111 11.5 8.1 707 NV NY8...ausvnwms 1220 10.1 24.4 66 Nd UIA. sv mas wie 111 11.5 6.9 707 NV Linco. ..:cses + 1220 10.1 15.4 87 NJ Morris. «oo... .. a9 230 24.9 707 NV Esmeralda. ...... 1220 10.1 52.6 87 NJ Sussex 99 23.0 17.1 707 UT Kang ...samu 50s 1290 10.1 21.8 oo ’ 707 UT Garfield. . wv «uuu. 1290 10.1 27.9 93 NJ Hunterdon. ...... 113 28.0 29.1 707 UT Pie ius cuiws 1290 10.1 48.7 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 720 NV EKO: vonncnss 1233 29.8 27.7 88 NY Schenectady . . . .. 198 18.4 10.6 720 Ny rer sreameess po a 542 94 NY NewYork ....... 114 11.9 11.0 720 WIR cu sinnsin s : 94 NY BromX.......... 114 11.9 12.8 805 NV Waite Ping. «.. .. . 4065 32.0 35.0 105 NY Tompkins . ...... 128 12.2 13.1 10 NY Albany ......c.. 11 10.0 8.8 105 NY Cortland . ....... 128 12.2 11.0 10 NY Rensselaer ...... 11 10.0 9.4 . : 113 NY Kings. viv:ninss 166 19.4 21.8 10 VT Bennington ...... 141 10.0 20.9 113 NY Richmond . . ..... 167 19.4 8.0 19 PA Bradford. ....... 21 11.7 16.0 19 NY Tioga. ......... 21 117 11.6 133 NY Rockland. ....... 235 156.3 16.3 19 PA Sullivan. ........ 21 11.7 53.0 24 OH Belmont ........ 26 8.7 8.7 19 NY Broome ........ 84 14.7 53 24 WV OHIO +. samamnss 26 8.7 53 19 PA Susquehanna. . . .. 84 147 31.0 24 OH Montoe. ....«w 26 8.7 23.4 21 NY Ontario. ........ 23 16.8 14.6 £4 WY Wettel un snas a 45 2 20.7 24 WV Tyler .......... 46 8.7 22.0 21 NY Seneca......... 23 16.8 20.1 24 WV Marshall 142 87 55 21 NY Wayne ......u«- 151 16.8 i66 | ~~ TT rrr : ’ 21 NY Y88...:.00030 195 16.8 18.6 270 OH Hamilton. ....... 469 6.6 3.3 35 PA McKean . ....... a8 206 14.6 270 OH. Clermont. ....... 469 6.6 3.7 270 OH Warren. ........ 469 6.6 37.0 35 PA Polter.....veuus 38 20.6 19.4 270 OH Brown ..: swims» 469 6.6 15.3 35 PA Cameron. ....... 38 20.6 56.7 270 OH Adams ......... 593 6.6 21.5 35 NY Cattaraugus. . . . . . 89 20.6 24.7 35 NY Al 89 20.6 17.0 270 OH Buller. ....: 5:4 672 6.6 6.0 BGAN wiv 5 2 270 IN Franklin ........ 672 6.6 60.4 41 NY Westchester. . . . .. 44 14.2 12.7 276 OH TCHS. ov ee sees 474 7.1 3.1 41 NY Putnam. ........ 44 14.2 15.3 276 OH Wood... ..:: 04:0 474 7.3 10.5 41 NY Dutchess. , .. «wv +o 129 14.2 18.0 41 NY Ulst 129 14.2 17.2 276 OH Fulton. ......... 474 7.3 11.8 BRE cin BEE ! 4 276 OH Ottawa ......... 601 74 18.3 54 NY MOonwoe. ..: «xx « 60 4.7 52 276 MI Monroe. . ....... 636 7.1 15.6 34 NY LWngsion. ...... 50 47 128 281 OH Franklin ........ 479 6.0 49 54 NY EBB. .u: cnvwas 94 4.7 4.5 : 281 OH Pickaway. ....... 479 6.0 8.6 54 NY Genesee. ....... 94 4.7 3.7 ) 281 OH Madison. ....... 479 6.0 14.8 54 NY Wyoming. ....... 94 4.7 6.1 : 281 OH Fayette. ........ 600 6.0 9.6 sd NY Noga. crvosn tie 47 2 281 OH Delaware 654 6.0 1.2 54 NY Orleans. . ....... 112 4.7 34 | = 2M ARs ! ? 56 NY Onondaga...» 62 86 6.2 283 OH Jackson a re 481 16.0 22.4 283 OH Gala. ...s:05 4 481 16.0 10.1 56 NY Oswego ........ 62 8.6 13.9 56 NY Cavoas 187 8.6 11.9 283 WV Mason ......... 481 16.0 13.0 YUGR: i650 sink + ” g 283 OH Meigs. ......... 481 16.0 19.4 32 NY O80. onenese a 71 28 288 OH Cuyahoga... .... 48s 42 3.2 58 NY Delaware. ....... 64 17.7 13.9 : 288 OH Laks ..ivvnvups 485 4.2 4.1 58 NY Schoharie ....... 64 17.1 229 58 NY Chenango 181 17.1 315 265 Oil GORDA «ovum 425 42 85 troy ! : 288 OH Ashtabula . ...... 610 4.2 18.0 59 NY Oneida. ..ususe i 14 75 292 OH Allen .......... 489 12.7 75 59 NY Herkimer. ....... 65 11.4 10.9 , 59 NY Madison 65 11.4 30.1 292 OH Auglaize . ....... 489 12.7 12.7 PER gi SE Re ' ’ 292 OH Putnam. ........ 489 127 15.0 65 NY Sleubshy «o.oo 72 13.7 16.0 292 OH Hardin ....00 uns 489 12.7 20.1 65 NY Chemung ....... 72 13.7 10.7 292 OH Mercer. ........ 570 12.7 13.5 65 NY Schuyler........ 72 13.7 13.6 292 OH vanWett........ 570 12.7 20.2 76 NY FOR. ..c0 cman 86 19.0 13.3 295 OH Montgomery. . .. .. 492 10.5 5.7 76 NY Montgomery. . . . .. 86 19.0 23.3 295 OH Greene. ........ 492 10.5 13.8 80 NY Jefferson. . ...... 91 12.3 10.3 oe Sib D198 yumusmzes pi 39s poy 80 NY Lewis.......... 91 12.3 15.3 TR 70 Som mone : 80 NY St. Lawrence ..... 206 123 13.2 337 OH BOSS ...vssmwany 534 20.0 22.7 81 NY Clinton... ... 92 112 73 337 OH Pike ol Br enn ca 28 534 20.0 20.5 337 OH Scioto. +. .s: v0: 647 20.0 11.9 81 NY ES80X. sss vsass 92 11.2 17.0 337 KY Lewi 647 20.0 59.6 81 NY Franklin ........ 140 11.2 10.8 BYOB. x x aman 4 83 NY Nassau. ........ 95 14.0 9.0 345 PA Mercer oi ® el ag 199 7.8 8.8 345 OH Mahoning ....... 541 7.8 8.9 83 NY Suffolk ......... 95 14.0 6.9 245 OH Trumbull 541 78 58 83 NY Queens. ........ 177 14.0 21.6 PATON i; vom ve “3 ; 86 NY Orange. ........ 98 19.2 145 be MH Qari £9LE EE REE Se buy ne 86 NY Sullivan. ........ 98 19.2 24.7 BIPBGR «imi y 86 PA PKB. i :ioowmewsn 98 19.2 49.1 347 OH Stark os vusmsnn 543 135 9.1 88 NY Saratoga... ..... 100 18.4 32.3 347 OH TUSCOIWES . ou vv + 543 B35 88 : 347 OH Camoll .....o 25 254 543 135 18.7 88 NY Washington. . .... 100 18.4 17.0 88 NY Warren. ........ 100 18.4 9.3 S47 Of WORE. 10 cine vo S54 135 26 88 NY Hamilton. . . . .. .. 100 18.4 68.1 347 OH Holmes... «onus 564 13.5 174 352 OH Summit. . ...... 547 14.5 10.9 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 352 OH Portage. ........ 547 14.5 22.8 440 OK Beckham. ....... 720 23.5 22.6 352 OH Medingd.. ,.. cuss 624 14.5 28.6 440 OK Roger Mills . ..... 720 23.5 26.7 354 OH Marion ......... 549 18.2 9.4 445 OK Muskogee. ...... 726 22.0 14.3 354 OH Wyandot. ....... 549 18.2 30.0 445 OK Wagoner. ......x 726 22.0 32.8 354 OH Crawford. ....... 606 18.2 15.9 445 OK Mcintosh. ....... 726 22.0 28.7 354 OH Morrow. . ....... 607 18.2 23.9 445 OK Cherokee ....... 774 22.0 17.7 354 OH Union... ...ss +5 630 18.2 24.8 445 OK Adair... ws :5+5 774 22.0 33.8 356 OH Defiance. ....... 552 27.7 225 451 OK Washington. . . ... 732 17.7 16.6 356 OH Henry. ....-us4s 552 27.7 42.9 451 KS Montgomery. . . ... 732 17.7 15.6 356 OH Paulding. ....... 635 27.7 35.6 451 OK Nowata......... 732 17.7 14.7 356 OH Wiliams ........ 675 27.7 18.7 451 KS Chautauqua. . . . .. 732 17.7 15.8 360 OH Jefferson. . ...... 556 34.1 30.8 451 8 Bin onven anion 782 ii $52 360 OH Harrison. ....... 556 34.1 51.8 466 OK Pitsburg. « + +a + 749 24.7 23.9 366 OH Efie........... 566 15.4 7.3 a hater, get 2a 7 27 22 366 OH Huron. ......... 566 15.4 12.8 USIDAIAAR. + vv + > 366 OH Sandusky ....... 611 15.4 27.1 469 OK Woodward. . ..... 752 29.4 25.0 366 OH Loraif, ..con03555 669 15.4 16.3 469 OK Big ..omms:nas 752 29.4 15.0 368 OH Fairfield. ........ 569 31.1 30.6 449 OI HUET. 43 six ves 752 204 252 ; 469 TX Lipscomb ...«.: 752 29.4 50.8 368 OH Hocking .::v:.4: 569 31.1 19.3 469 OK Dew 864 20.4 392 368 OH Athens ......... 670 31.1 31.9 OWEY iez@me fem : 368 OH VION . 0:00:00 670 31.1 65.4 472 OK Payne. ..:..:u:5 756 28.3 20.4 369 OH Richland . ....... 572 19.0 19.9 $2 ox Painss wham a = 2s 369 OH Ashland ........ 572 19.0 16.6 DIG + wre ims : , ar2 OH Clinton. ........ 576 28.1 22.2 474 Ok Cafehuunwsnvns 753 188 133 370 OH Highland 576 28.1 34.9 474 OK LoVe «wo sna h uu 758 18.8 16.4 Tos : : 474 OK Johnston. ....... 847 18.8 28.1 376 OH Miami. ......... 580 18.3 17.5 474 OK Marshall ........ 848 18.8 23.3 376 OH Shelby . :v55.:4. 580 18.3 20.5 475 OK Custer .......5. 760 39.7 29.3 377 OH Muskingum . . .... 581 15.9 58 475 OK Washita ........ 760 39.7 B57 377 OH Guernsey ....... 581 15.9 14.9 476 OK Grady. ......... 761 38.7 29 1 377 OH Poly ..covwasss 581 15.9 28.5 476 OK Cadd 761 38.7 47.7 377 OH Morgan. ........ 581 15.9 34.2 BGO «5.2 viv rae : ; 377 OF: Noble., ov iwsmes 581 15.9 31.3 478 OK PomotoC. . ws v5 sw 763 26.7 17.1 377 OH Coshocton. ...... 631 15.9 18.5 478 OK Coal.y cvvvis +255 763 26.7 18.0 380 OH Columbiana. . . . . . 584 25.9 203 ae x Saas He oug nen hi os 2 380 WV HancogK. :. su + ++ 584 25.9 16.8 hes: «vu vivwin it : 2 380 WV Brooke. ........ 584 25.9 64.6 483 OK Jackson ........ 770 18.2 16.2 386 OH Hancock. ....... 594 20.6 16.1 an on Joh errs ears £4 32 3 386 OH Seneca......... 595 20.6 238 | Tv 0 V HHEEReeeresee : ’ 388 OH Ys ose sn 612 26.3 23.6 493 OK Comanche RIE par 787 20.5 10.5 388 OH Licking 613 26.3 27.8 493 OK Kiowa, «vn 2545 787 20.5 28.0 BE li ’ ’ 493 OK Cotton ......... 787 20.5 39.3 398 OH. Logam. « « wu ew us 677 33.1 33.1 493 OK THMAN. cova vue 831 20.5 30.4 414 OK Tulsa. ...oxamss 693 8.4 55 497 OK Stephens ....... 792 31.1 27.0 414 OK Creek. ......... 693 8.4 9.0 497 OK Jefferson. ....... 792 3141 46.5 414 Of BOBI%. un smenns 09s 24 58 515 OK Kay........... 894 25.0 16.3 414 OK Mayes ......... 693 8.4 11.6 515 OK O 895 25.0 44.0 414 OK Ottawa ......... 782 8.4 28.1 BODE cmrwr ante : 414 OK Craig.......... 782 8.4 10.0 529 OK Okmulgee. ...... 912 34.5 30.3 47 OK Oklahoma . . . .... 696 73 5.6 529 OK Okfuskee ....... 912 34.5 47.1 417 OK Canadian ....... 696 7.3 8.4 535 OK Beaver......... 918 48.3 48.3 417 OI LIBERA, + v0 tewns 030 73 112 689 OR Multnomah . . .... 1201 4.6 3.9 417 OK Pottawatomie . . . . . 773 7.3 14.5 ; 417 OK Lincol 773 73 19.6 689 OR Washington. . . . .. 1201 4.6 53 BOM: rv vio ? : 689 OR Clackamas. . . .... 1201 46 5.7 429 OK Garfield. . ....... 708 16.5 8.4 689 WA Cowlitz. ........ 1253 4.6 53 429 OK Malor:. ++ v0 708 16.5 16.0 689 OR Columbia ....... 1253 4.6 3.5 429 OK Alfalfa. ......... 708 16.5 16.2 689 WA Wahkiakum . ..... 1253 4.6 6.6 429 OK Grant. ......... 708 16.5 25.9 689 OR Clatsop. ........ 1318 4.6 5.4 429 OK BRING. . ws vusve + 815 16.5 32.6 689 WA Clark . ...o055 45 1333 4.6 4.7 3 Oh Ker wom ae Io oy 705 OR Marion ......... 1218 23.0 23.7 OOS + wa ema lis } : 705 OR Polk........... 1218 23.0 13.0 430 OK Cleveland . ...... 709 26.2 23.5 705 OR Yamhill. ........ 1366 23.0 26.0 450 OK MOBI «5x 29 20 709 26.2 32 717 WA WallaWala . .. ... 1230 12.6 95 430 OK Garvin ......... 789 26.2 22.2 : 430 OK M 789 26.2 39.9 797 WA Columbia ....... 1230 12.6 13.8 HRY. av it $mon : 77 OR Umatilla ........ 1299 12.6 14.8 66 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked ————————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 77 OR Momow. ....... 1299 12.6 16.8 48 MD Frederick. . . ..... 216 13.0 13.7 719 OR Deschutes. ...... 1232 12.1 12.0 52 PA Venango........ 58 22.9 11.6 719 OR Crook. . «us sh wows 1232 12.1 125 52 PA Clarion. ....\..: 58 22.9 35.1 719 OR Jefferson. ....... 1232 12.3 11.4 52 PA Forest ... seme 58 22.9 62.2 72 x Vege: es on wos oy 2 £57 57 PA Cambria ........ 63 9.2 5.8 Lis ALAR AAR ER ; 57 PA Somerset ....... 63 9.2 13.1 727 OR LA... mame soem 1240 16.2 17.5 57 PA Bla. ..s:s05 5010 120 9.2 6.6 727 OR Benton.......:« 1240 16.2 12.8 57 PA Bedford ........ 120 9.2 24.0 729 OR Malheur ........ 1243 25.7 22.0 72 PA Franklin... » wow 80 9.6 9.1 729 ID Payoite. .:.. ux» 1243 25.7 25.4 72 MD Washington. . . . .. 80 9.6 9.2 729 ID Washington. . .... 1243 25.7 23.6 72 PA. Fulton. .. uv «x ve 80 9.6 16.8 729 ID Adams ......:.. 1358 25.7 48.8 78 PA LUZBIE « «vv smn 88 10.8 7.9 738 CA DelNorte ....... 1252 20.9 32.8 78 PA Columbia ....... 88 10.8 22.3 738 DR CUNY. :o.umves x 1252 20.9 32.2 78 PA Wyoming. ....... 159 10.8 25.9 738 OR Coos .......... 1313 20.9 14.1 84 PA Lehigh ......... 9% 9.1 46 743 OR Kamath ......:- 1258 12.8 11.5 84 PA Northampton . . . .. 96 9.1 5.0 743 OR Lake .......... 1258 12.8 26.2 84 PA. ‘Carbon. ... wv» « sss 175 9.1 30.2 748 OR Wasco ......... 1264 213 18.2 ou PA. MODEOB. + uvi0s 22 181 9.1 182 748 WA Klickitat. . ....... 1264 21.3 19.0 100 PA Washington . . . ... 122 25.0 21.1 748 WA Skamania . ...... 1264 213 37.8 100 PA Fayette. ........ 122 25.0 29.1 748 OR Sherman. ....... 1264 21.3 19.2 100 PA Greene. ........ 150 25.0 29.2 748 OBR Gillam : us vcvwn 1264 21.3 31.5 7 : 106 PA Erie........... 131 6.7 4.9 748 OR Hood River ...... 1341 21.3 22.6 106 PA WEB. oo on on 131 6.7 12.0 752 OR Jackson . » + s+ ss» 1269 11.2 6.8 106 NY Chautauqua. . . . .. 186 8.7 7.7 7% 08 iy #4 hn ee ne ad 110 PA Huntingdon . . . . . . 152 20.0 31.7 YOU ix vis nix ss ’ : 110 PA Mifflin... ...... 153 20.0 9.6 759 OR Union. ......... 1277 21.4 19.9 110 PA: JUNBIA ,.. s0:ms 0 153 20.0 272 759 OR Wallowa ........ 1277 21.4 22.2 759 OR Baker.......... 1387 21.4 223 We PA CIoWIOH, oe voir oe 219 24 22 782 OR Douglas ...... = 1359 9.4 13.2 125 PA. BRwrnnrrens ae 227 265 258 782 OR L8h8 ..» oo wits 1360 9.4 7.8 128 PA Tioga... .s caw 230 32.6 32.6 783 OR Lincoln. w.cunmwos 1363 35.0 38.4 129 PA Lawrence ....... 231 16.4 16.4 783 OR Tillamook . ...... 1364 35.0 30.5 139 PA Berks. ......... 241 13.7 137 805 OR Grant.» cvvwse 1407 28.8 28.8 140 PA Lancaster ....... 242 7.0 7.0 8 PA Schuylkill. . ...... 9 16.5 18.8 : 8 FA Northamberian « . . 9 165 10.0 20 Rl Providence . ..... 22 5.8 37 20 Rl Kent. ......co.. 22 5.8 3.9 8 PA Snyder......... 9 16.5 23.1 : 8 PA Uni 9 16.5 17.0 20 RI Bristol «ow swuwy 22 5.8 12.0 : fa i sass y bi ore 20 CT New london ..... 104 5.8 12.4 TROY cies Ei : 20 RI Washington . . . . . . 104 5.8 8.5 26 PA Clearfield. . . ... .. 29 16.1 14.7 160 SC Richland . . . ..... 262 8.0 7.4 26 PA Jefferson. ....... 29 16.1 17.4 : 26 PA C 139 16.1 17.2 160 SC Lexington ....... 262 8.0 4.6 os EER : : 160 SC Fairfield. . . ...... 262 8.0 11.5 28 PA Philadelphia. . . . . . 3 6.0 59 160 SC Newberry ....... 388 8.0 17.4 28 PA Monlgomatye.s» «= 31 80 55 176 SC Orangeburg. . . . . . 279 21.9 19.6 28 PA Bucks... ..ssxvs 3 6.0 9.5 176 SC Bamberg. ....... 279 21.9 25.8 28 PA Danae bo ¢ ti faa 84 3.5 176 SC Calhoun ........ 279 21.9 22.4 2 DERG ms wie veins 200 8.0 20 176 SC Allendale. . . . . . .. 426 21.9 28.1 42 PA Allegheny... x: « 45 50 go 182 SC Greenville . . . . ... 285 59 5.6 42 PA Butler. ......... 45 5.6 18.1 182 SC Anderson ....... 285 5.9 5.6 42 PA Armstrong. ...... 45 5.6 5.4 : 182 SC Pickens. .....,.. 285 59 4.3 42 PA Westmoreland . . . . 110 5.6 8.6 182 SC Oconee 285 5.9 10.7 42 PA Indiana. ........ 110 5.6 LZ J EE ! 42 PA Boaver...... sss 185 5.6 6.2 184 SC Florence. ....... 287 15.0 6.7 : 184 SC Darlington. ...... 287 15.0 5.8 i n Dap A 2 ol 184 SC Williamsburg . . . . . 287 15.0 34.1 bi oh Coll SHAND. ox 5s + a Se 184 SC Chesterfield. . . . . . 380 15.0 36.2 43 PA LeBaron 202 74 6.5 184 SC Marlboro. ....... 450 15.0 23.1 al PA Lycoming « «veo 48 6.6 6.0 187 SC Greenwood . . .. .. 290 21.3 9.1 24 PA Clint 48 66 83 187 SC Laurens ........ 290 21.3 23.7 MRO: #; WHE Jo we : 187 SC Abbeville. . ...... 290 21.3 25.4 47 PA Lackawanna. . .... 51 9.1 6.8 187 SC Saluda......... 290 21.3 52.9 47 TR 51 9.1 22.0 187 SC McCormick ...... 290 21.3 40.9 48 PA York. «voce ves 52 13.0 12.5 191 SC Spartanburg. . . . .. 296 9.8 8.5 48 PA Adams ......... 52 13.0 14.4 191 SC Cherokee ....... 296 9.8 10.4 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 191 NC PolR. ...cuims su 296 9.8 24.1 610 SD Brookings. ...... 1004 29.8 255 191 SC Union. «: «oss oa 399 9.8 6.9 610 SD Kingsbury . ...... 1004 29.8 38.0 196 SC Sumter......... 306 23.9 16.4 615 SD Davison ........ 1011 24.3 13.2 196 SC Clarendon. ...... 306 23.9 28.5 615 SD Charles Mix. ..... 1011 24.3 20.8 196 BC 188 .,..u:vvussas 306 23.9 50.8 615 SD Douglas ........ 1011 24.3 35.6 199 SC Marion ......... 309 19.8 21.4 B18 SD AUR vu snvems ton 245 Cig 199 SC Dillon 309 198 17.7 615 SD Hanson.......s. 1011 24.3 37.0 BEE ’ : 615 SD Sanborn ........ 1011 24.3 51.5 212 SC Charleston. . ..... 322 7 4.4 615 SD Brule ums visi 1150 24.3 16.3 212 SC Berkeley. ....... 322 7.7 6.7 615 SD Lyman ......... 1150 243 46.8 212 SC Dorchester. . ..... 322 77 8.7 615 SD Buffalo... vu .v vn 1150 24.3 29.7 i Se Sen one ii 2 ae 623 SD Shannon. ....... 1020 22.1 20.7 POM: sit» : ! 623 NE Sherdan....:.«. 1020 22.1 20.2 244 SC York. iar: onninas 375 24.7 24.1 623 SD Bennett. ........ 1020 22.1 26.8 244 SC Chester. ........ 375 24.7 26.4 623 NE Dawes «vu uw 1123 22.1 20.4 246 SC Homy.......... 378 20.1 18.6 623 NE SIOUX ...oovesns 1123 22.1 50.0 246 SC Georgetown. . . . .. 378 20.1 23.9 629 SD Hughes. :.....uss 1029 2558 22.2 249 8C Kershaw. ......-. 400 25.4 24.0 629 SD Haakon. ........ 1029 25.5 32.7 249 SC Lancaster . ...... 401 25.4 27.1 629 SD Stanley......... 1020 25.5 122 629 SD Ziebach ........ 1029 255 68.4 544 SD Minnehaha ...... 927 13.9 5.0 629 SD Sully ....u:.0ns 1029 25.5 35.3 544 SD Lincoln. ........ 927 13.9 9.7 629 SD Jones. ......... 1029 25.5 23.1 oe 13. % su A lyn Sor 5 20:2 640 SD Perkins. ........ 1047 24.5 20.1 544 SD Union. .....c un 927 13.9 57.6 640 ND Bowman........ 1047 24.5 14.4 544 SD TWINGE... ss 927 13.9 21.2 ; 640 ND Hettifngsr. « . .. . 1047 24.5 42.8 544 8D Moody .::«:0::a 927 13.9 18.7 640 ND Adams... ....... 1047 24.5 11.6 544 SD McCook ........ 927 13.9 32.8 7 ; 640 8D Harding ........ 1047 24.5 65.7 544 SD MNBL: ov csv ain is 927 13.9 39.4 640 ND SI 1047 24.5 341 544 SD Lake .......... 1125 13.9 7.5 BPR « wine wine : : 544 MN Rock ..u:vuviss 1155 13.9 6.9 648 SD Walworth, . ...... 1055 39.1 36.2 544 MN Pipestone . ...... 1169 13.9 11.4 648 SD Dewey ......... 1055 39.1 32.8 558 SD Pennington . . . . . . 943 13.0 9.9 648 SD COON. .s neue 1035 89.1 2.9 : 648 SD PoMer...... so 1055 39.1 26.5 558 SD Fall River. ....... 943 13.0 127 a8 sD GC bell 1055 39.1 59 558 SD Custer ......... 943 13.0 12.8 2 BOBO weep me : 3 558 SD Jackson ........ 943 13.0 60.7 141 TN Hamilton. ....... 243 57 50 558 SD Meade ......... 1032 13.0 14.1 141 GA Walker ......... 243 57 5.6 558 SD Bute. ....:s5.u 1032 13.0 16.2 141 GA Catoosa ........ 243 87 7.0 568 SD Codington. . . . . .. 953 27.2 16.4 141 TN Maden ...uvsvns 243 27 0.2 . 141 GA Dade ...vs sus 243 57 8.2 568 SD Hamlin. ........ 953 27.2 27.8 141 ™ Ss thi 243 57 211 568 SD Deuel.......... 953 27.2 422 BURGE: + fr on « 3 : 568 SD TlaK . .oomawn uw 953 27.2 39.6 146 TN Shelby... :vu:ue 248 5.3 2.7 568 SD Grant. ...e:nsin 1149 27.2 32.9 146 MS DeSoto ........ 248 53 3.3 572 SD Brown ......... 956 13.6 10.7 146 TN Tipton. ......... 248 53 25 146 MS Marshall . ....... 248 5.3 17.3 572 SD Edmunds ....... 956 13.6 8.4 146 TN Fayette. ........ 248 53 37 572 SD McPherson ...... 956 13.6 9.5 | 146 MS Benton... ...v.ss 248 53 58.1 572 SD Spink. . «cad sus 1069 13.6 16.4 146 MS Panola .... sms 807 5.3 25.7 B72 BD FAK. cain n sions 1097 13.6 17.4 146 MS Talo. ...ovouwins 807 5.3 2.9 572 SO Day.....sswvws 1162 13.6 19.3 146 MS Tuni 836 5.3 230 572 SD Marshall . ....... 1163 13.6 17.9 RIGA, 53 tom vin 2 2 : 578 SD Yankton ........ 963 28.3 11.0 347 IN OK cvene ening 29 29 43 147 TN Sevier... ci -vses 249 5.9 6.0 578 SD Clay. : vu:asnnss 963 28.3 26.8 147 TN Loudon. ........ 249 5.9 4.4 578 NE Cedar.......... 963 28.3 62.6 : 147 TN UnION,: so vo sie ins 249 5.9 18.5 578 NE Knox ...:vs0533 963 28.3 36.3 147 IN McMi 353 5.9 14.2 578 SD Hutchinson . . . ... 1022 28.3 31.3 bpd - A mpl sity Se pp oo 578 SD Bon Homme ..... 1022 28.3 12.2 147 IN Meigs. ..... .... 353 59 27.2 590 SO Roberts. .......: 976 30.2 38.5 147 TN Blount ......... 389 5.9 3.2 ig Stone . . . .... 97 30.2 23.7 " > Ms 5g Sine 59 » : es 148 TN Davidson. . . . . ... 250 5.2 41 Co ’ ’ 148 TN Williamson. . ..... 250 5.2 45 595 8D Todd ....ovswoe 985 273 24.6 148 TN Robertson. ...... 250 52 7.4 595 SD Tripp .......... 985 27.7 20.3 148 TN Cheatham... .... 250 5.2 9.2 595 SD Mellette. . . ...... 985 27.7 48.1 148 TN Dickson ........ 304 5.2 4.9 595 SD Gregory ........ 1002 27.7 25.7 148 TN Hickmam........ 304 5.2 7.3 595 NE BOYH «5 2igoomnmn 1002 27.7 40.3 148 TN Humphreys . . .... 304 5.2 18.8 600 SD Beadle......... 991 243 20.3 151 VA RUSE. «55 «5% ws 174 10.7 20.0 600 SD Jeraud......... 991 24.3 16.5 151 VA Smyth ......... 204 10.7 7.6 600 8D Hand. ...,u:us uu 1143 24.3 26.2 151 TN Sullivan. . x. «ovis 253 10.7 8.4 600 SD Hyde .......... 1143 24.3 50.7 68 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked ——————————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 151 VA Washington. . . ... 253 10.7 5.7 248 TN DYsr. ....ove00 0 393 39.3 33.2 151 TN Hawkins ........ 253 10.7 21.8 248 IN Lake sv ovsvnwn 393 39.3 57.7 151 YA Scolt ...r sve sno 253 10.7 59 248 TN Lauderdale . ..... 394 39.3 41.7 169 TN Washington. . . . .. 272 12.0 14.1 252 TN Cocke ....: cms 430 20.1 35.2 169 TN Canter... ovo ous 272 12.0 12.8 252 TN Greene. ........ 431 20.1 11.5 189 WN UROL soon 272 120 27 260 TN Bedford ........ 459 31.2 31.2 173 TN Bradisy. . co cv 276 20.3 13.6 173 GA Fannin ......... 276 20.3 27.8 “68 IN Soo ..ros aso 62 25 id 173 TN POR: wwawymwews 276 20.3 24.0 404 TX -BOWIB. . wx wns ni 2 683 12.5 7.8 173 GA Gilmer ......... 446 20.3 31.7 404 AR Miller. ......... 683 12.5 6.6 181 IN Obion. . ...... .. 284 26.0 217 404 AR Little River. ...... 683 12.5 8.0 404 AR Lafayette. ....... 683 125 49.6 181 KY Fulton. ......... 284 26.0 28.0 181 IN Weak Ha 26.0 29.3 404 AR Hempstead . ..... 740 12.5 8.9 BEKEY + wir oe : : 404 AR Nevada......... 740 12.5 21.1 188 TN Madison. ....... 291 11.9 5.8 404 OK McCurtain. ...... 747 12.8 24.7 188 TN Gibson. ..:«: «uv. 291 11.9 9.2 404 AR Sevier. ......... 747 12.5 6.1 188 TN Henderson ...... 291 11.9 8.7 404 AR Howard. ........ 880 12.5 13.7 188 TN Crockett .... 4. 291 11.9 11.4 405 TX Potter. ......... 684 10.9 5.4 188 TN Chester... ...... 291 11.9 19.6 405 T™X Randal. ovo: oes 684 10.9 5.2 188 TN Decatur. : «oon vw. 291 11.9 10.3 : 405 TX Hutchinson . ..... 684 10.9 12.1 188 TN Hardeman. ...... 433 11.9 29.4 . 188 IN H ood 447 119 23.1 405 TX Swisher .:...w. + 684 10.9 29.6 BYWODD ts « + 51 3 : : 405 TX Carson. ........ 684 10.9 47.2 208 TN Mauty. . :: suv» + 318 225 17.4 405 TX Donley. ...:«c«xu 684 10.9 20.5 208 TN Marshall ........ 318 22.5 25.5 405 TX Oldham ........ 684 10.9 5.8 208 TN Lewis. .:svowusws 318 22.5 26.7 405 TX Armstrong. ...... 684 10.9 73 208 TN Giles .......... 434 225 27.6 405 TX Briscoe. .....++: 684 10.9 66.4 211 TN Rutherford. . . . . .. 321 25.2 30.5 403 TR Moo ..uimssns 717 12.9 7% 405 TX Dalam ....o69 4 717 10.9 13.3 211 TN Cannon ........ 321 252 15.6 405 TX Hartley......... 717 10.2 8.8 211 TN Warren. ....:... 357 25.2 20.7 405 IX Sherman. ....... 77 10.9 107 21 TN DaKally cii00004 357 25.2 24.4 211 TN VanBuren. . ..... 3s7 25.2 335 408 TX Hansiord.,..enws oH 109 75 405 TX Ochiltree. . ...... 842 10.9 16.6 215 TN PUAN 26: 0p ves 326 25.6 29.3 406 TR LubBook. «a ss 685 7.0 48 215 TN Overton ........ 326 25.6 14.4 : 406 TX Hockley ........ 685 7.0 6.8 215 TN Pickett . cc...0 001 326 25.6 24.8 406 TX Dawson ...::ssv 685 7.0 18.3 215 TN Jackson ........ 390 25.6 20.7 . 406 TX Bailey. ......... 685 7.0 19.6 215 I™N Clay... :vmevseo- 427 25.6 16.1 515 TN White . ) 451 25.6 39.1 406 TX yma coervvns 685 7.0 4.0 oe ’ : 406 TX OCHZA...: 5:04 685 7.0 8.9 217 TN Hamblen. ....... 328 34.3 26.6 406 TX Motley ......... 685 7.0 39.1 217 TN Jefferson. ....... 328 34.3 40.9 406 TX Borden......... 685 7.0 63.6 17 TN Grainger. ....:... 328 34.3 52.6 406 TX KING: iwsssswe sn 685 7.0 333 217 TN Hancock. ....... 425 34.3 20.4 406 TX Lamb. ..cicenany 833 7.0 11.0 223 TN Cumberland. . . . . . 337 18.5 15.3 408 TR Coo wes wn wn 25 74 4 406 TX Dickens ....:%: 4 861 7.0 11.4 223 TN Bledsoe ........ 337 18.5 23.1 406 TX Coch 223 TN Fentress. ....... 408 18.5 21.7 OTARA 14 4 +i + 885 7.0 27 226 TN Anderson ....... 341 26.8 28.4 408 TX HES cuvunenes oe7 87 4.8 408 TX FortBend ....... 687 8.7 20.3 226 TN ROANE aww sme 341 26.8 19.7 226 TN Morgan. ........ 341 26.8 39.5 408 TX Monigomery.....s gaz 87 59 408 TX Liberty ......... 687 6.7 17.3 226 TN RhBR ...vvcncws 416 26.8 40.6 408 TX Waller 687 67 141 226 TN Campbell ....... 440 26.8 21.9 408 ™ Austin 687 67 25.2 231 TN Sumner. .... asks 351 25.1 30.0 408 TX SandJacinto...... 687 6.7 30.7 231 TN Trousdale ....... 351 25.1 9.4 408 TX Chambers. ...... 882 8.7 28.9 231 WIHISON. ob nino 6 = 8 71 25. : 2 NN Wison 3 5.1 289 410 TX Bexar. ......... 689 6.2 38 231 TN Smith. ..con0us 371 25.1 15.7 410 TX Atascosa 689 en 2.2 TN Macon: icv: ons 25.1 24.8 | T.r 20 LEU TYRerweisdien > ™ acon 565 6 410 TX Medina. ........ 689 6.2 4.1 232 TN Wayne ......... 352 34.9 40.5 410 TX Wilson ......... 689 6.2 4.0 232 TN Perry.......... 352 34.9 43.1 410 TX Kendall. ........ 689 6.2 20.3 232 TN Lawrence ....... 443 34.9 31.2 410 TX Zavala ......... 689 6.2 58.2 410 TX Bandera........ 689 6.2 41.7 234 TN Montgomery. . . . .. 361 19.1 19.0 234 TN Stewart......... 361 19.1 19.3 410 TX McMullen ....... 689 6.2 25.0 234 TN Houston ........ 361 19.1 19.4 210 YX -Guadolung,...x +n 779 62 135 410 TX Gonzales. ....... 779 8.2 28.2 239 TN Coffee ......... 367 26.8 25.4 410 TX Frio. .......... 832 6.2 7.2 239 TN Franklin ...o: cen 367 265 175 410 TX LaSalle ........ 832 6.2 315 239 TN Grundy. ........ 367 26.5 45.2 239 IN Moore ......... 367 26.5 56.8 413 TX Jefferson... ...... 692 9.0 6.4 413 TX Orange. ...: ux. 692 9.0 7.9 245 TN McNairy . ....... 377 40.1 46.1 413 TX Hardin ......... 692 9.0 7.6 245 TN Hardin .....s 454 377 40.1 34.6 69 Table II. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 413 TX Jasper ......... 692 9.0 23.0 437 TX Brooks....c:0 +0 716 6.9 8.8 413 TX Newton......... 692 9.0 18.6 437 TX Kenedy. ........ 716 6.9 26.7 413 TX TF cnevwennss 358 20 14 438 TX Lamar. ......... 718 143 75 415 TX EIP2a80...s:w0 0x 694 4.5 4.2 438 OK Choctaw. ....... 718 14.3 20.9 415 TX Culberson. ...... 694 4.5 26.2 438 TX RedRiver ....... 718 14.3 17.7 415 TX Hudspeth ....... 694 4.5 23.3 438 TX Della ....cocomne 718 14.3 31.4 415 TX Loving «: 0:0 694 4.5 50.0 441 TX Smith. . ..... 701 18.9 8.6 420 TX WICHRE. . v5.05 699 12.1 7.0 441 TX VanZandt....... 721 18.9 47.5 420 TX Montague . ...... 699 124 20.2 441 TX Wood. ..u.we sus 721 18.9 17.6 420 TX CY... cows ws nn 699 121 8.0 441 TX Henderson ...... 819 18.9 22.0 420 TX ARPS uunanans 090 12 103 444 TX Ector .......... 725 16.3 135 420 TX RAO oc 46 sa wom 765 12.1 42.6 444 TX Reeves......... 725 16.3 28.7 420 TX Baylor .....cxvs 765 1241 7.6 444 TX Ward .:.uviwsvis 725 16.3 18.3 420 TX YOUNG ovo cnn ne 785 12.1 14.3 420 TX Throckmort 785 12.1 36.9 444 TX. Craft. nme nn 725 16.3 26.7 CORMONON; = « 2 » : : 444 TX Winkler. ........ 872 16.3 29.2 42s TX TE anne 704 193 32 aa7 TX Nacogdoches. . . . . 728 23.4 13.1 425 TX Williamson. ...... 704 10.3 123 447 TX Shelby. ... uv: 728 23.4 30.1 425 TX BaSHOP. vo vv vows 704 10.3 12.3 5 447 TX San Augustine . . . . 728 23.4 25.0 425 TX Bumet...:vs sus 704 10.3 16.7 447 TX Sabine 826 23.4 45.7 425 TX Hand .scsvemsss 704 10.3 237 { C0202 TEEEruremeaeee : ’ 425 TX Lee........... 884 10.3 32.8 449 TX Angelina........ 730 26.7 12.3 426 TX Tom Green ...... 705 11.7 7.0 49 x Pte CIEE 730 20.7 42.0 449 TX Tiny. vo nvens 730 26.7 30.8 426 TX McCulloch. ...... 705 11.7 16.2 449 TX Walker 825 26.7 36.7 426 TX Reagan......... 705 11.7 227 4 Tr Mh MEER ! ’ 426 TX Sutton ......... 705 11.7 51.2 452 TX BeHiuws iwews sans 733 12.8 6.9 426 TX Crockett .....:.. 705 11.7 17.6 452 TX Coryell... einen 733 12.8 20.6 426 TX Coke .......... 705 112 9.7 452 TX Milam. sve vins ows 733 12.8 17.3 426 TX Schleicher. ...... 705 11.7 4.6 452 TX Lampasas....... 859 12.8 27.9 426 TX Concho: w......:s 705 11.7 8.9 453 7X Dallas... ....... 734 10.5 9.4 426 TX Menard. ........ 705 117 16.5 + : 453 TX Collate ani 734 10.5 10.5 426 TX Won. ..:.isevme 705 11.7 4.7 4 453 TX Bliss ations ows 734 10.5 12.1 426 TX Runnels ........ 851 11.7 19.9 426 TX Sterli 896 117 15.5 453 TH Hunt os. .o. ens 734 10.5 8.3 BAG. + 25 sm ge : 4 453 TX Kaufman. ....... 734 10.5 18.7 427 TX Hidalgo. ........ 706 14.0 14.6 453 TX Rockwall. ....... 734 10.5 6.6 427 TX Slat «conn omes 706 14.0 6.6 453 TX Rains.......... 734 10.5 34.2 428 TX Taylor. ......... 707 15.7 9.1 453 TX HOPKINS 500s sun 834 10.5 20.7 428 TX Eastland... .::.= 707 15.7 29.6 454 TX Gregg «sons nme 735 23.7 16.3 428 TX Callahan. ....... 707 157 18.9 454 TX: Upshur. uiwesns 735 23.7 22.4 428 TX Stephens ....... 707 15.7 30.6 454 TX Rusk .......... 874 28.7 32.2 428 TX Shackelford . . . ... 707 15.7 8.8 454 TX Panola. ...us + 892 23.7 36.7 423 TY OMS. .esweunes 789 157 8.2 458 TX Wilbarger . . . .... 739 36.0 40.0 428 TX Haskell. .... «5s 769 15.7 15.9 458 TX Hardeman. ...... 739 36.0 26.9 428 TX Stonewall ....... 873 15.7 10.6 458 TX Foard 739 36.0 43.9 428 TX Kents ain bsisps ww 873 15.7 44.7 iii ii i a ’ A 49 TX Vlora... 72 105 50 | ie TX Wheeler... 7a 27 284 433 IX DoWilt..,u; uses 712 183 £5 460 TX Hemphill. ....... 742 27.9 55.5 433 TX Jackson ........ 712 10.5 14.1 460 TX Roberts 742 27.9 49.0 433 TX Goliad .... 0... 712 10.5 07 | = 20 Teemmeemeaaran ’ ' 433 TX Refuge. :«v:ev a0 852 10.5 29.4 462 TX McLennan. ...... 744 20.2 13.4 433 TX Llavaca......... 883 10.5 12.8 462 TX Limestone. ...... 744 20.2 21.1 433 TX Calhoun .....: = 897 10.5 11.3 462 TX Fag. vv cos wirwns 744 20.2 45.2 462 TX Bosgue.....- ... 772 20.2 222 434 TX Tamant......... 713 7.9 75 462 TX Hamilton. . . . .... 772 20.2 56.2 434 TX Johnson. ....... 713 7.9 8.4 462 TX Hill 850 20.2 18.7 434 TX. PAIKBE «iv vv ui 5 713 7.9 03 | 0 rrr ’ : 434 TX Hood.......... 713 7.9 12.4 464 TX Brazos ..... wes 746 20.2 10.0 434 TX Somervell ....... 713 7.9 9.0 464 TX Washington... ... 746 20.2 25.3 5 Rema 74 20. . 436 TX Grayson ........ 715 135 8.6 484 IX |[Doberen g op ue 464 TX Burleson. .:..:=: 746 20.2 19.5 436 OK Bryan. .....c..s 715 135 9.6 464 TX Grimes. ........ 871 20.2 33.3 436 TX Fannin coos 715 13.5 254 436 OK Atoka.......... 828 13.5 30.0 465 TX THUS sovwsehess 748 28.5 21.0 18: 5 deosiincn: ammmons 74 28. : 437 TX Nueces. ........ 716 6.9 49 i = MoT Sa y o 2s at 437 TX San Patricio. . .... 716 6.9 8.3 465 TX Camp 890 285 27.8 437 TX JimWells ....... 716 6.9 gg | TT 2 Th THFhmewrweaees ’ i 437 TX Kleberg. ........ 716 6.9 7.0 468 TX Brown ....weow. 751 209 14.5 437 TX Aransas ........ 716 6.9 12.9 468 TX Coleman. ....... 751 20.9 32.0 437 TX Duval. .e:wecmb 716 6.9 55 468 TX Mills. .......... 751 20.9 26.2 437 TX LiveOak........ 716 6.9 52.8 468 TX SanSaba ....... 827 20.9 26.8 70 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of— 800- 1400- area by residents of — unlinked UOHNKOE Simmmimimimsreimp me emtmmmtess unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 479 TX PaloPinto....... 764 32.4 27.7 524 TX Collingsworth . . . . . 907 49.1 49.1 479 ™X Jack.......... Lu 824 43.9 526 TX Brewster. ....... 909 44.6 39.0 481 TX Hales. ....xiusie 767 38.8 38.4 526 TX Presidio ........ 909 44.6 50.0 481 TX Floyd.......... 767 38.8 39.7 526 TX JeliDavis ....... 909 44.6 50.8 il TX Casho......... 878 328 $8.5 531 TX Dimmit......... 914 41.4 41.4 485 TX C88... conus ome 776 39.9 45.0 485 ™ Maron 776 39.9 34.9 532 TX Galveston ....... 915 28.0 28.0 485 TX Harrison ........ 840 39.9 36.2 536 TX Uvalde ......... 919 36.0 36.0 489 TX Navarro. ........ 781 22.0 18.3 538 TX Webb...o.. ous 921 13.5 9.2 489 TX Freestone . ...... 781 22.0 31.9 538 TX Zapata ......... 921 13.5 26.3 490 TX Fayette. ........ 783 39.3 48.3 356 TX SIHOgY nr ene 921 Bs 497 490 TX Coloiadn. ... «vis 783 39.3 28.7 703 UT Utah. ...nvvoves 1216 13.9 12.5 491 TX Deaf Smith. . . . . .. 784 427 31.0 703 uT Sanpete WIE Tw BE 1216 13.9 18.2 491 TX Parm 784 42.7 68.6 703 UT Sevier. ......... 1216 13.9 18.9 BIBL vv 0%. 5 8 * # 703 UT Juab .......... 1216 13.9 9.2 492 TX Midland ........ 786 21.3 14.2 703 UT Wayne ......... 1216 13.9 29.5 492 TX Martin. ......... 786 21.3 44.0 703 UT Millard ......... 1370 13.9 19.0 452 IY Uh eine ren 728 21.3 $7.4 708 UT Saltlake........ 1221 7.2 6.5 492 TX P2008: .i:oomnwan 891 213 36.4 492 ™~ T 1 891 213 725 708 UT Tooele ......... 1221 7.2 5.6 RBH tec see ene oe 4 ? 708 UT Summit. ........ 1221 7.2 27.4 495 TX Demon. ««uvenes 790 23.8 22.2 708 UT Uintah ......... 1259 7.2 9.6 495 TX Cooke . ovum wos 790 23.8 20.7 708 UT Duchesne ....... 1259 7.2 11.0 495 TX WISE ws vsumzmns 881 23.8 32.6 708 UT Wasatch. ....... 1325 7.2 17.5 501 TX BON uve vu vnnn 797 51.0 59.3 715 UT Caph® w:wovwons 1228 16.3 13.5 501 TX Madison ........ 797 51.0 38.0 715 ID Franklin ........ 1228 16.3 19.3 503 TX Nolan. ......... 800 34.1 30.6 715 ID Bearlake....... 1228 16.3 29.3 503 TX Mitchell. . ....... 800 34.1 35.7 744 UT Davis.......... 1260 13.8 21.0 503 TX Fisher. ......... 888 34.1 41.7 744 UT Weber ......... 1260 13.8 7.7 505 TX Matagorda. . . . . .. 803 34.8 24.3 744 UT Morgan. coven 1200 a8 33 744 UT Rich........... 1260 13.8 57.5 505 TX Wharton ........ 803 34.8 21.9 Te OT Box Elder bres 55 oi0 505 TX Brazoria ........ 822 34.8 42.7 0% BOR oo» ci 4 506 TX Kerr. ..oovoo... 804 25.8 19.5 Is 7 Son ARERR Tor 0% i) 506 TX Gillespie . ....... 804 25.8 18.0 BEVBE 1.» wd a2 5 8 A 506 7X Benco......... 804 25.8 64.6 809 UT Carbon......... 1410 22.7 20.3 506 TX Mason ......... 804 25.8 42.2 809 UT Emery ......... 1410 22.7 30.5 506 TX Beal. .ooasneues 804 25.8 58.3 5 VA Newnoit News 8 7.5 53 506 TX Edwards........ 804 25.8 45.1 ? 506 TX Kimble 857 25.8 21.5 : VA Hampton Chy. ».. 75 70 SEE aE : : 5 VA York. .......... 6 75 4.9 507 TX Erath.......... 805 29.1 31.3 5 VA Gloucester. . . .... 6 75 9.6 507 TX Comanche. ...... 805 29.1 26.4 5 VA Mathews. ....... 6 7.8 11.9 509 TX Anderson ....... 808 24.6 20.9 5 VA James Cily Ju... 214 75 130 509 TX HOUBION . + vv ov vs 808 24.6 81.1 6 VA Virginia Beach . . . . 7 8.4 5.0 509 TX Cherokee ....... 893 24.6 24.4 6 VA Norfolk/Portsmouth . 7 8.4 3.5 510 TX Hays .......... 809 37.3 423 2 VA Chesansake City. . 7 8.4 8.5 510 TX Caldwell . ....... 809 37.3 315 8 NG CUBR. «vs vv ss 7 8.4 52.5 6 VA Nansemond. ..... 135 8.4 5.4 512 TX Childress. . . ..... 829 45.5 35.4 6 VA Southampton. .... 135 8.4 12.4 512 TX Cottle. ......... 829 45.5 58.3 6 VA Isle of Wight. . . . .. 135 8.4 100.0 12 TR Hal.oousnnnnna ga0 45.5 53.1 14 VA Roanoke. ....... 16 6.2 41 513 TX Bee........... 869 47.4 41.1 14 VA Franklin ........ 16 6.2 6.8 513 TE Karnes uve wv wes 870 47.4 56.7 14 VA Botetourt. ....... 16 6.2 12.6 514 TX Tery .......... 875 49.9 49.9 " a ka Ce 3 gg 832 14 TX Yoakum .....o.. 7 49, 497 |] 13 YA MBB vrs#Ercase e . 5 Day 376 98 14 VA Campbell . ...... 82 6.2 48 516 TX Maverick. ....... 898 30.7 38.8 14 VA Bedford ........ 82 6.2 6.4 516 TX ValVerde ....... 899 30.7 22.0 14 VA Amherst ........ 82 6.2 6.1 516 TX Kinney ......... 899 30.7 44.7 14 VA Appomattox. . . . . . 82 6.2 19.3 517 TX. Comal wo :voswes 900 30.4 30.4 25 VA Frederick. . . ..... 28 13.5 6.3 518 TX Howard. ........ 901 20.1 19.1 25 VA Shenandoah ..... 28 13.5 19.9 518 TX Glasscock . . . .... 901 20.1 70.8 25 WV Hampshire. . ..... 28 13.5 36.4 25 VA Clarke ......... 28 13.5 5.3 519 TX Seurry ......... 902 a7 417 25 WV Berkeley. ....... 106 13.5 8.4 520 ™ Cameron. . ...... 903 9.8 9.5 25 WV Jefferson. ....... 106 13.5 13.3 520 > Willacy Ca 903 9.8 14.1 25 WV Morgan we ee 188 13.5 252 25 VA Warren. ........ 236 13.5 10.4 71 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 33 VA Henrico. ........ 36 10.5 3.4 135 VA Spotsylvania. . . . . . 237 19.8 11.1 33 VA Chesterfield. . .. .. 36 10.5 33.9 135 VA Stafford. ........ 237 19.8 22.0 33 VA Hanover ........ 36 10.5 4.5 135 VA Caroline ........ 237 19.8 49.0 33 VA Westmoreland . . . . 36 10.5 47.3 135 VA King George . .... 237 19.8 15.9 » » Rope: I > oe Tor 137 VA Accomack. ...... 239 24.0 28.3 a3 VA Goochland. . . . . .. 36 105 43 137 VA Northampton . . ... 239 24.0 15.0 33 VA NewKent ....... 36 10.5 24.8 49 VT Chittenden. . ..... 53 7.0 8.7 33 VA King William. . . . .. 36 10.5 35.1 49 VT Franklin ........ 53 7.0 8.5 33 VA EBSBX....ocuivs 36 10.5 9.0 49 VT Addison ........ B53 7.0 11.5 33 VA Amelia ......... 36 10.5 5.0 49 VT Grandlsle....... 53 7.0 9.6 3 VA Biohmond sss sews 50 105 5.6 103 VT Caledonia . . . . . .. 126 23.4 30.1 5 VA ChatesCRy...... 35 105 21.2 103 VT Lamoille . . ...... 126 23.4 26.4 % VA Wingend Queen... 56 jos 335 103 VT Orleans. ........ 192 23.4 14.7 33 VA Lancaster ....... 158 10.5 9.5 33 VA Northumberlan . . . . 158 10.5 72 122 VT Rutland. ........ 224 11.5 11.5 33 VA Middlesex . . ..... 158 10.5 46.8 698 WA Spokane. ....... 1211 6.2 5.2 33 VA Prince Edward . . . . 205 10.5 12.1 698 WA Lncon......... 1211 6.2 6.5 33 VA Charlotte. ....... 205 10.5 47.0 698 WA Pend Oreille. 1211 6.2 08 33 VA Cumberland. . . ... 205 10.5 10.0 698 WA Stevens . ....... 1974 6.0 6.2 55 VA Wise ..;.ouvins 61 26.4 20.5 698 WA Fefty ...... 00: 1274 6.2 8.9 55 VA Dickenson. ...... 61 26.4 31.9 698 WA Adams ......... 1382 6.2 24.7 55 VA lee........... 220 26.4 35.1 702 WA Benton. ........ 1215 16.6 16.6 63 VA Rockingham. . .... 70 16.8 10.9 702 WA Franklin ........ 1215 16.6 16.7 63 VA Page .......... 70 16.8 20.2 736 WA King. .......... 1250 5.6 5.2 63 WV Pendeton ....... 70 16.8 48.6 736 WA Snohomish . ..... 1250 56 47 69 VA Fairfax ......... 76 16.7 18.5 736 WA Skagit. «v0 vis 5 ars 1250 586 6.1 69 VA Arlington ....... 76 16.7 15.5 736 WA island. ......... 1250 5.6 8.3 69 VA Alexandria City . . . . 76 16.7 13.5 736 WA SanJuan ....... 1250 5.6 29.2 69 VA Loudoun. ....... 211 16.7 21.3 736 WA Kitsap. . ........ 1282 5.6 7.8 70 VA Montgomery. . .... 77 16.6 18.8 739 WA. 'YBKIMB. . : 05 04 54 1254 8.2 7.4 70 VA Giles .......... 77 16.6 20.6 739 WA Kithias ...... ov. 1254 8.2 15.8 70 VA Pulaski......... 130 16.6 9.1 747 WA Grant. ......... 1263 16.1 28.3 70 VA Wythe.......... 130 16.6 19.6 747 WA Chelan. ........ 1263 16.1 9.4 71 VA Halifax ......... 79 24.8 19.0 747 WA Douglas ........ 1263 16.1 10.1 71 VA Mecklenburg . . . .. 79 24.8 23.0 747 WA Okanogan. ...... 1324 16.1 15.4 7 VA Lunenolrg. «xx a3 526 758 WA Thurston... ..... 1276 17.9 13.4 73 VA Camoll ...ooivns 81 28.2 43.4 758 WA Lewis, ...o vues 1276 17.9 15.3 73 VA Grayson ........ 81 28.2 19.7 758 WA Mason ......... 1276 17.9 33.2 73 NC Alleghany ....... 81 28.2 25.0 758 WA. PacliC ...o0 000s 1319 17.9 33.2 77 VA Dinwiddie . . . . ... 87 258 18.2 758 WA Grays Harbor . . . . . 1352 17.9 13.6 77 VA Prince George . . . . 87 25.8 17.6 784 ID: LBA «vw bi 1371 34.2 25.1 77 VA Sussex......... 87 25.8 63.5 784 WA Whitman. ....... 1372 34.2 42.1 7” VA Sumy.......... 87 25.8 57.7 785 WA Clallam. . ....... 1374 22.1 17.7 79 VA Hemty......q a+ 90 21.2 19.0 785 WA Jefferson. ....... 1375 22.1 35.1 Ie VA Patrick ......... 90 212 28.1 794 WA Pierce. . ........ 1395 11.2 11.2 97 VA Auguym oman v7 55 148 815 WA Whatcom. . . .. ... 1416 9.2 9.2 97 VA Rockbridge . ... .. 117 16.8 21.9 97 VA" Highland. ....... 117 16.8 36.1 278 WI Brown ......... 476 7.7 52 278 WI Oconto. ........ 476 7.7 17.2 99 VA Albemarle ....... 121 21.9 6.3 278 WI Kewaunee. . ..... 476 7.7 87 99 VA Culpeper. ..,..: 121 21.9 17.3 278 WI Door .......... 609 77 8.9 99 VA Orange. ...«:+s 121 21.9 24.2 99 VA Llouisa ......... 121 21.9 39.0 280 WI Milwaukee. . ..... 478 3.8 32 99 VA Buckingham. . . ... 121 21.9 63.5 280 WI Waukesha. . ..... 478 3.8 5.5 99 VA Nelson ......... 121 21.9 50.0 280 WI Washington . . . . .. 478 3.8 6.7 99 VA Fluvanna. ....... 121 21.9 21.5 280 WI Ozaukee. ....... 478 3.8 6.6 9 YA Madson........ 27 21.9 8.1 282 WI Wood. ......... 480 10.3 4.8 pe YA Groen, uess.ws yo! 219 74 282 Wi Cla. econo rans 480 10.3 20.0 29 VA Bappehapnock.; » 2 299 52.3 282 WI Taylor.......... 480 10.3 12.2 109 VA Prince Willia. . . . . . 136 22.7 23.5 282 WI Marathon. . ...... 588 10.3 7.8 109 VA Fauquier. ....... 136 22.7 21.3 282 WI Lincoln. ........ 588 10.3 14.7 130 VA Brunswick. . ..... 232 36.2 63.0 284 WI Oneida......... 482 26.0 16.1 130 VA Greensville ...... 232 36.2 18.5 284 WI Vilas.....cc00:0 482 26.0 19.3 } ) 284 WI Forest ......... 482 26.0 39.2 i VA PHONE op ue gs 208 174 284 MI Gogebic ........ 659 26.0 33.2 132 NC Caswell. ........ 234 20.8 48.4 284 WOR. 659 26.0 407 72 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area— Con. Percent of routine stays outside 800-unlinked Percent of routine stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of— unlinked unlinked ——————————————— unlinked unlinked area no. State County area no. Area County area no. State County area no. Area County 284 MI Ontonagon ...... 660 26.0 32.2 7 WV Nicholas . ....... 134 13.1 27.6 200 ETL 487 103 47 7 WV Webster ........ 134 11.1 27.3 290 WI Monroe. ........ 487 10.3 7.6 30 VA Alleghany ....... 33 22.9 11.5 290 WI Vernon......... 487 10.3 6.5 30 VA Bath, coiovsmsws 33 229 21.2 290 MN HOuSION ... «+ «xv 487 10.3 14.2 30 WV Greenbrier. . ..... 154 22.9 22.9 290 IA Allamakee . . ..... 487 10.3 14.7 30 WV Monroe. . ....... 154 22.9 53.2 290 WI Trempealeau . . . .. 629 10.3 253 31 WV Monongalia . . . . . . 34 14.0 7.0 298 WI EauClaire....... 495 14.4 8.2 31 WV Marion ......... 34 14.0 9.3 298 WI Chippewa . ...... 495 14.4 9.5 31 WV Preston. ........ 34 14.0 24.4 298 WI DU pv cvvnnes 495 14.4 14.0 31 WY Taylor... ...u:45 197 14.0 34.7 pe y puftalo nr amd a 124 Se 34 WV Wood. ......... a7 10.3 46 BPR: 3 5.8 is x se 2 on ’ ? 34 OH Washington . . . . .. a7 10.3 10.8 301 WI Dang ....osvnsx 498 13.3 6.3 34 WV Ritchie . ........ 37 10.3 13.9 301 WI Columbia ....... 498 13.3 125 34 WV Pleasants ....... 37 10.3 9.6 301 WE Sauk ; ove winnwe 498 13.3 4.7 34 WY Will. ons saan 37 10.3 5.0 301 WI OWE... corms mas 498 133 5.7 34 WY ROBNG & 4.0 0'vs vx 118 10.3 24.9 301 WI Marquette . . . .... 498 13.3 45.1 34 WV Calhoun ........ 118 10.3 12.9 301 WI Grant. . ccs 567 13.3 36.2 34 WV Jackson ........ 156 10.3 22.4 301 WI Richland. ....... 567 133 18.3 a9 WY Merger . . oct a2 17.1 10.8 306 WI FondDulac ..... 503 16.4 12.8 39 VA Tazewell........ 42 171 11.1 306 WI Dodge ......... 503 16.4 29.0 39 WV McDowell ....... 42 174 27.9 306 WI Greenlake ...... 503 16.4 11.8 39 VA Bland.......... 42 174 58.0 306 WI Winnebago . ..... 565 16.4 10.8 39 VA Buchanan....... 173 17.1 21.3 306 WI Waushara . ...... 565 16.4 29.1 46 WV Cabell . ........ 50 14.5 9.4 314 WI Washburn... .... 511 36.4 31.9 46 OH Lawrence ....... 50 14.5 8.5 314 WI Burnett. ........ 511 36.4 43.1 46 WV Wayne ......... 50 14.5 57 315 WI Marinette. . . . .. .. 513 217 25.6 46 WY Linco. «wi weve 50 1s 422 LE 46 KY Boyd .....: co... 102 14.5 55 315 MI Dickinson ....... 513 21.7 14.0 : 46 RY Greenup. ...u «qu. 102 14.5 24.3 315 MI Menominee . . . ... 513 21.7 26.7 46 KY Canbl. .o.conras 102 14.5 40.5 315 WI Florence...,...x- 513 27 18.3 315 Ml lon 673 217 20.6 46 KY Lawrence ....... 102 14.5 20.8 SER : : 46 KY Matin. ......:.x 102 14.5 47.4 326 WI Green. ....rsw» 523 17.5 19.3 50 WV Randolph . ...... 54 19.3 11.4 326 WI Lafayette. ....... 523 17.5 36.1 326 Wl Rock . .. 665 175 14.9 50 WV Barbour ........ 54 19.3 19.7 : 50 WV. TRICKS wv: 5 bs sein 5 ox 54 19.3 31.1 344 WI Outagamie. . . . . .. 540 18.2 17.2 50 WV Pocahontas . . .... 164 19.3 315 344 WI Waupaca. ....... 540 18.2 19.7 60 WV Raleigh. ........ 66 18.1 11.4 355 WI Manitowoc. . . . ... 550 11.7 8.7 60 WY Fayefte......+.. 66 18.1 18.5 355 WI Calumet ........ 550 11.7 35.3 60 WV Wyoming. ....... 66 18.1 35.5 355 WI Sheboygan ...... 671 11.7 9.2 60 WV Summers ....... 194 18.1 27.7 357 WI Ashland ........ 553 28.8 12.7 92 WY Harmison ........ 109 18.9 13.5 357 WI Sawyer......... 553 28.8 44.4 92 WY Lewis... comsun 109 18.9 127 357 Wi Bayfield « «von se 553 28.8 21.0 92 WY Braxton. . uv. «ws 109 18.9 24.4 357 WI Price .... 000: 608 28.8 46.8 92 WY Gilmer ... cox 109 18.9 57.8 370 WI St.Croix........ 573 36.1 29.9 82 i Doddridass. «aes 109 189 29.6 370 WI Pierce. . ........ 573 36.1 457 22 Upshur. ........ 198 18.9 19.9 370 MN Goodhue. . ...... 1167 36.1 36.3 586 SD Lawrence ....... 972 25.0 22.4 586 WY Crook... «.:m:.05 972 25.0 38.0 374 WI Barron ......... 578 21.9 19.9 374 WI Rusk .......... 578 21.9 29.3 586 WY Campbell ....... 1348 25.0 24.1 726 WY Natrona ........ 1239 10.3 10.0 379 WI Shawano. ....... 583 34.7 41.9 379 WI Langlade. . ...... 583 34.7 24.0 726 WY Converse ....... 1239 10.3 12.1 379 WI Menominee . . . ... 583 34.7 24.4 741 NE Cheyenne....... 117 16.3 27.2 x 741 NE Deuel.......... 1117 16.3 66.2 382 WI Racing: .:.u:u « 587 14.0 13.4 741 WY Laramie ........ 1256 16.3 9.8 WI K Nis voi sias 87 : 14. a Sangha 5 140 9 741 NE Kimball. ........ 1256 16.3 18.7 387 WI Jefferson. . ...... 596 30.1 26.5 749 WY Pam... census 1266 24.8 21.5 87 WI Walworth. . ...... 597 30.1 34.0 3 749 WY BigHom. ....: .. 1266 24.8 30.0 51 Wl AEBS... petosnn of aa 415 770 WY Sheridan 1291 18.6 19.5 Wi Juneau.... .eva 45.2 4B: | She i he oo ai Sine e 2 351 t dunsay 623 Bs 770 WY Johnson... ..... 1291 18.6 15.9 WI Jackson ........ 67 36.5 1 5 Baan 8 36.8 77 WY Albany ......... 1295 19.3 16.5 400 WI Portage. . ....... 679 24.3 24.3 779 CO Jarkeon : «ows +» 1295 19.3 46.4 7 WV Kanawha... ..... 8 11.1 6.9 77 WY 1C2roof. us sue 1296 19.3 20.9 7 WY Putnam. ......«.. 8 11.1 127 775 WY Linco. ove sms 1314 40.6 43.6 7 WV Boone ......... 8 11.1 9.8 775 WY Teton.......... 1315 40.6 31.3 7 WV Clay. .......... 8 11.1 38.7 775 WY Sublette ........ 1315 40.6 53.5 73 Table Il. List of health service areas for 800-area and 1400-area unlinked solutions and percent of routine Medicare stays by residents outside 800-unlinked area—Con. Percent of routine Percent of routine stays outside 800-unlinked stays outside 800-unlinked 800- 1400- area by residents of — 800- 1400- area by residents of — unlinked unlinked ——rrrrrr—— unlinked unlinked —————r+———"= area no. State County area no. Area County area no. State County area no. Area County 777 WY Hot Springs. ..... 1322 16.9 18.9 797 WY Platte: .«. wa vas 1398 27.5 275 7 Washoe SE pv 1323 i ger 799 WY Sweetwater . . . . .. 1400 205 20.1 ONE +s iwi 3 t ! 799 UT Daggett. ......: 1400 20.5 45.5 792 WY UintZ..sasmeaes 1893 304 Li 804 WY Weston. ........ 1405 34.0 34.0 74 Appendix IV Obstetric service areas Obstetric service areas were defined using the same methodology as for the health service areas except that natality data for 1984-86 were used to calculate the distance matrix that was input to the cluster analysis. Since the 800-unlinked solution was the preferred solution for the health service areas, that solution is presented for the obstetric service areas. As for the health service areas, modifications were made to the initial solution to ensure contiguous obstetric service areas. Two zero-production counties (Grant, Ne- braska and Grainger, Tennessee) were not contiguous to the cluster where the greatest number of their residents’ births occurred. These two counties were reassigned to the cluster where the second largest number of births oc- curred. Seven other clusters were found that contained noncontiguous counties. Solutions for two of these clusters were not changed as the noncontiguous counties were separated from their clusters only by water (Marin, Cali- fornia; Island and San Juan, Washington). Four noncon- tiguous counties were reassigned to contiguous obstetric service areas based on examination of their travel patterns (Rio Blanco, Colorado; Ontonagon, Michigan; Clatsop, Oregon; and Jefferson, West Virginia). The obstetric service areas for the 800-unlinked solu- tion are presented in table III. Appendix tables IV and V present information on travel patterns outside the obstet- ric service areas comparable to that presented in text tables E, F, and H for health service areas. 75 Table lil. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked eet SoS Sees unlinked ee area no. State County Area County area no. State County Area County 140 AL -Jefflerson. ... vex 2.7 0.5 257 AL CUIIMBN & ovr » 00 mh 4130 21.8 21.8 140 AL Shelby ............. 2.7 1.7 140 AL St.Claif ............ 2.7 15.3 252 AL JBCKEOD i xewr runs 308 208 140 AL Blount ............. 27 17.4 409 AR Sebastian ........... 6.9 2.1 140 AL Chilton... ows vvs uss 27 21.5 409 OK LeFlore ............ 6.9 13.5 142 AL Montgomery ......... 49 12 409 AR Crawlord. .. .s sss 0s 00 6.9 29 409 AR logan ............. 6.9 13.2 142 AL. EIMOIS: i wen ws senna 4.9 8.4 3 ; 409 AR Franklins . oom va nwwe 6.9 125 142 Al. Covinglon . .:s «onus 4.9 13.1 409 OK. Haskell, ......onvens 6.9 28.3 142 AL AUBUEA . .. vivir snare 4.9 12.1 409 AR Scoft 6.9 32 142 AL. Buller... sos enna 4.9 43 § 2c 202020202 SusirahisaseNrruce ’ ’ 142 AL Crenshaw ........... 4.9 9.7 413 AR Pulaski............. 4.1 0.9 142 AL Lowndes. ..: «. sus ums 4.9 13.9 413 AR- Saline. .... suas 9 0m 4.1 3.0 165 AL TUSCAIBOSR . vv as vm an 9.0 4.7 413 AB LONOKD. , uo scum wai 4.1 57 + 413 AR Arkansas. ...ou evens 4.1 7.3 165 Al. Pickens. . i: ses sv 5m 9.0 15.4 : 413 AR Grant. .... o0svvinns 4.1 29.1 165 AL Bibb. .............. 9.0 30.5 1 A 165 AL Greene 20 15.1 413 R MONIOR. . ..oo cvs viva 4.1 29.6 Troe ’ ’ 413 AR Prairie ....v ms amsms 4.1 23.3 180 AL HOUSION . co cw 5 vues 8.4 3.1 413 AR Perry .............. 41 57.3 330 2 Do ALR ag a 3s EA 427 AR POPE .............. 17.2 15.5 OBL; nasil SAEs ons : : 427 AR Johnson. ........... 17.2 19.0 180 AL Geneva. ... vv 55m 8.4 5.8 427 AR Yell 17.2 19.7 180 FL: “HOWNBE. . 4 2505 4 5 mln 8.4 544 [| T7000 TR TER mtmamaameaaaes E : 180 AL HHOINY. vn swum sm en 8.4 9.8 429 AR Bader ............. 71.7 12.2 181 AL Dallas. ............. 15.8 43 429 AR BOOMS .vvumsmsmn sus Wy 64 429 AR Carroll ............. $1.7 16.4 181 AL Clarke: wi: vos wis mn ssn 15.8 27.3 429 AR Marion ......c3 cusns 11.7 5.4 181 AL Marengo. ........... 15.8 114 429 AR Searcy ............. 11.7 33.2 181 Al. HA: ss name sivsn smn 15.8 88.2 429 AR Newton 11.7 6.5 181 AL Perry.............. 15.8 43 | TTT 2 rhein abishis ewes ’ ’ 181 AL. WHEOX svn ss brim sma ae 15.8 1 433 AR. Craighead. . vs vw vuws ns 15.6 7.9 193 AL LRudRrias « «ooo 8.6 a5 433 AR Greene wh wh Bod ALR EI 15.6 4.6 433 AR Poinsett ............ 15.6 41.9 193 AL Colbert. ..... us swan 8.6 3.6 433 AR ChaY.u ou sasms owe ms 15.6 29.5 193 TN Lawrence ........... 8.6 17.7 . 433 AR Lawrence ........... 15.6 16.3 193 AL Frankly oo vi v0: 9:0 8.6 14.8 433 AR Randolph 15.6 8.9 193 TN Wayne ............. 8.6 20.4 Ph pwn ew caine : 3 198 AL Madison ............ 7.4 2.41 hs BR a jas 198 AL Marshall oo vim mmsn 7.4 28 | 0 TTmorrrrrrrrrrrrts ’ ! 198 IN Lincoln. ..... oven 7.4 20.8 441 AR Hempstead .......... 275 17.2 204 AL Mobile ............. 59 05 a 8 Foe Ry er rms sean iL 32s 204 MS Jackson ............ 5.9 09 | Fr 0020 TESA Eetueieiehemre- i ! 204 AL Baldwipy ....ooni. uns 5.9 7.0 475 AR “Union. : vos ws sv suns 12.8 4.0 204 AL Washington. ......... 5.9 28.0 475 ABR: ‘Ouachita. : . ... 05 sas 12.8 9.0 204 MS George. = «vv .uv vse 5.9 11.6 475 AR" Dallas. cos vann «rime s 12.8 66.4 205 a 33.0 38.5 475 AR Cahoun............ 12.8 22.8 205 AL Monroe. ....ssas swiss 33.0 24.2 487 AR White. ............. 28.8 18.1 205 AL Conecuh............ 33.0 35.4 487 AR Cleburne. ........... 28.8 38.9 215 AL Walker ............. 35.0 33.0 487 ARS: WOOT oon els 4 ii 6 215 IL 35.0 41.1 493 AR Jefferson. . ....co vei 18.9 19.5 219 AL Etowah... .......... 13.8 6.8 493 AR Desha ............. 18.9 22.6 493 AR Drew ...:vvevsuvens 18.9 123 219 A. DeKalb .c:uvvmeninin 13.8 23.0 + 219 AL. Chaiok 13.8 305 493 AR LINCO... co osunsss 18.9 23.0 RYOKSE cs «= 23g ce 2 2 493 AR Bradley. ............ 18.9 9.6 226 AL Morgan. . ...o.mvsmss 15.5 13.7 493 AR Cleveland . .......... 18.9 22.3 28 2 Snavins hab LE bg 1 496 AR Faulkner. ........... 47.6 53.6 BWISIIOE ow wie be £ iw 0 : : 496 AR Conway ............ 47.6 36.4 227 AL L166, rv vy summa 17.8 17.4 496 AR VanBuren........... 47.6 36.9 227 AL Chambers. .........s 17.3 14.5 497 AR Garland ............ 25.7 17.8 227 AL Tallapoosa. .......... 17.3 18.4 5 497 AR HotSpring........... 25.7 37.1 227 AL Macon 5 cs scien sma 17.3 156.3 207 AL Bullock 17.3 26.3 497 AR Clark ...onnsusmmsns 25.7 29.6 ROOK 5.2 sll tinlele mint : 497 AR Pike. .............. 25.7 46.1 241 AL Calhoun «uu. wives wen 13.0 6.5 497 AR Montgomery ......... 25.7 20.6 241 AL Talladega . : vss saan 13.0 15.8 505 AR Crittenden. . . ........ 32.1 42.7 241 AL Clay. .............. 13.0 8.5 . 505 AR St. Francis. .......... 321 18.4 241 AL CIBbUING. «wo. wii sas 13.0 31.0 505 AR Cross 32.1 25.0 241 AL Coosa ............. 13.0 522 | YT TRE ’ ’ 28 AL PKe. 352 27 | 20 A Lemeten. 57 7s 248 AL" Barbour . . «ux «no 0 tne 35.2 39.2 248 GA Qulman ......s +: sus 35.2 52.6 515 AR Ashley ............. 38.7 47.0 76 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of— 800- by residents of— unlinked unlinked area no. State County Area County area no. State County Area County 515 AR BhICHE «i «ovvn wamns 4 38.7 27.8 764 CAR YOIO. co vvvnmasnmsin 21.9 38.9 561 AR Washington . . ........ 9.1 2.1 704 CA “SUBBL. cvs umn anna 21.9 7: 764 CA Yuba.............. 21.9 8.8 561 AR Bemon; sv: vo vss zs 9.1 2.9 764 CA Coluss 219 120 561 OK Delaware. . .......... 9.1 437 | 0 T TEEEErrrrrrrrras : : 561 MO MeiDonald.. . . . «x «ails 9.1 36.7 773 CA San Bernardino. . ...... 14.2 14.6 561 AR Madison. : x nnam sas 9.1 10.7 773 CA Riverside. . ..ouvvsvsis 14.2 13.7 596 AR Independence ........ 19.6 9.8 776 CA LosAngeles. ......... 1.1 1.2 596 AR JACKSON . .. isos sma 19.6 10.6 776 CA Orange. ..:«so6 smsms 1.1 0.9 596 AR Sharp. ............. 19.6 55.8 776 CA Ventura. ............ 141 0.8 2 49 a GALLE EERIE 1s os 789 CA Butte .............. 7.8 7.6 DP 813 P10 Bown to : 2 789 CA Tehama ............ 7.8 10.8 647 AR. Mississippi. . . vv ws swe 12.9 12.4 789 CA Glen... sovnvss vos 7.8 4.3 647 MO Pemiscot. . .......... 12.9 14.4 701 CA Sarin Baibars . ....... 23 25 714 AZ Graham .:..s.65 955 19.6 235 791 CA Sen Luis Obs. ...:4:.. 23 1.9 714 AZ GIoRMBS. «cs vkw vn mes ws 23 794 CA Lassen. ............ 27.2 29.5 727 AZ PNR soi6s sa vwivase 31:1 1.0 794 CA Plumas............. 272 24.0 727 AZ BEtaCAR .os tune na 1 28 800 CA Humboldt ........... 1.8 18 72 AZ MalCopd.s «even owen 12 05 801 CA INYO. «even, 19.0 10.8 728 AZ Pinal. ..cnss aime sunns 1.2 11.3 708 AZ Gil 12 6.1 801 CA MONO. ..:sv55nsnews 19.0 26.5 Win 2 0 8 wri rs ram oy : : 801 NV Esmeralda. .......... 19.0 52.6 762 AZ COCONIND ... +» v5.55 5.010 10.6 9.5 762 UT Kane .............. 10.6 255 50 OA KBEY wis suse smo 59 59 762 JT. Garfield. wi: wn sms nme a 10.6 14.8 811 CA Tulare. ............. 10.1 10.1 783 AZ Navajo « : «: ctcm rumen 46.7 47.2 815 CA Momersy. . .c.es ves 587 57 783 UT Sandu coevvvssnns 46.7 44.3 818 CA Imperial ............ 6.3 6.3 797 AZ YUMA. ......000 cn 14.4 10.2 . 797 AZ Mohave «ot 144 227 819 CA SanDiego. ..: us +0: 0.9 0.9 . 701 CO Denver............. 13 05 803 AZ COCHISE ..v isin 4350 13.3 13.3 701 CO Jeflerson. . o.oo. 13 16 824 AZ Yavapai. ............ 10.5 10.5 701 CO Arapahoe ........... 1.3 07 698 CA Sacramento. ......... 2.9 25 Li OD Adah. seven rime sus Ey 2 701 CO Douglas . .....o.00:0¢ 1.3 2.2 698 CA PlECBI. «ois vv ve vw wns 29 4.4 698 CA Neved 29 75 701 CO Elbert. ............. 13 19.8 BVBORB. «wi vs w3 Wns Han ’ 701 CO Clear Creek. ......... 1.3 1.2 720 CA: "FIEBNO i: « a wi wis wie wmv 0 25 1.5 701 CO PK. .o vv vanmssmrnn 1.3 19.5 720 CA KINGS ...vs so sms ihsn 25 4.2 701 CO Gilpin.....s:05 5002 1.3 29.6 20 CA Model. ,ucntvmennea 2.5 52 707 CO Alamosa. ........... 3.9 2.6 725 CA San Francisco ........ 11.1 22 707 CO RioGrande . ..s: 5:0 3.9 2.8 725 CA SanMateo. .......... 11.1 21.7 707 CO Conejos .....oooov nv 3.9 2.8 725 CA Matin, o.. seen vsaons 11.1 8.7 707 CO Saguache . . vs cms» 3.9 13.4 730 CA ElDorado........... 19.6 245 7% o Si FARRER wanes ye i 730 NV Ormsby ............ 19.6 12.5 YB 5 wn comet 0 : s 730 NV Douglas ............ 19.6 11.4 708 CO Puehlo . sun sus wims os 3.3 29 730 CA Apine, .....:cosvusu 19.6 25.5 708 CO Huerfano............ 3.3 10.9 734 CA Sonoma ............ 9.4 8.8 732 CO KitCarson..:::: +540 18.0 12.8 734 CA Mendocino .......... 9.4 4.0 732 CO Lncoin............. 18.0 20.9 734 CA Lake . ... 9s 6055s 9.4 24.3 732 CO Cheyenne... .u.u.ss 18.0 25.7 736 CA Stanislaus. . .......:. 3.4 2.8 758 CO Otero. ............. 7.9 6.8 736 CA Merced. ............ 3.4 3.4 758 CO Prowers ..... : sv: 7.9 9.0 736 CA Mariposa. . .......... 3.4 29.6 758 CO Bent. o.ov vv nnnmmoms 7.9 2.4 737 CA SanJoaquin ......... 10.5 10.2 7 2 S OMG) wsvsvmeminns ? 2 3 2 737 CA Tuolumne ........... 10.5 7.1 OWE: ie Bama etiinee 4 737 CA Calaveras ....: «cys: 10.5 8.1 767 CO. Gadisld, . svn 5 uamn ws 16.0 8.9 737 CA Amador. ............ 10.5 30.3 767 CO Eagle. ..: 9 snnmmsnny 16.0 135 746 CA Shasta............. 43 3s Ie 22 Sa Braess ous Hips gu 746 CA Trinity. ............. 4.3 13.7 I sie we ma amma vm ; 747 CA Alameda. ........... 78 8.2 772 CO Weld .............. 25.2 28.0 747 CA Contra Costa 78 71 772 CO Morgan. ...::s:0+00- 25.2 7.6 nea ee E : 772 CO Washington. . ........ 25.2 37.7 TTY 17. ! 1 ga Sana ye 24 777 CO Logan ............. 12.3 105 REE EE ’ ’ 777 CO Phillips. ............ 12.3 8.1 756 CA SantaClara.......... 3.8 4.0 777 CO Sedgwick . .. cursor ovs 12.3 33.9 756 CA SantaCruz .......... 3.8 2.0 756 CA SanBento .......... 3.8 6.2 773 CO M882. covcrnwunvans 3 13 77 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked Ce unlinked i —— er r————— area no. State County Area County area no. State County Area County 784 CO EIPo80.... .5: +5: 2.8 2.0 166 FL PamBeach.....:.:... 7.6 55 784 CO Fremont............ 2.8 13.0 166 FL... SLLUCIB. .. o.com ems 7.6 9.9 784 CO Tellor.. :o.voevn ives 2.8 2.8 166 FL Matin. o.ocvvsevimse 7.6 3.0 784 CO "CUSter & «os vim bt vime ma 2.8 45.6 166 FL Okeechobee ......... 7.6 11.6 785 CO Montrose. . .......... 13.2 12.5 bi ig Bey wesw samy 79 as 785 CO Deita.......coonnuns 13.2 183 | TF 020202020 TESS vies % 785 CO SanMiguel .. v.55 «5 13.2 19.4 182 FL. Polk. uv vuiswairwzmsis 6.4 5.7 785 CO DAY. vs wn sims mm ems 13.2 7.0 182 FL. Highlands , . .v.ccvo vrs 6.4 4.0 787 CO ROUtt......''vvnn.. 8.7 5.0 1% i Peguie sme mire 24 25 787 CO Moffat. . ............ 8.7 12.3 ] HT 02020 hr SAR aaa 3 y 703 CO Yuma. ............. 39.9 41.2 214 FL Orange 2 a RE 1.8 0.8 703 NE Dundy 399 350 214 FL Seminole. ........... 1.8 0.9 pr fr : ’ 214 FL Lake &.,.osvsvnwuses 1.8 9.7 796 CO Chafles. ..vo.us wavs s 9.9 6.1 214 FLL Osceola ............ 1.8 1.7 78 CO Lake ...imennrmsipur Bs 138 223 FL Pinellas. ............ 15 05 812 CO Grand. «..un soins vais 31.1 31.1 223 FL Hillsborough ......... 1.5 1.8 223 FL. PasCO. .... 5:5 wwss 1.5 22 813 CO Boulder. ............ 16.8 16.8 223 FL Hemando 15 aa 3 00 Lofimet: vous sanases 63 68 228 FL Sarasota............ 4.9 2.7 832 CO GUNNISON .v + viv vx vw wie 11.7 10.3 228 FL Manatee............ 4.9 4.2 832 CO Hingdalg. .«.vvvuswus 11.7 38.1 228 FL Charlotte. ........... 4.9 15.1 834 CO RioBlanco........... 35.7 35.7 234 FL. D808 o:nsmn ins mbis 2.0 2.0 7 CT Hartford . ........... 5.7 4.2 234 FL Monroe. ............ 2.0 2.0 7 CT TOWBAR. ..s oa vie wmv om 5.7 4.4 235 FL Jackson . «ovo insane 27.2 29.5 7 CT Windham .....u.o0s x 5.7 19.2 235 FL Calhoun ............ 27.2 18.7 76 CT Fairfield ............ 6.5 3.8 238 Fl: be8 us vv sive snus £600 2.2 2.1 76 CT" NewHaven........«. 6.5 3.8 238 FL “COMBE .iv sin viwm women 22 2.6 76 CT Litchfield. ........... 6.5 18.4 . 76 CT Middlesex . . . . . ...... 6.5 26.2 244 FL. MaHON « - ov mws vms ws en 28.7 28.0 244 FLL CRIUS. ... cvccmemn ra 28.7 13.9 87 DE NewCastle .......... 11.5 6.8 244 FL Sumter............. 28.7 63.4 il MD: ‘Cagis. ur wu se mrnens Hs 78 247 FL StJohns ........... 35.9 325 97 DE SUSSEX: suv cunvnas ub 4.7 47 247 FL. PUN. own v3 550% 20 35.9 39.6 97 DE Kent............... 4.7 6.3 97 MD Wicomico . .......... 4.7 3.0 253 FLL. Brevard. .....cvonsunn 4.0 4.0 97 MD Worcester . . ......... 4.7 4.2 254 FL IndianRiver.......... 9.2 9.2 97 MD. Somerset ... vss 020es 47 26 262 FL Volusia. ............ 115 11.2 139 FL *BaY cuss voi sms nme 7.9 3.2 262 FL. Flagler. . .. + «sms sv sn 11.5 18.0 139 FL. Washington... « eu ous 7.9 40.1 139 FL auf... 79 35 263 FL Broward : ; coos mwas 6.8 6.8 139 FL. Franlihv ooo v0 smass wu 7.9 32.2 137 GA MUSCOgRR. . «+ +s vv s+ 5.2 17 137 AL, Russel. . ...ouviaws us 52 4.8 14 he Alachua alia bast 49 11 137 GA Chattahoochee . . . ..... 5.2 1.6 143 FL “Columbia . . cu ome. 4.0 3.5 : 137 OA "Hams. : cosmos wsmams 5.2 17.4 143 FL Suwannee........... 4.0 14.7 137 GA Randolph ........... 52 33.3 143 FL. L&W... cvs amems nme 4.0 9.4 137 GA Tabel.. .vsvvsniones 5.2 18.6 143 FL Bradlord.. .: «sswsims 4.0 12.7 4 137 GA Stewart. ............ 82 14.7 143 FL. .UAIBA. . .. v5: 2% worse 4.0 6.4 . Sot 137 GA Marion... ....:veuvss 5.2 46.8 143 FL. DMB wv vsnessnans oy 4.0 15 137 GA Cla 5.0 42.2 143 FL Gilchrist ............ 4.0 0.0 ns 2 ip i eg oi 138 GA Richmond... sims 4.1 0.8 14s Fl. DValyyrsesnwimimaes 18 29 138 GA Columbia ........... a1 0.7 145 FL Cla. ...cimews amemme 1.6 5.6 138 GA .BUIKS. . v2 ws vpewe su 4.1 2.5 145 FL Nassau. ....:vims ves 1.6 1.4 145 FL Baker 16 10.4 138 GA McDuffie ........... 4.1 1.5 ps EEE ’ i 138 GA Washington. ......... 4.1 26.5 156 FL Escambia .... «-5355+ 25 1.7 138 GA: Jofferson. .. wus wavs ss 4.1 35 156 FL. OKBIOOSA wz wv savin ob 25 24 138 GA Screven ............ 4.1 34.5 156 FL SantaRosa .......... 25 1.0 138 GA Jenkins. .....soawsus 4.1 13.9 156 Fl. Walton vcovsvs oman 25 22.5 138 GA Lincoln............. 4.1 5.9 161 Rr 70 16 138 GA WEIR, ...usvmamany 4.1 8.7 138 GA Glasscock. ........... 4.1 7.0 161 FL Gadsden. .....:«wsvu 7.2 3.7 138 GA Taliaf a1 32.1 161 FL Taylor. ............. 7.2 11.3 PRIBHO x a ginal . : 161 FL Madison .....ssuesns 72 37.0 141 GA Floyd: :usvamigmwss 15.7 3.5 161 FL. Wakulla, . cranes 7.2 2.4 141 GA Bartow ............. 15.7 34.3 161 FL Jefferson. ........... 7.2 16.2 141 OA POI. ©. onvvssnmss ns 15.7 5.1 161 FL ‘Lafayefle. ......«voeu 7.2 75.3 141 GA Chattooga. . « «ovo 0.44 15.7 274 wi Flo “IBRY 54 wip wii rw 2 813 147 GA Lowndes. ....cscvvws 99 25 78 Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area by residents of — 800- by residents of — 800- unlinked Ll unlinked ee ——————— ES —— area no. State County Area County area no. State County Area County 147 GA COOK: vives orwws as 9.9 20.5 190 GA. POIESKY, «5.0 3058 283 22.4 38.7 147 GA Berrien ur ts 0 tt 1 9.9 34.8 195 GA Clarke o.oo 21.0 4.0 147 FL Hamilon....:cevwwsos 9.9 25.1 ; 195 OA WEHON . . oss vmenin 21.0 47.0 147 GA Lanier. ......vewe sos 9.9 5.1 147 GA Echol 99 65 195 GA Bamow..........e0n 21.0 46.4 CAKE + 2s me sitims ws : 195 GA Madison. ........... 21.0 174 150 GA FUlton .......cvcon: 3.0 24 195 GA 'OCONBE «vu «vein nuns 21.0 4.1 150 GA DeKaD o o.o0.05 sie wins 3.0 2.3 195 GA Morgan. . .: «sine suis 21.0 23.7 150 GA Gwinnett. ........... 3.0 21 195 GA Greene. ............ 21.0 14.1 150 GA Clayton. . ..«.. vieidsws 3.0 2:2 195 GA Oglethorpe . ......... 21.0 10.5 10 GA PaYBHB. on vsrsvsves 35 138 196 GA Whitfield . . .......... 12.4 5.6 150 GA: Hen. . ccs vs vai sam 3.0 10.5 150 GA Forsyth 3.0 18.5 196 BGA GOIAON. «iv vu ww v sv vs an 12.4 23.2 Toor ’ ! 196 GA Munmay............. 12.4 4.3 160 GA Dougherty. .......... 7.2 25 196 GA GIMBE «5 + vs wn xv vs www 12.4 39.3 160 GA Mitchell. vc ccc: «dus 7.2 23.2 197 GA GYAN... voerenn. 195 6.3 160 GA Worth. . ............ 7.2 10.7 197 GA Camden............ 19.5 42.8 160 GA Lee............... 7.2 10.4 197 GA Mc Intosh 19.5 28.7 160 GA Temell. ..... cavum 7.2 63 § TX 0 ARES rdiahgiewan ’ ’ 160 GA Baker.............. 7.2 25.2 202 GA CHSD .ovnuwvssvwvaes 45.1 30.3 162 GA Sumter. ............ 17.2 5.0 Zz 2% Doug seseppusnsne £ oe 162 GA Macon ............. 1722 08 § T0020 TEMA Rrasmaea wean ! : 162 GA Sohigy «ocx wsimiws mms 17.2 19.4 208 GA: 'ColieB . . + civ ws sms s 25.3 9.1 162 GA Webster . ........... 17.2 39.0 208 GA ApPING. « « six ws vvinin we 25.3 50.4 164 GA Bibb. ooo 94 40 208 GA JeffDavis ..... own ams 25.3 23.3 208 GA Bacon +: «ovr runs 25.3 26.3 14 GA. LIDEON us amean avian 94 2 208 GA Atkinson 25.3 46.9 164 BA JONBB. . vo toms swrms 9.4 68 | TF 200 HE EER assays : ! 164 GA MONIOB. « . outs snvun 9.4 6.1 212 GA Lawens ...:.. vu. 21.6 10.3 164 GA Lamar... vosusvarny 9.4 21.4 212 BGA DOUG «wv wav umn 21.6 19.9 164 GA Wilkinson .........s 9.4 41.6 212 GA Telfair. ............. 21.6 33.0 164 GA TWIGEE. + «vv vrvena 9.4 11.7 212 GA JONSON . «ven wmv an 21.6 29.7 164 GA PKB. ..cisvvmsvunny 9.4 45.8 212 GA Troutlen ....csvvruss 21.6 41.9 164 GA “TaYIOF. «ss 52% ds www 9.4 36.5 212 GA Wheeler .. ..couvuwmun 21.6 47.9 164 GA Crawford. . . ..c v2 wvne 9.4 18.4 220 GA Decalur. ............ 26.4 25.6 169 GA CObD...cvimsusismsna 43.4 48.9 220 GA Seminole. ........... 26.4 29.1 169 GA. "OHBIOKEE , . vv vv vsivs 43.4. 48.3 222 GA Rockdale... ......... 38.8 478 169 GA Douglas ....ris sms 43.4 20.6 iy 222 GA Newlon. . ...csvieimus 38.8 24.2 169 GA Paulding: . ome cui 43.4 20.3 200 GA Jasper 38.8 74.3 169 GA Pickens. ....... 0:5: 43.4 230 | = 2 SMeESERLRLLecRarErbee ’ 170 a 20.3 43 229 GA Early 3 RE oe 44.1 35.5 4 229 GA MBF. ..:coo353 345 44.1 43.8 70 GA. PIBICS vues vnsnns 20.5 89 229 GA Calhoun ............ 44.1 66.5 170 GA Brantley . . cu vores 20.3 58.5 170 GA Charlton. «ise ss sa08 20.3 63.5 230 GA Chatham. ........... 3.5 0.9 170 GA Clinch ............. 20.3 20.8 230 GA Liberty ............. 35 2.0 172 GA Bulloch. ............ 32.0 27.1 230 GA . BHRGRAIM . «oc vse 05s 28 20 230 GA Wayne ............. 35 6.6 172 GA Emanuel............ 32.0 38.1 230 GA Tattnall. ...... 0000 35 38.1 172 GA EVANS. ... cies ensnn 32.0 38.0 230 GA B 3 9.6 172 GA Candler 32.0 28.9 Wamp ameme susan : EE I SR angus ’ ’ 230 BA Long : icv axvnma sme 3.5 6.0 175 GA Thomas ......sxssms 173 5.1 232 GA TIOUP. «vee een. 21.2 9.9 175 GA Grady. ............. 17.3 9.2 175 GA Brook 173 56.6 232 GA ‘Cowela. : + svi swine an 21.2 24.1 TOOK: vx titi ini 8 : : 232 GA Meriwether . ......... 21.2 39.2 177 GA Hall... .oouvsnns onus 17.3 9.6 232 GA Heard. ...vevsvvavnn 21.2 39.0 7 GA JBCKSON our vrunsnnss Vs #11 249 GA Caroll «vue. 26.4 23.5 177 GA Habersham .......... 17.3 18.8 249 Al. Randolph . .. ce usu 26.4 33.0 177 GA Lumpkin :,... oceuss 17.3 6.2 249 GA Haralson. . . ) 26.4 28.8 177 GA WEB. . ov cvs us sms an 17.3 gg | "7 20 v0 Mester tasmree ! ’ 77 GA Banks. ............. 17.3 23.3 252 GA WIKBS +. ,0+00 v0 smas 74.9 74.9 Ww GA. DBWSEN oss vs nama 173 992 259 GA Toombs ............ 36.7 34.9 178 BA TH. on venaiy ot Boss som = 9.2 7.4 259 GA Montgomery ......... 36.7 44.6 178 GA Ben Hill. «comin sau 9.2 8.0 : 267 GA. Spalding... «vx omen aes 22.4 19.0 178 GA TUBE . «veo vivinren 9.2 16.1 7 GA Butts .............. 22. =] 178 GA IWIN «oor 9.2 11.6 % . oe 24 is 268 A QUI. oo sion nz omens 187 184 GA Baldwin ............ 20.2 26.3 olquit 5 1B 184 GA Putnam. ............ 29.2 26.3 320 IA: DUbMOUB. , « + «uv v uve 16.1 6.2 184 GA Hancock. ........... 29.2 435 320 IA CMON «cvs nwsmsmus 16.1 27.7 320 IL JoDaviess........... 16.1 36.6 190 GA Housion ....: ve emswa 22.4 16.4 2 1A KSOM + si 0% vw 4 wu 4 , , 190 GA PBABH ...icsusinsas 22.4 30.5 520 Jagkson 15 92 190 GA Bleckley ............ 22.4 53.3 345 IA. DesMoines.......... 19.5 12.8 Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 80 800- by residents of — 800- by residents of — unlinked A ——— py unlinked i esse ie area no. State County Area County area no. State County Area County 345 IB. ABB i wnivm vii noes vn En 19.5 13.0 623 IA: AUGUDBON.. + v5 vw 0s ome 17.6 222 345 H.: HANCOCK. ser s5 sma 2s 19.5 29.2 637 A JlrEOn. ovina 36.6 41.1 345 IA HeMY. ....cc cnsiws ru 19.5 17.2 637 A Van Buren 36.6 27.4 345 IL Henderson .......... 19.5 81 | TT TEE ’ : 345 MO Clark . ............. 19.5 54.9 645 IA. SIOUX. . oc nnemsinims 21.4 1.7 547 A PO... 45 1.4 vi A ig Sroka: Bele aie 2 2 po 547 IA Warren. ............ 4.5 258 | Yr 202020 HO wesrasawabaas ’ ’ 547 A JBSPRF + vu vs nnn ans 4.5 16.8 653 IA Winneshiek . ......... 255 14.5 547 IA: MBO 2 vv hv vd via 4.5 11.8 653 IA Allamakee........... 25.5 37.6 547 A Dallas. . vs vans» duis 4.5 10.4 653 IA. Howard. ............ 255 30.3 547 IA Madison. ...:««:2:++ 4.5 10.5 655 IA Clayton. ............ 475 50.4 547 JA- LUCES. . + .ovov mon wen 4.5 20.4 655 A Delaware 475 44.8 547 JA Cloke + worn snsmass 4.5 987 1 000 hy hEmeesheme maaan ’ 3 ; 657 IA Clay... ..iasemsms ens 17.5 16.5 548 IA," RIB. wo ci 508 2 fs rn a 73 3.9 bi 548 IA Benton............. 7.3 16.5 657 iad LL LLL ws 180 548 IA. JONBB.. vous vsenranin 7.3 28.1 659 IA. Appanoose .......... 32.1 26.7 551 Wh JOBRBOI «oe ee ss se tn 136 42 659 IA. Wayne ..... cuss ems 32.1 43.4 551 IA. MuSCalNg . « vas vs vvn a 13.6 6.9 660 A Caroll... conivnwnvms 19.2 12.5 551 JA Washington. ......... 13.6 3.3 660 IA Calhoun... .os 2505s 5x 19.2 36.3 53} Ih" Cotaf. ys upsnumrunes 185 422 663 IA Guthrie. ............ 47.4 46.7 pi IA TIOWR « wover niin Hoe joa 32 663 A: OIBENB. v5 5s smn ntnn 47.4 48.0 551 IA Keokuk. ............ 13.6 47.0 551 IA Louisa ............. 13.6 36.3 667 IA HEI oo ov eimin mewn 40.9 39.2 667 A Frankin ..ocus ness 40.9 43.6 552 IA Woodbury. : «cvs vs sn 7.2 2.9 552 IA Plymouth. ........... 7.2 11.0 675 IA. Marshall .... co cou vas 20.7 10.0 552 NE Dakota ............. 7.2 1.8 675 IA. Poweshiek........... 20.7 23.6 552 JA. Monona ............ 7.2 18.2 675 IA. Tama. ............. 20.7 42.1 552 SD Uniom.............. 72 26.9 677 A. SHEIOY ..cviniinnn 29.0 29.0 552 NE Thurston. ........... 7.2 4.6 552 NE Dixon. ............. 72 29.0 689 IA Crawford. ........... 20.4 20.4 567 A Wapslle : snzes wen dh 18.5 14.6 691 A° SColt....:0:svsmesy 17.8 17.8 567 IA Mahaska............ 18.5 25.8 702 ID Cassia............. 17.4 16.8 567 IA Davis.............. 18.5 17.3 702 iD: Minooka. « : owns sas 17.4 18.0 567 IA. Monroe. ............ 18.5 15.6 703 ID TwinFalls........... 4.0 3.2 572 IA Black Hawk . ......... 13.8 8.2 703 ID Jerome. ............ 40 5.3 572 IA Bremer............. 13.8 4.4 703 ID Gooding. ........... 4.0 3.6 572 IA Fayette............. 13.8 20.0 703 ID Lincoln. ............ 4.0 14.7 572 IA Buchanan........... 13.8 21.6 572 1A Butler. . ............ 13.8 29.9 713 ID Nez Perce RE 5.4 83 572 IA Chickasaw. . ......... 13.8 22.8 713 WA Asotin. ............. 5.4 4.9 572 IA Grundy. ............ 13.8 37.8 713 ID Idaho.............. 5.4 7.3 713 ID Clearwater. .......... 5.4 4.1 584 IA URIOR. ... sos s was ne. 32.0 14.6 713 ID Lewis... ..... ooo... 5.4 4.8 584 IA Adair .............. 32.0 40.9 713 WA Garfield. . ........... 5.4 8.5 584 IA. TOYIOE. 2.5.59 4 505 mn lima 32.0 56.8 584 IA Adams. ............ 32.0 26.0 716 ID Ada............... 1.3 0.7 716 ID Canyon, .. ues conve 1.3 0.8 591 A Story.............. 11.9 9.6 716 OR Malheur ............ 1.3 1.4 591 IA- BOONG ..uvvvemsvien 11.9 17.3 716 ID Payette. ............ 1.3 1.2 591 A Halon. Liu ess 0 ene 11.9 14.2 716 ID Gem ... oo 1.3 2.1 592 IA. Cerro Gordo ......... 17.1 10.0 716 ID Owyhee ............ 1.3 16.4 592 A FIO i sss sme ma sins 17.1 24.4 716 ID Washington. ......... 13 3.1 592 IA Hancock. ........... 17.1 175 716 ID Boise.............. 13 0.8 592 IA Winnebago . ......... 17.1 29.4 724 WD = NOUEY. : viv nin ma wm vs 226 215 592 IA. Worth. ............. 171 258.2 724 ID Adams ............. 22.6 25.2 603 IA Webster, .c.vvivesmas 19.8 15.3 726 ID Bonner. ............ 24.6 22.1 603 IA Wright ............. 19.8 30.0 726 MT Sanders ............ 24.6 45.5 603 IA Humboldt . . .ovn0 wns 19.8 15.7 726 ID Boundary ........... 24.6 7.7 603 IA Pocahomtas .......... 19.8 29.6 744 ID Kootenai............ 12.7 124 619 IA BuenaVista.......... 20.2 14.7 744 ID Shoshone ........... 12.7 5.8 619 IA. Cherokee ........... 20.2 20.9 744 ID Benewah. ........... 12.7 27.7 619 JA BAC. ssi cmrwe swe 20.2 33.4 ) 619 TI 20.2 145 759 ID. Madison. ........... 8.5 8.8 759 ID. Fremont .. ...i5 casas 85 7.8 620 IA Page.............. 225 15.1 759 ID Teton.............. 8.5 23.2 620 IA. Montgomery ......... 225 30.5 2 620 1A Fremont . . .......... 225 24.9 761 ID Bonneville, . un. vs ims ms 16.5 5.7 761 ID Bingham. ........... 16.5 20.6 623 IA. Cass .............. 17.6 15.2 761 ID Jefferson. . .......... 16.5 50.5 Table lil. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked unlinked area no. State County Area County area no. State County Area County 761 ID Bulle... scr esman 16.5 8.4 322 I. Woodlord ...suewn ass 8.3 28.7 761 ID Cak.............. 16.5 45.0 322 IL. Mason ...c.cusvs@s 8.3 44.7 7 ID Bannock. ........... 65 47 3 x a Eee nie nwnmey oe 22 771 ID Canbou . ...usvmenns 6.5 208 | Yer = 0 MER ! : 771 ID POWBK ...oonssmewns 6.5 6.9 325 IL KOR as insme smrmumnn 7.2 6.9 826 ID Elmore. ............ 13.9 13.9 92¢ he: Waltel. .ouuy snsmn es 7.2 20 830 ID Blaine. ............. 7.6 6.9 3 lk Mo Donough #30 wns 39 2 830 ID Camas............. 7.6 20.7 ARE win wade vi 5 h 833 ID Lemhir ooo, 29.0 5.7 328 I. Efingham .... cows wes 20.4 127 833 ID Custer 29.0 61.6 328 IL. Faves. . .c sos nmunss 20.4 19.9 Corry ’ ; 328 HE © JBEPBE iv vnie is wig ane» 20.4 47.6 251 I. Champaigh «connie ms 50 14 329 IL Rockisland .......... 12.6 10.0 281 IL Coles. ... cv couwms ens 6.0 4.3 329 Hh. HENRY. wwe sms min mmm 12.6 175 281 I. Douglas. ces wwsins «wa 6.0 13.3 329 WL Mercer 12.6 206 281 PE... eran 6.0 26.2 RIRE «coc uv is ten win 8 : 281 IL. Ford s.scnsvwnnsans 6.0 16.5 335 I. MBOOUPIN « «5 vs wiv vio 46.4 58.7 281 IL Cumberland. ......... 6.0 37.4 335 IL Montgomery ......... 46.4 27.3 285 I. Winnebago . ..«.c cau 77 3.2 340 IL Stephenson. ......... 32.6 25.8 285 I. Dekalb .....s0susan 77 15.4 340 WH. Cawoll .c.vovvmwsans 32.6 53.8 ps3 3 Og vE HEHE re vs 2 27 348 IL Jefferson. . .......... 24.9 215 ONG wt ems wiin nen ’ : 348 IL Wayne ............. 24.9 14.7 286 IL Madison............ 23.3 24.1 348 IL - Richland. ........... 24.9 24.1 286 I: Jersey ...wu cmon ome 23.3 41 348 IL Clay, .c: us rwsms wana 24.9 21.8 286 IL Greene. ............ 233 39.0 348 IL Hamion : « «ssw sss sn 249 50.6 286 IL Calhoum « cv ovo vv wis 23.3 12.4 348 IL EQWArdS . « : v4.0 0 5 nn ne 24.9 49.8 290 J WHE. canine ea de 41.5 44.2 350 IL Kankakee . «xs vio 5 owes 9.6 6.2 290 IL Grundy. ............ 41.5 11.8 350 IL HOGUOBIS: « x 4.4 2:05 v0 200 4 9.6 21.4 292 I. JACKSON vous sve ae nn 10.7 3.9 354 le. Malton s v0.0 svvwn aes 20.1 14.7 292 IL Williamson. .......... 10.7 5.1 354 IL Washington. ......... 20.1 36.9 292 ho BrOnRiR oc unens mune lo7 102 368 IL Whiteside ........... 25.9 24.0 292 L PBAY. .oompisnsnsnme 10.7 14.8 368 iC Jee 25.9 29.0 292 Ie Union. cose cosims cms 10.7 28 | “7 Troon ’ ’ 292 I JohnSOB.....uies 40s 10.7 48.0 369 IL. Vermilion. . . .ccoxvwnw 21.9 12.8 369 . EBdgar..:vienemsmsis 21.9 30.0 29 i gar AaB Eames 3) od 369 IN Fountain. ........... 21.9 437 297 IL Christian. ........... 5.1 20.2 569 IN VOITION «os sve s wns ad 441 297 Ie LOORD :rovns wiv m es me 51 16.8 372 He SEC avai vv nie 17.9 15.6 297 WH. 0388... consis woh de 5.1 12.9 372 IL Randolph... osu ve ans 17.9 13.9 297 IL Menard «iq «x 505 ve sen 5.1 1.2 372 IL LT PR 17.9 23.8 297 IL Sool. ...oasncniems 5.1 12.0 372 He. MOIVOB. ¢ iv vw x sa wnimn 17.9 40.0 302 IL Adams. ............ 105 2.0 37 Lo Bond..cesenims suse vs 523 302 MO Marion ............. 10.5 2.2 272 IN Vanderburgh ......... 8.7 0.7 302 Lh PRE. vocuisarmsvns 10.5 14.3 272 IN Wao. «.i.ovu sms whos 6.7 1.9 302 MO PRE... .oomrmnsmsmns 10.5 26.9 272 IN Gibson............. 6.7 5.1 302 MO Lewis.............. 10.5 8.5 272 Hh Saling..... sss snes 6.7 15.4 302 MO Bally ;.vivsaovomsans 10.5 12.9 272 IN- POSBY. .cv ivan amnna 6.7 0.9 302 MO Shelby ............. 10.5 49.7 272 IN SPBNBRr . «sw vv ww swims 6.7 59.1 302 i.- Brown csswsesswen ns 10.5 59.7 272 IL. White... .ovwsvnimus 6.7 13.3 308 IL LaSalle ............ 14.8 11.4 ore he Wabash + vewsanrurss id 15 i 272 HL Galan. ..... oncom: 6.7 6.6 308 IL. LIVINGSION ; vs osm wn 14.8 28.0 279 IL Hardin 6.7 333 308 IL Bureau............. 14.8 11.3 | #20200 METER eh skeen ! ! 308 iL PUINAM: 4 sw smn sos nm 14.8 83 276 IN Marion............. 1.8 0.5 309 Lh COOK. ans 14 14 276 IN Hamllon......ceeusx 1.8 3.9 276 IN JONSON. oo ws va wins wn 1.8 5.6 309 i. DuPage... iwswy sms 1.4 0.6 a 276 IN Hendricks ........... 1.8 1.8 309 WL: BBKSL. «vom ns omg 9h 26 1.4 0.6 276 IN. MOIGaN. conan nae ns 1.8 14.0 309 IL YANG vv 00 5 wh mw nn 1.4 1.4 276 IN Hancock. ........... 1.8 3.1 309 IL. McHemMy ....csowens 1.4 4.0 276 IN Boone 18 3.0 309 IL Kendall. ............ 1.4 9s | +7 0M EE ; 313 15 29 282 IN Bartholomew . ........ 9.4 5.6 282 IN JACKSON , uu va swis in sa 9.4 16.0 313 IL MCLEAN ..c cid sa vine wus 115 7.1 282 IN Jennings 9.4 8.6 313 IL Shelby ............. 15 63.4 8% it elim 2a ee i : 313 IL DeWitt. ............ 11.5 19.8 294 IN Tippecanoe . ......... 8.5 2.1 313 IL MOUlFI® «ow sows sms wi 5% 11.5 33.5 294 IN Clinton............. 8.5 16.1 300 IL ROBIE = ve evr ooze i 8.3 07 294 IN White. ... esensnams 8.5 16.1 294 IN Cafoll . ..onvvcvrnrns 8.5 16.9 322 I Tazewell .....:eneuss 8.3 3.1 204 IN Benton 85 76 322 i. FOR osm 2b 405 wo 8.3 220 | TY TEE Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked ——————————— unlinked SE tt ty area no. State County Area County area no. State County Area County 294 IN Warren. ............ 8.5 44.3 403 IN Putnam. ............ 41.4 41.4 295 IN Migoessnssnvmemnvas 11.0 2.2 539 KS Baling. .:..onsnsensn 8.5 27 295 IN Clay. ..c.vsvnsnssns 11.0 58 539 KS Dickinson ........... 8.5 19.7 295 IN- Sullivan. . com smvn eas 11.0 12.0 539 KS Elsworth. . :::e555, 4 8.5 26.0 295 IL Cla... ova swans 11.0 40.8 539 KS Oftawa............. 8.5 6.5 295 IN Parke; csverwsswsws 11.0 53.1 539 KS Linco. vuwssmwsies as 8.5 17.4 304 IN Alen .............. 4.5 1.7 545 KS Riley .............. 5.0 37 304 IN Noble. ..iuonssernas 4.5 22.4 545 KES Beary. soonsnmemsss 5.0 3.5 304 IN DeKab ............ 4.5 2.8 545 KS Pottawatomie . . ....... 5.0 13.6 304 IN Whitley. . .«.oscnsumss 4.5 7.0 545 KS CIaY. cs suismy wus ms an 5.0 12.3 304 IN Steuben ............ 4.5 11.9 549 KS Shawnee. ........... 8.6 2.7 310 IN Lawrence ... ous uu wes 17.6 14.8 549 KS Jefferson. ..: wv sus sv sn 8.6 325 310 IN| DUBOIS... . «oils wpm we 17.6 6.8 549 KS Osage ............. 8.6 28.2 310 IN Daviess... covscnsnes 17.6 27.1 549 KE Jackson . cuenew pase 8.6 16.0 310 IN OFanEB. .t cen ss wom mm 17.6 14.5 549 KS Wabaunsee. ......... 8.6 49.0 NS BN Blepasrrarsosnanse 2 57 553 KS Finney ............. 6.3 3.9 Ee a EEE ’ ’ 553 KS Kearny............. 6.3 5.7 315 IN Delaware. ........... 10.1 5.7 553 KS Haskell, ..:vvuouimwses 6.3 25.9 315 IN Randolph ........... 10.1 20.2 553 KS Hamilton. ........... 6.3 20.0 315 IN Blackford. ...«. cuss. 10.1 25.8 558 KS Ford .. o.oo. 10.4 5.9 318 IN Madison............ 10.7 10.6 558 KS Gray .............. 12.4 35.4 318 IN Howard... ..coscmsimss 10.7 3.0 558 KE: KIOWA. . : ss ss cnims va 12.4 25.7 318 IN Miami. ............. 10.7 225 558 KS Clark .............. 12.4 13.6 318 IN. TIBION ue ws menoswns 10.7 21.0 558 KS Comanche. . uw: wasn 12.4 7.5 330 IN St Joseph. .......... 5.2 41 558 KS Hodgeman .......... 12.4 20.6 330 IN Marshall. ........... 52 11.5 568 KS Johnson. .::s sess 21.9 29.5 332 IN Ripley. . o.oo... 438 46.5 568 KS “WYandofie. ...v seu sus 215 136 332 IN Decatur. . . .. 43.8 40.8 568 KS Leavenworth ......... 21.9 8.3 568 KS Miami.............. 21.9 17.1 333 IN Lake .............. 4.8 3.2 568 KS Linn.ssws covma anew 21.9 58.2 9% IN POHL snr rmemeirmen is e7 570 KS Seward. ............ 11.3 5.4 333 IN Jasper............. 4.8 16.9 333 IN Newton. .... .... .... 48 26.7 570 OK BBaYBr: us cvsvun sums 11.3 24.3 : 570 KS Stevens... :.asivoxs 11.3 19.2 337 IN BOX; cau: wns ms wie 16.5 10.8 570 KS Meade ............. 11.3 18.6 ey % J ous ow ain 4 7 078 i i 574 KS Grant. ............. 29.2 34.0 Crore B ’ 574 CO Baca .............. 29.2 23.6 338 IN Mone. wu. vuvmn owas 17.4 4.4 574 KS Stanton. :: uses vsews 29.2 20.2 338 IN Greens. ..:.viav usa 17.4 30.6 578 KS Elis... 71 21 338 IN Owen......coinvimasn 17.4 27.8 338 IN Brown 17.4 64.6 578 K8 RUSSBH.....coeom own 71 18.2 Toy ! 578 KS ROOKS . uv omsnmons 71 8.8 351 IN Elkhart... .. ccm saan 14.4 9.8 578 KS Trego.............. 7.1 10.9 351 IN Kosciusko... ... uous 14.4 20.7 578 KS Grahath .. «Gown smewa 7.1 19.3 351 IN Lagrange ........... 14.4 22.7 586 KS Sedgwick ........... 24 12 360 IN Jefferson. ........... 37.1 25.3 586 KS Bufler...ssssssswsne 2.4 6.9 360 KY Cartoll .vos55 550s nas 37.1 65.8 586 KS Sumner ............ 24 9.4 360 KY Trimble............. 37.1 44.2 586 KS Prfl cover vvnssvvns 2.4 14.1 361 IN Cass .....oovvvnn.. 25.2 225 586 KS Kingman............ 24 126 361 IN FUHOO «ooo vo dt sti wi 25.2 28.3 589 $8 Bardon ..:usvessvvens 9.4 5.3 361 IN Pulaski..cocninvcnse 25.2 28.2 589 KS Pawnee ............ 9.4 4.1 371 IN Adams . . o.oo. 18.5 16.0 589 KS Stafford. . ........... 9.4 27.4 589 KS. Rush... :ssnsumems 9.4 29.7 371 IN WBIB uu us vars wwii 18.5 17.2 89 KS Edward ou 18.1 371 IN Jay ooeeeeeeeeen 18.5 23.9 5 HOEISL lsh vedi i : 7 597 KS Mitchell. .: cusnssnsiws 22.7 18.3 379 IN" LaPoRB , «usw vows 14.9 10.4 379 IN Starke . . ........... 14.9 36.6 597 KS Osbomne............ 22.7 29.0 384 IN Grant. ............. 13.2 11.8 22 is Pavey PRUE TY Hey A hig 384 IN Wabash ............ 13.2 16.1 0 BHO i mn 1% 8 Gls 386 IN Shelby ............. 51.0 52.2 606 KS Thomas ."..coasswews 15.3 10.7 386 IN RUSH «ooo... 51.0 48.3 606 KE Gove.............. 153 275 606 KS Logan ............. 15.3 9.3 391 IN Wayne............. 8.8 8.8 606 KS Sheridan. . ooo sueims 15.3 20.8 392 IN HEY. svc vimsmnswen 36.6 36.6 611 KS Sherman............ 32.8 32.1 395 IN Montgomery ......... 30.8 30.8 611 KS Cheyenne........... 32.8 34.4 : 614 KS Phillips. ............ 18.4 172.5 $8 4 WR .0 5. 936 IN © Huntingion 8s 554 614 KS NOHON . ccvwrws vmsms 18.4 13.0 82 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked a ii ss unlinked ———————————— area no. State County Area County area no. State County Area County 614 KS Smils:ssssvmeecanns 18.4 26.9 22 KY Lincoin............. 272 20.1 615 KS RENO .............. 14.8 10.7 2 Ny Josey pe anzad ensues a ii 615 KS McPherson . ......... 14.8 238 BHR erent eters Ho : ? 615 KS RICE. «vn nu mows wv inn 14.8 16.3 31 KY Calloway. i «is. usms on 20.7 7.2 616 KS Montgomery ......... 25.4 20.8 a IN Hof cxmeusaninpis Sy 22 616 KS Wilson ............. 25.4 8S | TT FEET : : 616 K8- BIR + oovmpinismbrmen 25.4 62.0 33 KY McCracken. ......... 13.3 2.8 617 KS SGOft.............. 219 18.8 23 RY Moshilowcusmermans 138 $44 TL 33 IL. Massst. . i... onvmens 13.3 10.2 617 KS Whit. «wv mw smal vous 21.9 14.5 To 33 KY LvINGSION . «5s von vas 13.3 4.8 617 KS Waliage, . vues wwii 21.9 53.6 : 617 KS Greeley 219 92 33 RY ChHenden. .s conve 13.3 14.4 Torry ! 33 KY Ballard... ..00vv009: + 13.3 8.4 625 KS" Lyon zones sav we man 11.4 3.3 33 KY LYON cos ounniannsvmes 13.3 49.7 625 KS Coffey . oon wn vw wus 11.4 23.7 33 KY Carlisle. ............ 13.3 26.0 625 KS Greenwood .......... 11.4 23.5 33 I. POPE wy ums wnsws nada 133 34.0 oes Ps Yer ht EE ne 3 a4 KY Bell ..oooovonnnn... 26.4 11.3 Torry ! : 44 TN Claibome ..oe serie so 26.4 49.8 028 KS Douglas ......ee.v'ss 152 as 45 KY Rowan ............. 33.4 12.7 628 KS Franklin us ccc nwsw 15.3 20.2 628 KS Andeteon 15.3 20.1 45 KY MaSON : vrvsnmansne 33.4 16.9 A: ’ ! 45 KY. FIBMIAG «vv eminem ona 33.4 17.0 633 KS ‘Labelte.. . . «vv ve msin win 14.9 16.2 45 KY MOIOBN. wm ow vik w vou 33.4 44.0 633 KS Neosho. ......:.r:v2. 14.9 11.2 45 KY Bathicuvincinssnies 33.4 41.9 633 KS AGH ....vowwsvvivuesy 14.9 15.2 45 KY Bracken ............ 33.4 54.7 633 KS Woodson ......«:wve 14.9 22.8 45 KY WolB.:::uninsensmess 33.4 72.0 636 KS Marshall ............ 34.8 315 po i ox woo an wk ER Ss Se 636 KS Washington . ......... 34.8 41.8 Lo REAR LEE EERE ; “a 641 KS Crawford. ........... 14.9 19.0 48 KY HUD. wuvanwrmnss 152 $3 641 KS Bourbon. ........... 14.9 6.3 45 KY: Meade, cusms smomems 102 27.1 46 KY Grayson ............ 15.2 14.3 648 KS .COMBY... .. sosmn oman 16.6 12.0 46 KY Breckinridge ......... 15.2 34.6 648 KS Chautauqua. ......... 16.6 54.5 46 KY Larue. ............. 15.2 16.7 656 KB Cloud. «ovcovnwiwsss 21.8 24.8 47 KY Laurel c:vunvvuansas 28.0 26.6 656 KS Republic. ........... 21.8 16.3 47 KY Whitley. ............ 28.0 8.1 668 KS Atchison . . .......... 20.6 113 47 TN ‘Campbell cue vs unas 28.0 55.5 668 KS Brown ............. 20.6 34.6 47 KY KOOX oo vr vmnms vans 28.0 24.2 672 KS Decatur. ............ 26.7 2256 as RY Phd. «rensnransonnn bid 100 672 KS Rawlins. ............ 26.7 31.6 03 BY FOV cuvrwsvmewrns . 138 5.1 63 KY JONNSOM... cwsiviv ams vs 13.8 20.0 674 KS Harper............. 15.1 17.1 63 KY Knott.............. 13.8 45.0 674 KS Barber............. 15.1 13.1 63 KY" MAgofiif.v ve sme enon 13.8 14.3 678 KS Nemaha cou su vivs si 18.2 18.2 83 KY Perry.............. 24.6 20.0 688 KS Ness .............. 35.1 20.1 83 KY Breathitt............ 24.8 43.8 688 KS Lane ...........u.o.. 35.1 66.4 83 KY LOBE: 115. 4 wise vrei 24.6 18.2 3 KY Warren. ............ 7.4 4.2 98 KY CiMon.svsvasssvwns 45.1 46.2 3 KY Barren ............. 7.4 21 98 KY Cumberland. ......... 45.1 43.6 3 KY Ha. .: covers nmrwna 7.4 21.7 100 KY Logan ............. 58.8 58.8 3 KY Alen .............. 7.4 6.4 ) 3 KY Simpson. ........... 7.4 15.2 101 KY Hopkins ............ 11.1 10.2 3 KY Monroe. . ........... 7.4 5.2 101 KY Muhlenberg. ......... 11.1 12.5 3 KY" Buller. .:us ss vws ews 7.4 16.2 118 RY Clay. civisnonmssnsns 39.7 35.0 3 KY Edmonson. ....o.sos sas 74 12.9 118 KY Owsley. ............ 39.7 61.1 5 (Y MSeafS.. w «yar muse 4 5 124 KY Haran ............. 14.1 14.1 4 KY Pulaski. ...ooooswins 10.9 5.6 a KY Taylor. ............. 10.9 9.2 125 KY Monigomery ........ou. 43.5 43.5 4 KY Wayne .......evseen 10.9 15.2 134 KY Henderson .......... 19.0 13.5 4 KY McCreary. ........u. 10.9 4.4 134 KY UPIOA.. «vvms mn sms es 19.0 10.6 4 KY Ada, onus amas mee 10.9 19.5 134 RY Webster .o:uevs cose 19.0 46.6 : Ny fuss am amet 2822 iss 730 189 KY Graves............. 265 56.8 trons : : 189 TN ObIon. : ws soizws sors 26.5 5.6 16 KY MadiBON ....«vusvmsi 34.0 26.3 189 TN Weakley . ........... 26.5 18.2 16 KY ES .....ocomsmwaes 34.0 19.0 189 RY: FURON wins vmammans 26.5 55 16 KY Rockcastle. . ......... 34.0 53.2 189 TN. lake cuacpnams emai 26.5 39.6 16 KY JacKsSON .. .....svoimu 34.0 49.1 189 KY HICKIBA « o vvn 4 2 wis 0 4 26.5 23.9 18 KY: M08 urasannnsmonmen 54.0 478 271 KY Fayette. ............ 2.7 1.1 22 KY: BOYR. ..vumu dawns oils 27.2 155 271 RY. Jessaming. - «sums vos 2.7 1.5 22 KY MBICBY + so0 tw +s oo wins 27.2 22.0 271 RY Clatk are 505 00 5 2 1 sean 2.7 2.6 83 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked unlinked area no. State County Area County area no. State County Area County 271 WY Scolt...::oinsms ows 27 5.0 445 LA St. John the Baptist. . . . . 2.1 8.7 271 KY Woodford ........... 27 58 445 LA Plaquemines ......... 23 1.7 7 KY BoUMON.s enemys zuma 27 29 446 LA Lincoln. ............ 50.2 51.3 27 KY Hamison .. . cone snes 27 1.6 446 LA Jackson ............ 50.2 33.4 271 KY Powell ............. 2.7 22.4 446 LA Winn 50.2 64.4 271 KY Nicholas. ....«::ivss 27 A ’ : 27 KY Robertson........... 2.7 21.2 451 LA Lafayette. ........... 2.2 0.9 287 KY Jefferson. ........... 3.4 13 451 lA SELaIY swe vr saves 22 24 451 LA Berd: evans vnnee 22 1.3 287 IN Clark. .ournmos noma 3.4 1.6 : 451 LA ACER... ones upnas 2.2 3.5 287 IN Floyo ooo cmsmmn man 3.4 0.8 Hi 4 451 LA Vermilion. . ..oxc uses 22 1.3 287 KY BU .oo0 a5 mm smn » 3.4 5.6 7 451 LA SEMEN criss» 2.2 1.7 287 KY Okham ...co05 sumnns 3.4 1.2 451 A Evangeline 29 10.8 287 IN Harrison ............ 3.4 1.7 BONS 70 0 ari : 4 287 IN Washington. ......... 3.4 11.3 459 LA Calcasieu ........... 8.6 1.4 287 WN Sool. .u:cermiamans 3.4 21.6 459 LA VOOM. vv vw svi soi 8.6 14.3 287 KY Hey. co: cmv soon 3.4 52.3 459 LA Beauregard .......... 8.6 2.4 287 IN Crawford. . .. ws smsms 3.4 44.9 459 LA Jefferson Davis. . ...... 8.6 29.0 287 KY Spencer............ 3.4 40.2 459 LA- AIBN oz snsnzsmenss 8.6 26.9 317 KY Franklin oo 355 25.9 459 LA Cameron. ........... 8.6 5.1 317 KY Shelby............. 35.5 39.8 469 LA SL. Tammany . ...: «ax 31.1 39.9 317 KY Anderson . ...s:sss sas 35.5 45.1 469 LA Washington. ......... 31.1 10.5 317 KY OWBHN.v.vsmenss mn 35.5 62.5 469 MS PeariRiVer. ...:..:v0x 31.1 215 377 KY Daviess. ....onxssapwws 8.7 2.8 518 LA Tangipahoa.......... 45.4 17.4 377 KY SONI +: 4 4t 5.0 51 00h om: 8.7 9.0 518 LA. Livingston. «oe omen en 45.4 84.5 377 IN POY «oon enim sme 87 36.4 518 LA StHelena........... 45.4 28.5 id Me win SEA LET AE 37 212 28 MA Middlesex . .......... 45 5.8 BACHE 2 we ie len 3 : 28 MA Suffolk ............. 45 1.2 390 KY Nelson... wins ening 43.9 59.0 28 MA Norfolk. . ...conue swiss 45 5.0 390 KY Marion............. 43.9 18.2 28 MA Plymouth. aco nvr wes 4.5 55 390 KY Washington. ......... 43.9 42.3 48 MA Hampden ........... 25 21 411 LA East Baton Rouge . . . . .. 2.4 1.3 48 MA Hampshire. . . voc os vse 25 4.2 1h #1 Asp van aus ve 24 71 88 MA Essex.............. 19.8 19.0 BOING ose on 2 bre dock mss : ! 88 NH Rockingham. ......... 19.8 28.0 411 LA Pointe Coupee . . ...... 24 4.1 88 NH Strafford 19.8 5.7 411 LA West Baton Rouge. . . . . . 2.4 co | 0 TTT ’ ’ 411 LA East Feliciana. . . ...... 2.4 8.8 89 MA Berkshire ........... 15.9 4.1 411 LA West Feliciana . . ...... 2.4 6.1 89 NY Columbia ...ovuseivn 15.9 29.5 412 LA Caddo ............. 2.8 0.7 5 NY B188h8......v nv init iso 345 412 LA Bossier. ............ 2.8 0.6 99 MA Worcester . . ......... 14.4 13.5 412 LA Webster ............ 2.8 1.0 99 MA Franklin . ....«ccoruse 14.4 246 472 LA, NOUR it 4 4 00s 28 i390 102 WA BHA wo ws nin ssn 22.6 23.9 412 LA D8 S03 .usvsnssnnns 25 15 102 RI Newport. ........... 22.6 14.9 412 LA Sabine............. 2.8 52 412 LA Claiborne ........... 28 5.2 121 MA Bamstable. . .....:4.5 22.6 22.6 412 LA Biswils.... sonnn smn 28 17.8 135 MA Nantucket. .......... 8.9 8.9 412 LA BedRWVer ...uies vss 2.8 0.5 136 MA Dukes ...::r 050s 0 7.8 7.8 418 LA Ouachita. ........... 4.2 1.6 418 LA Morehouse . ......... 4.2 13 27 MD Montgomery ......... 35 3.2 418 LA Franklin ............ 4.2 11.1 27 MD Prince Georges. . . . .. .. 35 5.0 418 LA Richland... ......... 4.2 1.6 27 DC The District . ......... 35 1.8 418 EA: Um, «oon cis mb sm 4.2 18.1 27 MD Charles. ............ 35 4.4 418 LA WestCarroll. ......... 4.2 3.9 30 MD Baltimore ........... 6.0 0.7 418 LA Caldwell ............ 4.2 2.8 30 MD Anne Arundel. . ....... 6.0 18.8 418 LA EastCarroll .......... 4.2 14.4 30 MD Harford. . ........... 6.0 15 424 LA Rapides ............ 5.6 2.2 30 MD Howard. .....sxs «vss 6.0 24.3 424 LA Avoyelles ........... 5.6 9.9 30 MD Carroll ............. 6.0 13.1 424 IA Graf. cos mame nme 5.6 6.7 70 MD Allegany ............ 7.7 3.1 424 LA LaSalle ............ 5.6 24.8 70 WV Mineral. «ooo zy 3.1 436 LA Terrebonne .......... 12.9 2.4 70 MD Garrett. ............ 7.7 17.9 436 LA Lafourche ........... 12.9 3.4 70 WV Grant. ............. 7.7 22.7 436 LA St. Mary SEE PEA ge 12.9 26.9 91 MD Queen Annes. ........ 16.1 35.6 436 LA Assumption. ......... 12.9 18.7 91 MD Dorchester. . .. ....... 16.1 8.7 436 LA StJames........... 12.9 52.4 91 MD Talbot. ............. 16.1 5.7 445 LA Orleans... cu .ms umm 2.1 0.4 91 MD Caroline ............ 16.1 12.5 445 LA Jefferson. ........... 2.1 17 91 MD Kent............... 16.1 14.7 445 LA StBemad.......... 21 0.7 94 MD St Marys ........... 23.8 19.9 445 LA -8t.Chanes . .. .u «ui wa 241 19.7 94 MD Calvert. ............ 23.8 30.9 84 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of— 800- by residents of — unlinked unlinked area no. State County Area County area no. State County Area County 114 MD: Freqeriok. .. . + « v5 00s 5s 37.1 37.1 323 MI Roscommon ......... 47.2 64.7 " WE Troost uo ox «ese. 0 " 323 MI Crawford. ........... 47.2 25.7 11 ME Hancock. ....:vv sun 3.0 1.3 336 Mi Houghton o. «iv. we vuisiv 3.8 2.6 11 ME Washington.......... 3.0 4.0 336 M -Baraga............. 3.8 8.4 11 ME Piscataquis . ......... 3.0 24 336 MI Keweenaw. .......... 3.8 5.4 23 ME Kennebec........... 15.1 14.5 339 MI Montcalm ........... 22.3 29.8 23 ME Somerset ........... 18.1 18.3 339 Ml Gratioh. us cnemas vw on 22.3 11.9 50 ME Cumberland. . ........ 13.9 3.3 341 Ml SL.CIRIE ...wxsnsvmis 24.0 22.4 50 ME York... ............ 13.9 29.2 341 Ml Sanilac. ............ 24.0 30.0 2 Ne Segadance ol 2 A en hi 352 MI Midland «o.oo... 15.2 8.7 COM: wuz nz isis rw " 3 352 MI Isabella... .......... 15.2 26.9 96 ME Androscoggin. . ....... 12.0 10.5 352 Ml Clare... cocveswrens 15.2 11.3 9% ME Oxford ............. 12.0 19.0 352 Ml Gladwin ............ 15.2 16.5 96 MB FoOiR w.avenenmiing 129 5.5 355 MI Muskegon. .......... 155 8.8 112 ME KNOX .............. 26.6 12.1 355 MI Newaygo. ........... 15.5 42.0 112 ME Waldo ............. 26.6 433 355 MI Oceana. ............ 155 18.5 132 ME Aroostook... ........ 3.5 3.5 357 Ml DERE voor vpn on 5.6 5.1 273 MI Grand Traverse. . . . . ... 25 1.4 id Wy Schools SASS EEA 2 2 2 273 Ml Leelanau. ........... 25 0.7 HOE: cncsive + Fazpnity Fait gi : 273 Ml Kalkaska............ 2.5 7.5 370 MI Gogebic............ 14.8 11.2 273 Ml Benzie............. 25 5.0 370 WI Ashland ............ 14.8 14.2 274 Mi Ingham. ............ 49 3.1 3 hi Joyisy SEBS Sn ws wen pu 2 us 274 MI Eaton.............. 49 Z4 | FEF RRR in ma ais ie Eu rR 3 274 MI CIIMON . os coves vitms 49 10.5 374 MI SLJOSEPN. , sisk v0 sme 25.3 26.8 374 MI Hillsdale . ........... 25.3 25.5 298 Mi DMG, crea enaer res 108 08 374 MI Branch. ............ 25.3 228 293 MI Cheboygan .......... 16.6 4.0 293 MI Charlevoix. . ......... 16.6 2.9 383 ML. Mason. ...o«ssvass 13.9 4.2 293 MI Antrim Le 16.6 67.5 383 Ml Manistee. ........... 13.9 26.7 293 Mi MagREE ov sesne dium 156 2s 385 MI Calhoun ............ 18.2 9.4 296 Ml Alpena. ............ 21.3 6.1 385 ML Bamy .............. 18.2 43.6 296 M OIS800. ....0onorms un 21.3 32.4 296 MI Presque lsle. . ........ 21.3 32.0 33 ML MOB. «nsx uss inns 5.2 9.2 296 MI Alcona ............. 21.3 37.8 398 M Chippewa . .. «:rusnxo 13.0 13.0 296 MI Montmorency. . ....... 21.3 17.0 400 MW Huron 19.6 19.6 298 MI Mecosta . ........... 18.0 27.7 : 1 Ml Jackson ............ 4 298 MI Wexford . ........... 18.0 16.2 40 WI Jackson 172 72 298 MI OSCEOR . .. 5555553 18.0 7.5 402 Ml Lenawee. ........... 18.5 18.5 298 MI Missaukee. .......... 18.0 10.1 835 MI Ontonagon .......... 38.7 38.7 298 Ml Lake .............. 18.0 26.3 280 MN St Louis. ........... 1.0 0.8 303 MI Kalamazoo. .......... 8.7 2.8 280 WI Douglas ............ 1.0 21 90 I Boon. wu suan same 27 29 280 MN Cariton............. 1.0 0.8 303 MI VanBuren........... 8.7 8.2 280 MN Lake «ooo, 1.0 0.7 303 M Cassusncsmsnsanens 87 343 | 280 MN COOK... ovvrnnnn.. 10 25 305 MI Wayne ............. 1.0 0.4 289 MN Hennepin ........... 73 3.5 305 MI Oakland ............ 1.0 1.7 289 MN Anoka ............. 7.3 10.1 305 MI Macomb............ 1.0 07 289 MN Wright ............. 73 17.2 305 MI Washtenaw . ......... 1.0 1.5 289 MN Carver ............. 73 29.2 805 ME LVAgSton. « swnxs sons 10 23 289 MN Sherburne. .......... 7.3 41.4 306 MI Genesee. ........... 5.9 15 289 MN Isanti.............. 7.3 14.0 200 MSN sx sn enna 3 32 312 WN BORG. 1 50s 050 0 0 47.9 59.3 2s Ml LOB. nnn nna 59 200 312 MN Scott .............. 47.9 26.4 306 MI Shiawassee... ....... 89 25.1 312 MN RICE. . oo oo, 47.9 15.7 806 MI TUBER, oa uisins corms 59 20 312 MN LeSueur............ 47.9 40.4 307 Mo Bay. .............. 17.3 22.4 321 MS AITO cn somes mun 22.1 25.8 307 Ml 10SCO. «veer 17.3 5.2 201 MN Washington . . . . ...... ad 25 307 MI Ogemaw............ 17.3 71 321 WI St.CroX. ........... 24.1 12.1 307 Mi Aenac............. 17.3 10.2 321 WI Pierce ............. 24.1 52.0 307 MI Oscoda. ............ 17.3 43.7 01 MN Chisago . . «oo... 1 op mn Mi Marquette . .......... 22 1.5 364 MN Winona. ............ 428 345 31 MI Alger.............. 2.2 9.1 364 WI Buffalo . ............ 42.8 69.2 319 ML Kent. ............. 6.6 13 58 MN Beltrami « «ven... 53 3.1 319 M Oftawal; .. cosas savws 6.6 3.6 566 MN Hubbard. ........... 5.3 57 319 MI Allegan............. 6.6 23.4 566 MN Clearwater. . . ........ 53 14.8 319 Ml Jonia imi anime se mms 6.6 41.2 85 Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked ———————————————— unlinked ————————— area no. State County Area County area no. State County Area County 576 MN OImsted . cous nme 8.0 0.8 324 MO Cape Girardea . . ...... 12.7 3.0 576 MN Goodhue. .......:::: 8.0 14.1 324 MO Scot... .vvirsvenwas 17.7 6.0 576 MN Fillmore ............ 8.0 38.7 324 MO New Madrid. ......... 17.7 557 576 MN Wabasha. . .......... 8.0 25 324 MD POI ; v5 ws annww vi ni 7.7 15.9 576 MN DOdEe : + vnivnsms sn 8.0 13.8 324 MO Mississippi. . ......... 17.7 83 583 MN Becker. ............ 27.3 27.6 Le Sanne: “waren 77 302 583 MN Mahnomen .......... 27.3 25.8 304 IL Pulaski. 17.7 50.1 7 MN Case. 204 sor | 8 MO Jaer........... 8s 48 543 MO Newton. ............ 8.5 9.7 588 MN Noble. . «row vimenmses 18.0 17.7 543 KS Cherokee . uve s ens ns 8.5 13.4 588 MN Cottonwood. . ........ 18.0 14.7 543 MO Barton ............. 8.5 23.9 = IN Jaks: seams vaasn 120 ii 544 MO Phelps ............. 15.3 95 HEY Jo v.05 50 2 00 0 wim 18.0 21.7 588 A Osceola . ... 18.0 17.8 544 MO Crawford. ........... 15.3 29.5 544 MO. Dent : coms: srame vans 153 10.3 598 MN OtterTail. ........... 19.6 15.8 598 MN Wadena ............ 19.6 23.8 350 MO 010608... su suesmsmns 70 11 598 MN Grant. ............. 19.6 40.2 $50 MO LOWES 1 uirnn surmss 70 8.5 550 MO Christian. ........... 7.0 1.4 610 MN Kandiyohi ........... 7.9 6.0 550 MO Baty i .v:vwavsvurnans 7.0 26.6 610 MN Chippewa . .......... 7.9 23 550 MO Laclede. ............ 7.0 14.2 610 MN Yellow Medici. . ....... 7.9 15.7 550 MO: TaNBY... «ic «vow x make 7.0 3.6 610 MN SWIRL ovum sae es 7.9 9.3 550 MO WEDS & cco sw sous mn 7.0 6.0 610 MN Lac QuiParle......... 7.9 11.8 550 MO Polk. .............. 7.0 3.1 612 MN Douglas ............ 9.7 7.0 550 MO SIN8. 4 uueverunuvnn 70 41 612 MN POPE . ooo ooeeeee 9.7 16.8 550 MO Dallas. .....ccveuns. 7.0 10.3 550 MO Douglas s «x vues snes 7.0 Bia 613 MN Ping. ..ovvnivsn an sms 50.8 58.4 550 MEO: ‘DAAB . i vw sv ies 7.0 20.9 613 MN Kanabec............ 50.8 38.6 550 MO Hickory. ............ 7.0 31.1 622 MN Roseau............. 16.5 17.8 555 MO SLLOUS: .5:6: 05555 15 0.7 622 MN Lake of the Woods . . . .. 16.5 11.6 555 MO St louisCit. ......... 1.5 1.6 624 MN Freeborn. ........... 15.8 12.1 855 MO 8t.Charles .......... 1.8 0:8 : 555 MO Jefferson. . .: .s:+4555> 1.5 1.3 624 MN Faribault... cc co:vx sn 15.8 22.4 555 MO Frankin 15 61 631 MN LYON cs ps eennt vanes 15.3 11.7 555 MO Lincoln. ............ 15 2.7 631 MN Lincoln... ....vsueen 15.3 30.8 555 MO Warren. ............ 1.5 27 634 MN Blue Earth. .......... 11.6 5.2 535 MO GoS00REER «=v us suk iE $3 se 634 MN Nicollet. . ........... 11.6 18.0 559 MO “BOOB itv nssaiiss 6.8 1.6 634 MN Waseca ....coccnus» 11.8 18.6 559 MO Randolph ........... 6.8 1.3 634 MN Watonwan. .......... 11.6 13.2 559 MO Aural, vss vm sma 6.8 5.2 644 MN Brown . ............ 20.9 8.3 559 MO Cooper wt 50 | wt 2 vt Jed es 5 0 0 6.8 15.3 644 MN Renville. . ........... 20.9 35.4 559 MO Linn............... 6.8 17.7 644 MN Redwood ........... 20.9 25.9 559 MO Montgomery ......... 6.8 34.5 559 MO. Charlton ws vs os vw ss ais 6.8 2.6 651 MN Mcleod. ........... 36.0 239 559 MO HOWE, «555 0 05 50 5 5 6.8 1.0 651 MN Meeker. ............ 36.0 41.7 559 MO Monroe. ............ 6.8 a.7 651 MN Sibley... sneer snnnss 36.0 55.9 ) 577 MO Adair. ..ssnsssevnss 17.6 4.2 658 MN Stearns. ............ 11.6 6.8 577 MO Macon ............. 17.6 37.6 658 MN Morrison. . .......... 11.6 8.3 577 MO Salar, . .v vv vs wiv sn 17.6 223 658 MN Bemon......ovow ems 11.6 7.9 577 MO Putnam. ............ 17.6 21.7 658 MN Todd .............. 11.6 43.1 577 MO Scotland. ........... 17.6 17.0 664 MN Martin. . .... oo .. 16.9 9.6 B77 MO KHOR vs vweims smswe » 17.6 22.6 664 A ROSSUINY ...vvu ive us 16.9 25.6 577 MO Schuyler. ........... 17.6 10.0 664 IA Emmet............. 16.9 17.2 580 MO. JARED... oi 40505 0m» 8.3 6.8 664 IA PaloAld. ...cruemsns 16.9 17.2 580 MO Clay. .............. 8.3 5.1 666 MN ltasca.............. 20.4 17.0 580 MO Cass .............. 8.3 13.8 666 MN Aitkin. «oe 20.4 33.7 580 MO Palle... venvsvonss 8.3 15.8 580 MO. REY .... +o sis Bois 2614 £0 8.3 8.4 669 MN Mower............. 22.5 21.8 580 MO Clinton. ............ 8.3 34.4 669 IA MEE 5 vw sine mvs 22.5 25.8 580 MO BABE. ....uusvinsss 8.3 37.4 670 MN Pipestone . .......... 21.4 24.8 585 MO Livingston . .......... 25.0 14.8 670 MN ROCK = gw amis nis misma 21.4 17.7 585 MO Grundy. ............ 25.0 9.2 680 MN Millelacs ........... 22.1 22:4 585 MO Daviess. ............ 25.0 26.9 585 MO Caldwell . ........... 25.0 50.8 683 MN Steele ............. 28.6 28.6 585 MO Mercer. ............ 25.0 48.9 685 MN Stevens ........+w054 13.2 13.2 590 MO St Francois. ......... 18.7 18.4 692 MN Koochiching. . . ....... 15.3 15.3 590 MO Washington. ......... 187 24.3 590 MO Madison. ........... 18.7 9.7 86 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked ai unlinked i area no. State County Area County area no. State County Area County 590 MO. ON... + vs wn 5 50s 4s ww 18.7 18.2 417 AL Choctaw. ........... 16.0 25.2 593 MO Butler... ........... 19.8 24 417 Al. Sumter. ... srs ums 16.0 30.2 X 417 MS Clarke ............. 16.0 26.3 593 MO Dunklin. ............ 19.8 32.2 47 MS Kemper 16.0 9.1 593 MO Stoddard. . ...u:vvsman 19.8 803 | FT 200 TERE aera : ’ 593 MO RIDIBY. . : vornssninse 19.8 8.8 420 MS Pike. ..........c... 21.1 8.0 593 MO. Wayne. . coun smmaim en 19.8 21.0 420 MS Lincoln. «oem onsimsme 21.4 25.5 593 MO Reynolds. . : ccccncwns 19.8 47.4 420 MS Walthall. . ........... 21.4 10.5 593 MO Carter... ........... 19.8 6.1 420 MS Amie: .:vsvsvsewans 24 40.8 601 MO Cole. ooo, 30.1 21.0 420 MS Lawrence ........... 21. 38.7 601 MO Callaway. ........... 30.1 45.7 423 MS Coahoma ... wo rus ne =x 9.6 5.9 601 MO Camden . «sus swiss « 30.1 37 423 MS Quitman............ 9.6 9.0 601 MO Miller. ............. 30.1 18.3 423 ME TuniCal. ; «ins imsmmpes 9.6 24.3 a0} MO Morgah. ..oomemrsnes 90.9 $8.5 431 MS Lafayette. ........... 26.2 38.6 601 MO Moniteau. . .......... 30.1 25.0 431 MS Grenada 26.0 75 601 MO 08808 . . ..onvmvnwn vs 30.1 14.8 431 MS Cahoun . 26.2 408 601 MO Mares .:c-s53 55x00 30.1 BOIS Nan. Nee Nebduce ren ’ : 431 MS Yalobusha........... 26.2 18.8 2 2 Hoy “os smne up nie 2 | ns 444 MS Harrison ............ 10.4 48 AMIR BEN 0 ded 8 OR An ’ : 444 MS Hancock... ..:uw.veew 10.4 25.3 604 MO Buchanan ........... 10.2 6.5 444 MS Stone........s:45.5 10.4 73.8 604 MO NoJaWay. cu. vv omnwe 102 27 447 MS Monroe. ............ 34.4 19.0 604 MO ANCIEW. . .. vu vu sv wn 10.2 3.2 447 AL Marion 34.4 54.6 604 KS Doniphan . ...cumaveis 10.2 35.9 447 AL Fayette. ’ : Welle. . vv vue mswmm 34.4 45.8 604 MO DaKalb ..vvw opus su 10.2 32.2 447 AL Lavan 344 292 604 MO AIChISON . «cvs nasms ns 10.2 2%: J A & ’ 604 MO Gently . .... «ocmwwnsus 10.2 7.2 449 MS 188. ..:vasuremennsy 11.5 27 604 MO HOR... ocsumessama 10.2 4.9 449 MS Acorn ............. 11.5 17.4 604 MO Worth. ............. 10.2 21.4 449 MS Prentiss ....ocnmemsn 11.5 2.0 621 MO Vernon. ............ 32.3 31.7 ie MS UBIO. cue smemn ines 1s 26 621 MO Cedar. 323 333 449 MS PontoloC.. ...ueunss 11.5 4.5 449 MS Hawamba. .:.-:.usemss 115 22.5 632 MO Howell . cm inssms sen 24.9 14.7 449 MS Tippah ............. 11.5 8.6 632 AR FUROR . ovo immm wn » ois 24.9 23.4 449 MS Chickasaw. . ....:s. s+ 11.85 34.4 632 MO OF8gon: «csi s Sa se ms 24.9 17.3 449 MS Tishomingo... «cu «vie 115 24.9 632 MO Ozaki. oo ow one ns sms 24.9 63.9 449 MS -Benlory. coos ww sms mus 115 29.0 32 MO Shannon... ssusxs sus 28 560 452 MS Oktibbeha. .......... 13.1 9.4 635 MO Johnson. ........... 40.6 24.5 452 MS Winston .....cc.o0n+ 13.1 20.0 635 MO Lafayette. ...cccms ces 40.6 56.1 452 MS Webster ............ 13.1 11.5 635 MOD Card vs cov omaman 40.6 68.4 452 MS Choctaw. ........u:« 13.1 14.4 639 MO. -Peltls . «vos » vam ums 81.7 17.0 454 MS: AGES. vous snsae ins 21.2 7.5 639 MO Saline. ......co00..04 31.7 46.8 454 LA Concordia. « vu. wo wv mwns 21.2 17.3 639 MO Banton. .....osvusss 31.7 50.2 454 JA Catahoula... ...o604+ 21.2 37.3 654 MO Harrison . ........... 38.9 55.3 454 WS JBTOINic 4 we one rs 21.2 2758 454 MS Franklin . .c:we ams ws 21.2 52.3 654 A, ‘Decatur. .., «vc vunwmsms 38.9 24.2 454 LA Tensas 212 35.3 654 JA Ringgold... coz 5000s 38.9 343 {| TT 20M WEEE awe ’ ’ 671 WO TOBE oor vee se + 57.1 58.7 457 MS Warren oa 0 3 5 5 5 TE OF 20.2 185 671 MO Wright «ooo. 57.1 55.3 437 LA Macisaht «ix STs 00 ve 292 29.1 457 MS Clalbome .......ccu. 20.2 18.2 679 MO Pulaski............. 27.9 279 457 MS Sharkey ........c0:+ 20.2 32.4 684 MO Ste Genevieve . ....... 47 “7 457 MS lssaquena........... 20.2 30.7 NGS. «ov eee 2 467 MS JONBS. «uw wwsns nus 22.4 18.7 i» by Has Co : : . 2 467 MS Wayne ............. 22.4 14.1 408 MS Madison . . .......... 36 05 467 MS Jasper fs swiss pmrwy § 22.4 23.9 408 MS Copiah......o...... 3.6 4.5 467 MS Sith... sv: ussmiins 22.4 51.8 408 MS YaZ00. ..:.wesnssnsws 3.6 115 489 MS LOWNOBS. «« «x vou vo vom wim 252 22.8 408 MS SCO c.55. 515 nor + ssn # 3.6 13.1 489 MS Clay... resco snsmg om 25.2 25.4 408 MS Simpson... ssww os wan 3.6 7.3 489 MS Noxubee. ........... 25.2 34.6 408 MS Leake. us ae mns asm 3.6 13.6 511 MS Holmes. . o.oo... 51.0 53.9 415 MS Forrest. .....cusus vs 13.4 3.5 511 MS Attala.............. 51.0 46.1 918 MS Lamak ..comeuesnens 154 36 517 MS Washington... ....... 10.7 7.8 415 MS Marion ............. 13.4 13.4 > I 817 MS BOlVEF ....oms vnswmns 10.7 12.1 415 MS Covinglon. ...«cvueims 13.4 28.3 517 MS Sunflower 10.7 15.2 415 MS Jefferson Davis. . ...... 13.4 338 ; 0M AEB amare i E 415 MS Ponty... onnivnnnn 13.4 9.2 522 MS Leflore. .ccinuvumnins 28.8 7.2 415 MS Greene. .....c..cvcun 13.4 46.3 522 MS Tallahatchie. ......... 28.8 51.0 417 MS Lauderdale .......... 16.0 5.3 Se i Rum Duays SHELL EEEE Ps 1 417 MS Neshoba.., .wswwmiwssa 16.0 DOH. I TES HS Imi nimi om ata nee tists : 417 MS Newton. ............ 16.0 17.5 525 MS Montgomery ......... 35.7 35.7 87 Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of— 800- by residents of — unlinked unlinked nm Ae) area no. Area County area no. State County Area County 835 MS WIKINSON . ....vviui 38.4 38.4 153 GA: Babin «univ 5s a 0s iP 15.3 44.2 153 NC Swain............ 18.3 5.1 694 ~~ MT Custer ............. 15.5 3.8 tr 694 MT Dawson . .. ... 155 16.6 153 NC. 'Grabam ; . «suis ew mii 15.3 30.0 694 MT Fallon. ............. 15.5 17.1 154 NC Watauga. «ovis ws 5 ms 6.1 2.3 694 1585 48.1 154 NC ASNB : sun nsnmenss 6.1 6.5 694 = MT Prairie ............. 15.5 2.1 154 NC Avery.............. 6.1 8.1 684 = MT Gardield. ..5 45m wun» 155 23.2 154 TN JONSON iv casi mmenas 6.1 12.6 6% MT WiBR.....esns sues 155 5.5 168 NC DUBE. oo. 450% wm 500mm 7.2 6.4 699 6.4 1.2 168 NC Vance. . cus co sms 7.2 3.0 699 = MT Rosebud. ........... 6.4 34.5 168 NC Granville. ........... 7.2 6.4 699 © MT BigHom.:.:veswooms 6.4 16.6 168 NC “POISON + ss voivms sais 72 16.6 699 MT Carbon............. 6.4 3.7 168 NC Warren. ............ 72 12.2 20 Mp mas eeeesada oa 2 171 NC Guilford ............ 8.5 4.1 599 6.4 111 7 NC Davidson. ........... 8.5 23.2 699 6.4 102 171 NC “Randolph « «.. vsnos s0s 8.5 7.6 EE EER : 471 NC Rockingham. ......... 8.5 6.2 I £m wa 22 a 174 NC Pitt ............... 8.8 2.4 704 65 35.6 174 NC Beaufort............ 8.8 6.4 a A : 174 NC Martin. ..ovwocmammes 8.8 3.1 705 Lewis and Clark 11.9 5.6 174 NC Hertford ............ 8.8 18.9 706 ~~ MT Jefferson. ........... 11.9 40.9 174 NC Bertls....:..c:os0mun 8.8 2.4 705 11.9 21.1 174 NC Greene. ............ 8.8 43.1 7066 MT Meagher. ........... 11.9 48.4 174 NC Washington. ......... 8.8 2.4 174 NC Chowan ............ 8.8 7.0 za Crore > oy 174 NC Gates. ............. 8.8 53.7 RE : 174 NC Hyde .:vesvivizmons 8.8 9.3 NEC ME CT8R, cme reavas wen 3s 7 174 NC Tyrrell 8.8 2.3 711 MT Chouteau ........... 35 19.8 IBY. 5p itm tg ie wen 179 NC Cherokee ..... v.45 30.4 28.1 712 MY Mis emsun snes 32 179 GA Union. ............. 30.4 35.2 78 MF Lake TT 25 NG 179 NC Clay............... 30.4 30.3 719 MT Mineral. . . 35 8.2 179 BA TOWNS ov uo vn vis cuss 30.4 30.2 719 ~~ MT Granite. ............ 3.5 23.5 183 NC Robeson. ........... 12.7 9.6 183 NC Scotland. ....:cneb us 12.7 10.9 738 2 MT Hill cov smsve swans ve 5.8 1.9 735 MT Blaine. . . ... ..... .. . 5.8 3.4 183 NC Bladen............. 127 29.5 785 2 MT PHINIPE. «cows msinma 5.8 14.0 187 NC Mecklenburg ......... 1.9 1.4 738 2 MT Ubeny ...::ss sesws » 5.8 28.3 187 NC UNION: : vss sa sms 25 1.9 1.7 740 MT Glacier . . . . 30.0 34.0 187 NC Anson ............. 1.9 10.7 7 9 MT Toole. we:uswwsmenns 30.0 13.4 192 NC Wake...v:veivivsin 10.4 4.2 192 NC Johnston. ........... 10.4 23.6 743 MT Gallatin. ............ 5.9 3.1 : 743 MT Pak... .. ... 5.9 25 192 NC Franklin ...ovinsimvss 10.4 521 743 ~~ MT Madison. ........... 5.9 28.7 206 NC Cumberland. . ........ 13.4 4.4 743 59 20.4 206 NC Hameft......ovvusws 13.4 49.7 760 MT Roosevelt ........... 22.1 27.8 206 NG HOB ssi sarvm ie vnivns 4 %e 760 MT Vallay..... ws ww aims.s 22.1 52 211 NC Nash... :mvnsrwsnn 18.0 15.8 760 @ MT McCone ............ 221 39.0 211 NC Edgecombe. ......... 18.0 20.7 21 NC Halifax ............. 18.0 10.0 769 MT Flathead. ........... 27 1.9 769 MT uncon... .. .. 57 5.1 211 NC Northampton . ........ 18.0 40.2 213 NC Alamance ........... 255 6.4 2 NEE oa 22 213 NC Orange. .:ivsspsnsas 25.5 49.8 Toone : ’ 213 NC 188. .0ninsse sme non 255 14.0 7819 = MT Fergus............. 20.9 10.2 213 NC Chatham. ..c.cnv5 suv 25.5 29.8 781 20.9 47.7 213 NC Caswell. ............ 25.5 47.6 i) aE 2 oo on 216 NC Catawba............ 11.1 8.3 ote : 216 NC: Bue. :...ncvvinzsn 11.1 4.7 817 22.6 22.6 216 NC Caldwell . ........... 11.3 5.1 216 NC McDowell. .......... 11.1 31.7 144 4.0 1.2 144 NC Brunswick . . ......... 4.0 6.7 216 NC Alexander ........... 11.1 28.9 144 NC Pender............. 4.0 11.7 218 NC Buncombe. ......s+us 4.6 3.8 218 NC Haywood. ........... 4.6 3.7 1p NO BORA... sunememnen 64 51 218 NC Madison. .....s vss 4.6 6.2 148 3 NC SUNY unc sm mn enn 6.4 29 148 NC Stokes ............. 6.4 97 218 NO Yanoey....vsmeswews 45 27 148 NC Yadkin ........... .. 6.4 5.9 218 NC Mitchell. ....c0 00s 4.6 15.8 MB = NC ‘David, .: un suis wm vw 6.4 255 221 NC Craven. ............ 8.9 6.6 153 NC Jackson ............ 15.3 1.4 201 Ng ape “aevnisB eens 2 oe 153 NC Macon ............. 15.3 7.7 AICO: ~ ¢ nsrntekede icns : : 88 Table lil. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area by residents of — 800- by residents of— 800- unlinked —_— eee unlinked ree area no. State County Area County area no. State County Area County 221 NC Jones. ............. 8.9 55.4 627 ND Pierce ............. 18.0 18.7 204 NC Moore «ooo, 11.9 95 627 ND TOWRBKr.... sums mawn 18.0 16.9 224 NC Richmond ........... 11.9 13.4 642 ND Sisman ...c corms ws 10.7 1 6.9 224 NC Montgomery ......... 11.9 14.6 642 ND. Wells oun smzms sms 10.7 25.1 225 NC Rutherford. .......... 23.3 21.4 642 ND: FOSIBr 5 «wis oo vin cium 10.7 5.3 225 NC Polk. .............. 23.3 34.7 643 ND Ramsey .....+s55 55s 18.7 11.5 231 NC Rowan ............. 18.3 11.8 on 2 Busca Ws woe wn wena n 57 pes 231 NC Cabarrus. :.......... 18.3 26.6 Ws ssanan em inm a b k 231 NG "Stanly. vw. cot swis smi mag 18.3 15.9 650 ND Bamnes............. 36.3 21.2 650 ND LAMOUIE + .s vusunas 36.3 56.8 250 NG Gost. coo rwrmmrnys 154 R04 650 ND Griggs . ............ 36.3 50.7 236 NC Cleveland ........... 18.1 73 236 NC Lincoln. ............ 18.1 29.1 662 NO Wah. ..x: enn vines 24.1 29.8 242 NC Henderson .......... 18.5 22.2 3 ND Pola EERE oe 22 242 NC Transylvania. . ........ 18.5 8.3 RYBUBY «vv xx ign 48 Sin w : ’ 673 ND Emmons. . ... «usw cms 30.8 36.1 4 TT PE a 10. 4, 24 Ne Wayne we a 673 ND Mcintosh. . .......... 30.8 24.1 245 NC Duplin «vcusiumsis san 10.6 28.9 686 ND Richland. ........... 25.2 19.8 686 MN Wl ow rvs wn m ao 25.2 24.0 251 NC ‘SAMPSON: « «i» iv 00 0 wives 24.6 24.6 686 ND Sargent ............ 25.0 53.7 256 NC Columbus........... 23.4 23.4 690 ND Stark o.oo 76 48 260 NC Iredell. ............. 15.3 15.3 690 ND Dunn. ............. 7.6 20.7 ; 690 ND Golden valley. . ....... 7.6 173 261 NC Wilson ............. 8.5 8.5 . ison 690 ND Billings. ............ 76 45 264 NG. WIKBS: = vw nam sown ns 22.4 22. . re 2 ? 540 NE Lincoln. ............ 12.1 5.0 266 NC Onslow. .....vvouswws 5.3 5.3 540 NE Keith .............. 12.1 2.9 269 NC Pasquotank .......... 26.2 8.5 540 NE Chase TT re “ 12.9 12.3 269 NC Dare .............. 26.2 } 35.0 540 NE Perkins. ............ 12.1 25 269 NC Curfituck. ........... 26.2 43.2 840 NE Pronfer..vswsnsenas 124 585 269 NC Perquimans. ......... 26.2 50.1 $40 NE Datitlsuiuvsusmonsin 124 $33 269 NG Comiden « «son nse 26.2 137 540 NE Hooker. ............ 12.9 12.5 . 540 NE Logan ............. 12.1 37.8 542 ND Burleigh ............ 7.7 1.3 540 NE Thomas ....uvmsss is 12.1 53.7 542 ND - Morton. . .asussvimss 7.7 3.1 540 NE Grant. ............. 12.1 62.1 542 ND Mercer............. 7.7 8.2 540 NE McPherson. ......... 12.1 5.0 542 SD Corson. ............ 77 31.5 540 NE Arthur. ............. 12.1 0.0 542 ND Grant.............. 7.7 35.6 542 ND SIOUX... .v'oneenn... 7.7 4.1 554 NE Hall............... 9.1 4.4 542 ND Kidder ............. 77 24.2 554 NE Hamilton... ...s. uses 9.1 13.5 542 ND Logan ............. 77 62.9 554 NE. MamioK.. . «ros 5000 sams 9.1 10.6 542 ND Sheridan. ........... 77 66.3 554 NE Howard............. 9.1 3.5 542 ND Oliver. ............. 7.7 24 554 NE Sherman............ 9.1 42.8 554 NE Greeley. ............ 9.1 416 546 ND Cass co oimv nmin 4.1 1.3 546 MN Clay. ..oooeeeenn.. 41 1.9 557 NE Douglas «ou ime wwen 2.6 0.6 546 ND Trail «ooo 4.1 20.2 557 NE Sarpy. .. EERE 2.6 0.5 546 MN Norman . ........... 4.1 223 557 IA Pottawattamie. . . ...... 26 4.9 546 ND Ransom ............ a1 16.6 557 NE Cass Hw wie Wie WB RE 2.6 25.2 546 ND Steele ............. 4.1 39.5 557 IR HArSON ,. 0 Ja5% 2 dena 2.6 15.2 557 NE Washington. ...cuvs evs 2.6 10.4 563 ND Wad .............. 8.0 1.2 557 A MIB. « «95 5s wep es 26 20.8 563 ND MCLBN wc: v5: 4: mans s 8.0 50.6 563 ND Bottineau . .......... 8.0 9.5 560 NE Adams............. 15.1 4.3 563 ND Mountrail. 80 85 560 NE Clay. .............. 15.1 435 563 ND McHenry. . o.oo 8.0 18.9 560 NE Nuckolls ............ 15.1 7.9 563 ND Remvile. 8.0 14 560 KS Jewell. . ............ 15.1 60.7 563 ND Burke. ............. 8.0 15.3 560 NE Webster ............ 18.1 7.1 575 ND Grand Forks. . ........ 3.7 1.6 562 NE Buffalo. ............ 5.4 4.7 575 MN POR.» oon mmm sie 3.7 55 562 NE POO. cnrvssmewnss 54 23 575 MN Pennington .......... 3.7 1.8 202 NE 90/0) ae tierce wins «a $4 81 575 MN Marshall, ..oouns es 3.7 4.0 gee NE HEE cue rmrmnrane £4 gs 575 MN Kitson . . ooo. 3.7 9.8 562 NE Frail cme vam we sme 5.4 13.8 575 MN Redlake ........... 3.7 1.8 564 NE ScottBluff........... 46 1.5 575 ND Nelson ....v:s5 9mm i” 3.7 32.1 564 WY GOSHEN «ov cv vi% 4a 4.6 3.0 599 ND Williams . ........... 8.4 6.0 564 NE Morllwswnsmnnmomes 4.8 2s 599 MT Richland. ........... 8.4 6.0 564 WY Niobrara . .: «s5.05 20 3:9 4.6 24.3 599 ND Mckenzie ........... 8.4 18.6 S04 NE Sioukemeis seswuismen 4.6 300 599 ND Divide. ............. 8.4 14.4 564 NE Donnef..nsvmsmn snes 4.8 182 627 ND Rolette . ............ 18.0 18.1 505 NE: MBSOl cus 4 on oie vam 86 43 89 Table lll. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked Le —— unlinked ee =r area no. State County Area County area no. State County Area County 565 NE Wayne ..,ceovervvsns 6.6 11.8 55 NJ MOIS... vvomswmsmsi 22.5 23.4 565 NE Antelope. ........... 6.6 9.2 55 NJ Sussex............. 22.5 17.8 565 NE PiBree us wcwmvms ons ws 6.6 4.5 55 NJ. Warten. ...:csses 99 22.5 25.9 565 NE Stanton. .....oo050ms as 6.6 11.7 56 NJ Camden ............ 13.1 12.2 569 NE Custer ...:cvvvrmuns 20.6 21.4 56 NJ © Burlington. + = vx wv v5 0 13.1 13.5 569 NE Valley: . ocx snins datals 20.6 13.4 56 NJ Gloucester. . ......... 13.1 14.5 569 NE Garfield. ons wvws s0ws 20.6 11.2 64 EI. 116 91 569 NE Wheeler. .....ahs9u4 20.6 69.1 : 64 NJ Union.............. 11.6 17.6 569 NE LOUD .cvvvvmm ame nus 20.6 14.3 569 NE Blain 206 16.0 64 NJ Mercer. . ooo: muammsmn 11.6 4.5 BING. «wats Zw els HBi2 4 4 64 NJ Somerset ........... 11.6 13.2 573 NE HOM. oowaswmonsy vs 18.1 10.7 64 NJ Hunterdon. .....s:u+u 11.6 18.2 578 NE" SOWA ounanion eran is: is 84 NJ Monmouth. .......... 14.0 13.6 573 NE Boyd e.osvasrnsnrsms 15.1 38.8 a4 N) Ocean 14.0 14.5 573 NE BROCK .....sswaawnsns 15.1 1e (| 7 TTT rrr ! : 573 NE KeyaPaha, ......«i us 15.1 30.8 85 NJ Bergen............. 17.7 17.4 579 NE ‘Lancaster .. ... we: onan 6.1 1.5 LL NJ Possaie....usun unives 17.7 LL 579 NE Saunders ........... 6.1 53.7 107 NJ Cumberland. ......... 17.6 9.3 579 NE Seward. ..:.:s:0euss 6.1 6.4 107 NJ Salem ....o0cnsmunn 17.6 38.1 579 NE Saline. ............. 6.1 14.4 115 NS ESS. «ses eee sess 10.9 105 595 NE Yolk. ..:.50msumenms 20.8 11.5 115 NJ HUdBON , crm: vwiex 10.9 11.4 595 i Fare on 1 oot une 5 vt OD 2 ip 132 504 NM Lea. .............. 15.6 6.4 585 Mes ems geinmzme ans : 50. 504 TX Andrews. ........... 15.6 20.2 605 NE Platte. ............. 18.2 10.4 504 TX GaAINBS.. iv vmeivionmsa 15.6 39.3 605 NE Coa... nronsvsws 18.2 19.7 504 TX YOBKUM cvs ws morn wows 15.6 31.4 605 NE Buller. » +x ms mini 2 18.2 178 693 NM CUNY. oo eee. 5.3 5.5 605 NE Boone ............. 18.2 40.6 693 NM Roosevelt . ......vevs 5.3 4.8 605 NE Nang .......cot04 18.2 38.2 } 618 NE Red Willow . +... 28.5 13.2 695 NM Bernalillo LE 80 pum oe: on rong 3.5 1.2 695 NM Valencia . ..cocwsxw 3.5 7.5 618 NE FOMas.. .cinsemsiwyny 28.5 62.6 ; 695 NM Sandoval... ...cs4040 35 18.5 618 NE Hitchcock ........... 28.5 34.4 ’ NE H 28 0 695 NM Socorro ............ 35 1.3 513 BYES ws sesmn tmsinne 5 50.0 695 NM Torrance. ........... 35 2.0 630 NE Gage.............. 21.7 22.1 721 NM San Miguel . ......... 16.8 16.2 630 NE Joflerson. , +: wn wu ens 21.7 9.6 721 NM Mora . ooo 16.8 5.1 630 NE Thayer............. 217 36.3 721 NM Guadalupe. . . . ....... 16.8 25.3 640 NE - Richardson . .: «ove os 26.1 19.5 721 NM Harding ............ 16.8 43.9 640 NE Johnson ........s6 +s 26.1 37.4 733 NM Oter0.... «+. cmuw 56 30% 6.6 6.3 640 NE Pawnee ....... sux 26.1 31.5 733 NM Lncoln. oven. 6.6 7.8 849 NE Do0gS ocnsmsnnsmns 228 130 738 NM SantaFe............ 6.7 6.2 649 NE Cuming ..c covuvews 22.8 20.9 i N NE B 22.8 7 738 NM RioAriba ........... 6.7 8.3 629 hs on newsnnenans 23. 738 NM Taos .............. 6.7 75 661 NE Cheyenne. .......... 12.6 9.8 738 NM Los Alamos .......... 6.7 3.7 661 NE Kimball... vs 55:05 «wn 12.6 18.2 750 NM Sanduan ........... 5.3 6.4 665 NE Otoe .............. 37.9 43.9 750 CO lLaPlata ............ 53 1.6 665 NE Nemaha....:us ens sn 37.9 26.4 750 CO Montezuma .......... 53 23 750 CO Archuleta. ........... 5.3 3.7 3 NE “Dobie owe enmue ws He hia 750 CO Dolores. ............ 53 26.8 682 NE- Garden. «cis sms sna 35.1 35.1 750 CO SanJuan uo. eon «bein 5.3 125 687 NE. Dawson ... vss sinew y 21.7 20.1 768 NM Ohaves. ... crows sn 5.9 4.9 687 NE GoSpsr....:vsimsnns 21.7 42.9 768 NM DeBaca.....:v:53:5 5.9 32.8 35 NH Grafton. . «+ amvms wwe 11.4 20.3 775 NM McKinley. . . ...ccouuu 125 3.9 35 VT Washington. ......... 11.4 11.4 775 AZ ApBChB. ..+ ws sms 12.8 20.7 35 VT Windsor ............ 11.4 5.0 775 NM Catron =. asmsweswei 12.5 63.0 > 45 Sukie wwe rene En swe Ti4 7s 780 NM Colfax. . ............ 14.7 14.9 TONG8. six mutmes wii ool ’ i 780 CO LasAnimas.......... 14.7 14.4 74 NH Hillsborough ......... 9.6 10.6 " 74 NH Merrimack. .......... 9.6 55 307 BM UBIO. owes sims vue 1.4 51.1 74 NH BeIRNAD ..us 50000 1 wias 9.6 6.7 808 NM Grant. ............. 14.0 5.8 74 NH Caroll «.....« +s 505% 5% os 9.6 17.4 808 NM Hidalgo... : sivas samen 14.0 47.4 105 NH Cheshire. ........... 31.8 33.9 821 NM Quay.............. 17.2 17.2 105 YT -WIndham «x ou swam 31.8 28.4 822 NM Eddy ........ovn... 7.4 7.4 129 NH CO0B....covvemenss 22.1 15.4 697 NV EIKO. ..:vovnomsnmii 24.8 24.2 129 VT ESS. vu von ums na vm 22.1 57.6 697 NV Lander. ......o. .o.. 24.8 27.0 52 NJ Atlantic. ............ 13.9 16.5 697 NV Eureka. ..ocnennonsa 24.8 25.5 52 NJ Cape May... vu ou :0-+ 13.9 7.3 Table lil. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of— unlinked unlinked area no. State County Area County area no. State County Area County 717 NV. C1arK ov vn wav 4 won 2.9 23 86 RY DUCHESS. «<5 ow mnini » 22.2 23.9 717 NY NYE. & is mms me san mn @ 29 28.9 86 NY Ulster... ...convewmen 22.2 19.5 723 NV Washoe . .: vas us wins 5.6 2.4 90 NY Cattaraugus. ......... 30.0 31.6 723 NV LYON own nmin bmw 5.6 39.0 90 NY Allegany ....ccvssmsus 30.0 26.6 723 NY COUEohi eo waits vn 50 12 92 NY Tompkins ........... 9.9 6.3 723 NV Humboldt . ..... cc: 5.6 11.7 92 NY Cortland 99 16.0 723 NV Mineral. ............ 5.6 wg | FF NH AREER ameter : ’ 723 NV Porshing....c..:.0.x 5.6 0.0 103 NY Chautauqua. ......... 15.2 14.2 723 CA: SBI. v.c0 uo vows mh hide 3 5.6 48.2 103 PA Waren... cos smunw 15.2 18.4 723 NV Storey ............. 56 40.2 111 NY Washington. ......... 14.2 18.0 809 NV White Pine... ........ 15.4 15.4 mM NY Wameh. ...csomvuevs 14.2 7.3 12 NY Albany ............. 53 0.9 111 VT Bennington .......... 14.2 18.3 12 NY Saratoga. .. «vm. ws ssn 5.3 13.4 130 NY Rockland... .v ems 23.2 23.2 12 NY Rensselaer .......... 5.3 8.3 5 OH Jefferson. . .: : «corms 8.3 7.5 12 NY Sohensciady «vue» ss 53 03 5 OH Belmont ............ 8.3 37 15 NY Westchester. . ........ 18.9 19.4 5 WV OND «ou vovis anisms 8.3 1.8 15 NY Putham.........:.0: 18.9 14.6 5 WV Marshall. cvownvanaise 8.3 2.4 19 NY Monroe. . ........... 25 0.3 2 WY Ba00sk. se snsnenn ss 25 Bld ° 5 WV Brooke............. 8.3 4.9 19 NY Ontario. ............ 2.5 1.0 5 WV Wetzel . wu. vi iwruis un 8.3 12.0 19 NY Wayne .. wus sovws cman 25 39 os 5 OH Monroe. ............ 8.3 22.1 19 NY Livingston ........... 25 13.2 5 WV Tyler 83 9.5 19 NY Orleans. ............ 25 6.7 Yel wressnnior snr : : 19 NY Seneca. ............ 2.8 21.6 275 OH Montgomery ......... 5.1 1.4 19 NY Ya8B.....covnvrssi 2.5 10.9 275 OH Clark. .ooovnsvynnmsns 5.4 2.9 36 NY Onondaga. .......... 33 0.8 > oA Shs Pf ABE ew mene : ! 36 NY Oswego ............ 33 5.9 275 OH oe 8 wemmnn sree > 1 > 36 NY Cayuga. ............ 33 10.4 BMDAIGN + 4 aie 2a woah 5 36 NY Madison............ 3.3 9.2 277 OH. Franky ...« wae ows mem 2.0 0.7 a2 NY OtSego............. 205 14.7 77 GH DORWBIS, , ss ts swans 29 7.2 277 OH Pickaway. . «+s cve vn 2.0 12.3 42 NY Delaware. ........... 20.5 152 y pn NY Schohari 20.5 426 277 OH Madison ...vsousms es 2.0 6.1 CAONBHR vs: sir wisi vs : 277 OH Union. ............. 2.0 17.7 St NY O5LOWISNES es «vg 44 58 2 278 OH Athens ............. 16.9 15.9 51 NY Clinton. vas enews ess 6.8 3.3 s 278 OH Jackson ............ 16.9 29.2 51 NY Frankhn . .....ccuvrs 6.8 25 . 51 NY Essex 6.8 24.9 278 OH Galllf...c:esvms usu 16.9 7.1 Torry : 278 WV Mason ............. 16.9 11.6 54 NY Fulton ............. 19.7 11.6 278 OH Meigs... .«.u: 63 sms «vis 3 5 18.9 13.6 464 OK Pawnee ............ 20.3 52.5 358 OH Seneca............. 18.9 20.4 464 OK NODIB. vo: us vmansmy 20.3 27.7 352 OH Wyerndat. .« cus sn sms 13s 295 468 OK McCurtain. .......... 33.0 35.9 365 OH Butler. ............. 31.5 30.8 468 AR POIK. si: vows mmawn v9 33.0 29.3 365 IN Fayslle. ...:«: 405509 » 31.5 18.4 468 AR Sevier. ............. 33.0 27.3 2 x fakin HAE nen Se iy 471 OK Muskogee. . ......... 217 12.8 SPREE i ee a ’ ’ 471 OK Cherokee ........... 21.7 9.0 366 OH Muskingum .......... 13.1 4.7 471 OK Sequoyah .....: es 5 21.7 46.1 366 OH Guemnsey ........:.. 13.1 5.8 471 OK Addir. ..omh vhs reswn 2.7 24.0 366 OH. POHY x «mn nin vo mimes in wee 13.1 30.3 471 OK MeintoSh.....cuens va 21.7 37.6 3 x Moin iibhibludib tala i i 474 OK Texas.............. 30.5 33.6 BRE Ee ne ’ : 474 OK Cimarmon. .. . « coins es 30.5 16.1 375 OH Efg.....ucussminmas 14.8 13.3 474 KE MOON. .vz03 3s mnwn 30.5 27.0 7 2 Serdusky Sita oes Je 478 OK Comanche. .......... 14.1 3.7 Torry ’ : 478 OK Grady: cn» ss resmasn 1441 37.4 376 OH Wayne . .. vwivss swees 22.6 27.2 478 OK Caddo ..... ox cvnmes 14.1 22.0 376 OH Holmes............. 22.6 11.4 478 OK THEN, « vcvsvssnanis 14.1 45.5 378 OH Miami. . . oo ooo. 29.4 35.3 478 OR CORO 16:55 55 ne io swien 14.1 29.3 378 OH Darke. ............. 29.4 26.5 479 OK Pottawatomie . . ....... 22.8 25.3 378 OH Shelby. . .u: ws vasna + 29.4 21.9 479 OK POMOIOC. : cvs sm vins on 22.8 8.5 381 OH Licking. . «vor... 30.6 31.2 i Dif Lined: cuss cutme ne 228 527 381 OH Knox .............. 30.6 28.5 479 OK Seminole. .: «zs «+vs 5 22.8 8.5 479 OK Hughes. ............ 22.8 22.8 394 OF: ROSS! 215 86 4 susie ® mom wie 19.7 19.7 479 OR Boal cms swe ni vie vas 22.8 21.5 397 OH Coshocton. . ......... 27.0 27.0 481 OK Washington. ......... 33.9 10.4 481 OK 08808 s cvsuvmpcsrns 33.9 63.7 OH Ashtabula ........... . 235 899 Bria 23.5 481 OK Nowa. .. vss 2020 33.9 37.6 OH Ashland . .c cas we sms 20.5 20.5 a0 Shigh 483 OK. JACKSON ... «wiv ain ss ais 14.4 8.8 405 OH Logan «. .u:vs ames 23.8 23.8 483 OR KiOWa: « omic nmems 52 14.4 31.0 406 OH SGIOtO ............. 17.7 5.1 483 OK Greer. ............. 14.4 10.7 406 OH PK. . ooo, 17.7 53.9 483 OK Hammon ...... es: uwes 14.4 32.8 406 KY Lewis. .....cucumres 17.7 29.7 499 OK Oftawa............. 36.8 35.6 407 OK Woodward. . ......... 16.7 9.1 499 OK Craig.............. 36.8 40.0 407 OK ENS. covivinsmasonn 16.7 13.9 513 OR Stephens ........... 30.0 25.9 407 OK Haper. ...onzmavmn. 16.7 8.4 513 OK Jefferson. ........... 30.0 59.0 IPSCOMY. vu wuss 16.7 407 ™ Upscomd 0 £82 524 OK Gavin ............. 53.3 53.3 410 OK Tulsa.............. 3.0 0.9 oo 410 OK Creek. ............. 3.0 9.0 532 OK Kingfisher ........... 46.1 46.1 410 OK Rogers............. 3.0 2.2 534 OK. ¥a¥ : unwnamsnssyses 11.4 11.4 410 OK - Wagoner. . . «.. zs» saws 3.0 15.5 410 OK Mayes ............. 3.0 12.9 696 OR Multnomah .......... 1.9 0.5 92 Table Il. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked ee e————————— unlinked me Ce area no. State County Area County area no. State County Area County 696 OR Washington. ......... 1.9 0.4 37 PA AdaMS. . «cxcos nas ms ns 16.6 11.1 8% OR Dlackamag. «suze 22s 12 28 40 PA Northumberland . . . . . . . 16.8 215 696 WA Clark . ............. 1.9 1.0 : 40 PA: Snyder. ...wsws op vee 16.8 15.5 696 OR Yamhill. ... vow we can 1.9 6.4 ’ : 40 PA Union.............. 16.8 2.6 696 OR Columbia ........... 1.9 22.0 40 PA Montour 16.8 20.3 696 OR Clatsop. :: : vv vw vvvns 1.9 58 | > Mh Ameena : 696 WA Skamania ........... 1.9 44.5 41 PA Cambria ............ 4.5 1.9 710 rE 15.5 14.9 41 PA Bigir, cov ovnn sums 4.5 2.6 710 OR Polk 15.5 18.6 41 PA Somerset ........... 45 6.1 Ofcis sis s suis wy we 3 41 PA Bedford ............ 45 16.2 7 5 pe J ss tneen.n os 22 43 PA Allegheny ........... 26 0.4 OD WRI. + visas 28s s 2 : 43 PA Westmoreland . . . . .... 26 1.8 718 OR Morrow. . ........... 6.5 9.6 1 718 WA Columbi 65 10.7 43 PA Washington. ......... 2.6 4.3 OITINE: = gic = # ri siwiuvry : 43 PA Fayette. ............ 2.6 9.2 722 OR Deschutes. . ....::+. 2.7 1.6 43 PA Buller... .. ssw vn umes 2.6 11.5 722 OR CrO0K. 50.5% snvsiin wo wi iow 27 3.6 43 PA AMMSIONG : + «a4 + 4 sia 4 2.6 8.8 722 OR Jefferson. ....... ir B27 24 59 PA Lycoming . .......... 6.3 5.8 722 OR Harney............. 27 10.1 59 PA Clinton 6.3 84 722 OR Wheeler ............ 2.7 286 | TT TETrrrorrrrrrrrs ’ : ; 62 PA Franklin ............ 8.2 6.2 751 OR Union.............. 6.6 3.1 62 MD Washington . . . ....... 8.2 99 751 OR Baker.:: :eievmevens 6.6 ¥1.3 62 PA Fulton 8.2 99 751 OR Wallowa . = + «ssn + 2% as 6.6 92 | Tm rrr ’ i 755 OR Wasco ............. 10.4 9.6 72 PA" LoNioh woes susmznses 29 21 72 PA Northampton . ........ 6.9 7.1 755 OR HoodRiver .......... 10.4 71.0 Co 72 PA Monoe. ....:.v:04n 6.9 14.7 755 WA Kickitat. . oc .o0 vu vmsw 10.4 8.5 70 PA Carbon 6.9 235 755 OR Sherman. ........... 10.4 89 | © 0 TUEErrrrrrrrrras : 755 OR Gilliam . ............ 10.4 31.8 75 PA LuZOME . oc vmpswsmy 9.5 6.2 763 OR UnN. «oo, 105 134 75 PA Columbia .... 6mm ews 8.5 26.7 763 OR Benton............. 10.5 6.5 77 PA Venango. : : uv.» sviswss 16.9 10.3 766 OR COOS . ouvir... Wy 7.8 2 oh Sa Ete ite ale oe pod 766 CA DelNorte ........... 11.7 27 | rT EEE : ’ 766 OR CUMY : vuive vos mw cwss 11.7 11.2 106 PA Erie............... 3.7 1.3 770 OR Jackson ............ 37 0.9 hi PA Mary nptsiis son ris 3 Zz 2 2 770 OR Josephine. . ...ux« +» «5 3.7 17 § TT TR AER ahem sme 770 CA Siskiyou . ........... 3.7 16.1 108 PA Beaver, ... ws. sun ven 275 34.5 774 OR Kamath ... . ........ 97 79 108 PA Lawrence ........... 275 14.1 774 CA Modoc ............. 9.7 15.7 109 PA BolkS...osnsowsmanis 18.1 16.5 774 OR Lake ....siussmiws 4 9.7 16.9 109 PA Schuylkill. . .......... 18.1 22.0 795 OR Linton. s .6:vccsmamw 29.3 31.9 110 PA -MIfllin. . .. soonssmsin 20.9 53 795 OR THEMOOK . . x: v5 505. 29.3 24.8 110 PA Huntingdon .......... 20.9 30.8 709 OR Lane ...... .... .... 24 13 110 PA Juniata. ...iuswsomses 20.9 357 799 OR Douglas . .. oct oncns 2.4 82 119 PA Bradford. ........... 16.7 14.5 828 OR Grant. ............. 16.6 16.6 119 PA SUlIVaR. . ci: vo ims amen 16.7 48.7 9 PA Dauphin............ 3.4 78 120 PA TIOGE us soiciwn meme sw 28.2 28.2 9 PA Cumberland. . ........ 3.4 41 122 PA Indiana. ....: venues 37.9 37.9 9 PA Lebanon. ........... 3.4 6.4 A Lancaster ..... cc. 40s 9 PA Perry... .. ......... 3.4 30 131 P Lancaster 6.6 6.6 24 Rl Providence .......... 5.4 3.5 13 PA McKean... ews shes 18.4 18.4 24 CT New London — vow. 5.4 11.2 13 PA EK .owvosns dons mms 18.4 18.2 13 PA Pott 18.4 235 24 RI Kent c..u:wesvswngs 54 1.2 > oh = BF: wn x Wu BE BAS : 24 RI Washington . . ........ 5.4 3.2 BATION.» 34 4 lems cae 184 8 24 RI Bristol ............. 5.4 12.6 18 PA ConitB vuwnems nnisims + £3 1090 159 SC Orangeburg. ......... 165 125 18 PA Clearfield. . . ......... 8.3 7.8 18 PA Jeff 8.3 6 159 SC Bamberg. ... +4 wae swe 16.5 11.4 BISON. wre oc pte 0 % 159 SC Calhoun ............ 16.5 257 26 PA Philadelphia. . ........ 29 0.6 159 SC Allendale, ..... v5. 16.5 44.7 2% PA Montgomery “®aenwan 2 os 167 SC Richland. ........... 19 0.9 DNBWBID. so of 3 it om ; : 167 SC Lexinglon ...«..v«ws ss 1.9 1.0 26 PA BUCKS. . i «wi vw vie nin 29 9.7 26 PA Chest 29 10.2 167 SC Kershaw ....: us 5a 0 1.9 6.1 BIOL. « wmitmy sm rE 2 167 SC Newberry ........... 1.9 8.4 29 PA Lackawanna. ......... 10.8 4.1 167 SC Fafield «vows vrusmains 1.9 5.6 25 PA. Woe. wns eusumus 104 5% 185 SC Sumter. ............ 10.9 8.1 29 PA Wyoming. ........... 10.8 22.0 29 PA Pik 108 66.6 185 SC Clarendon. . .. «vs 59 +5 10.9 9.9 KQuuix #3 5 was win as : : 185 BC LBB. ois cot vs memmes 10.9 29.1 ig PA: YOK. on coosvmsme nme 16.8 17.9 188 8C Florence...» sus: 9.8 3.0 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked A ——————E tn FE unlinked eer eerwsemreeeon feeekensrectt area no. State County Area County area no. State County Area County 188 SC Darlington. . ..vo ww wus 9.8 2.6 581 8D FAK. «ous vmws om win 10.7 13.7 i 52 seri Sati 3 os 562 SD Shannon. ........... 13.4 8.8 TREE i na ’ s 582 NE DOWBS «suv tmimm anew 13.4 10.5 191 SC Greenwood . ......... 18.3 1.6 582 NE Sheridan, . ..q..u va a0 13.4 12.8 191 SC Laurens ... .wiwa cus wie 18.3 21.7 582 SD Jackson ............ 13.4 46.7 191 SC Abbeville. ........... 18.3 13.9 582 SD Bennett. .......s:0: 4 13.4 12.5 181 SC EJOSRSI) we vawwumsns tes 473 594 SD Codington. .......... 187 4.3 191 SC Sauda ....:isvsn ew 18.3 41.0 191 SC McCormick 18.3 12.5 soa SD DBYusynvsemenansin id 532 Troy : : 594 SD Hamlin. ............ 18.7 18.5 200 SC Greenville . .......... 4.7 1.4 594 BO Detel: sou anivwvuiz dt 18.7 36.2 200 SC Piokens. ms wwis sav ons 4.7 8.0 594 SD Clark... usin swan msm 18.7 20.5 200 8C Oconee ............ 47 20.8 607 SD Lawrence ........... 22.2 V7.7 201 SC Charleston. . ......... 1.9 0.9 607 BD BUMS. ws wmiwrns soem 22.2 17.8 201 SC Berkeley............ 1.9 1.1 607 SD Harding «.«uxcimsmiss 22 58.0 201 8C Dorchester. . «vs «susan 1.9 4.8 607 MT Calf. ..... vc sss ms wwe 22.2 47.7 201 BC CollSlB vs wnvsv aman 13 75 608 SD Roberts. . ........... 22.4 20.7 203 SC Maron... :vovesvovns 15.8 16.6 608 SO Ghl. ... wins snus 22.4 23.9 203 SC DMON. + vnnasdidens 15.8 15.0 608 MN BigStone . .........» 22.4 6.8 207 SC Spartanburg. . ........ 9.7 10.0 608 MN. Traverse. . ..... «sv vs vu 22.4 20.7 207 SC Cherokee ........... 9.7 12.6 609 SD Too . wv sv vi alse vias 30.9 271 207 BC UNION: sc: sms mn mms ms 9.7 2.6 609 NE Cherry ............. 30.9 35.8 210 SC ANCErSON ...... ou... 21.4 20.1 609 SD Melee. . . +. valu vv mas 30.9 40.1 210 GA Stephens ........... 21.4 2.2 626 SD Beadle............. 8.5 5.1 210 GA Hart. .....smsns snes 21.4 14.3 626 8D Hand. u.osvaasls sunny 8.5 14.6 210 GA Elbert. ............. 21.4 29.6 626 SD Jeraud............. 85 22.2 210 GA Franklin . =u vw sv winm 21.4 30.0 629 SD Davisoh ..... dues 20.2 7.9 217 SC Beaufort............ 12.2 6.4 629 SD Hutchinson . ......... 20.2 28.5 217 SC ‘Hampton. . «us sus vo suse 12.2 25.2 629 SD Douglas. ,...dus 204s 20.2 19.5 217 SC Jasper... :s:sweus 12.2 28.5 629 SD MINBr:.: cswremsmss 20.2 58.2 233 SC AKeN. ............. 69.9 74.7 623 $0 AUD .ucusveswnnn 292 CLA 033 sc Barwell . . . 69.9 44.3 629 SD Hansow......cess ane 20.2 19.4 629 SD Sanborn. ........... 20.2 44.2 > Se Pa E282 4 wien we wine 3 208 215 638 SD Hughes. ............ 18.4 6.0 neaster . cvs amv we 20.8 28.2 239 SC Chester. . . . 20.8 71 638 SD Walworth. . .......... 18.4 22.1 638 SD DeWBY ... vim simu pw 18.4 22.3 240 SC Georgetown. ......... 28.5 17.3 638 SD Polar. :veswswa sisi 18.4 16.9 240 SC Williamsburg . ........ 28.5 41.8 638 SD: SIanBY.. .. sims wem 18.4 5.4 638 BD ZiBbach .. vw. vs vw usw 18.4 29.1 250 SC 'HOMYusneswsmnumsns 92 ed 638 SD Campbell . .......... 18.4 55.5 538 SD Pennington . ......... 5.1 2.1 638 BOD: BUY vous smams nrays 18.4 6.8 538 8D Meade .u.vvvnsnvrus 5.1 13.9 638 8D. HVE 5s nw ae sik tind 18.4 49.3 538 SD FalRiver............ 5.1 12.4 638 SD Jones. ............. 18.4 24.1 2 SD Heme LL... : ar | #8 moe... 156 16 646 SD Gregory ............ 15.6 20.6 = 2 Yinnghane #Fewe pumas op ou 652 SD Brookings . .......... 22.9 17.8 541 SD Turner ............. 4.1 19.7 652 SD Lake eR EEE 229 28.2 541 SD Moody... 41 223 652 SD Kingsbury ........... 228 32.1 541 SD McCook. ....vvewiess 4.1 253 676 SD BrUR .... wows so snes 41.7 48.1 ; 676 LLL 41.7 46.1 556 BD Poking. .:.ssws vevas 10.2 10.3 556 ND Bowman... . ........ 10.2 8.0 676 SD Buffalo............. 41.7 26.4 556 ND Hettinger. « uv wu ome 10.2 19.8 146 TN Shelby: ocwsnmems 2.3 0.4 556 ND Adams ............. 10.2 11 146 MS DeSoto ............ 2.3 0.7 556 ND Slope. ............. 10.2 15.6 146 TN Toon wx eovosemima sy 23 2.7 571 SD Yankton . . .......... 17.2 20 146 MS Marshall . ........... 2.3 31.5 571 SD Clay. .............. 17.2 18.8 145 VS PAIR «as seume we nne pa 239 571 NE Cedar... .ovvoo vu. 17.0 39.2 146 IN Paysite... vive vnuns 23 4.2 571 NE KNOX + oo ooo 17.2 32.1 146 MS TAS... conenvcormnre 23 3.2 571 SD CharlesMix.......... 17.2 10.4 149 TN Putnam, ...vws name 9.9 6.0 571 SD BonHomme ......... 17.2 7.6 149 TN Cumberland. ......... 9.9 6.5 581 SD Brown ............. 10.7 3.0 149 TN WHE, cu vuvmusnirmss 99 128 581 SD Spink. ............. 10.7 29.8 149 TN Overton ............ 9.9 2.8 581 ND Dickey ............. 10.7 19.6 149 TN Fentress. ..:..::v:en. 9.9 4.3 581 SD Marshall . . oo. 107 31.0 149 TN Bledsoe. . oma uns 9.9 35.1 581 SD. ECunds . o.oo 10.7 4.8 149 TN JACKSON .. cvs vuvan 9.9 11.0 581 SD McPherson. ......... 10.7 48 149 TN Clafsveevnevinronns $9 258 94 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked ese unlinked A —————————— area no. State County Area County area no. State County Area County 149 TN VanBuren........... 9.9 33.2 209 TN MeMinh. ...cww susms 20.9 10.4 149 TN Pickett ............. 9.9 7.0 209 TN MONO, oo vuivins suis 20.9 39.1 151 IN Madison . . . . ... . .... 73 35 209 GA Famiin............. 20.9 28.1 3 209 TN. POI. «ows eam ais sm 20.9 12.7 151 TN Gibson. ............ 73 8.7 209 IN Mei 20.9 20.1 151 TN Carroll ............. 73 11.2 AGB. 0 wre wri wo : : 151 TN MeNally.. iz 2m5 annie 7.3 11.4 237 TN “Dickson « vv vox wrens s 44.9 49.0 151 TN Henderson ........us 7.3 15 237 TN Hickman............ 44.9 58.5 151 TN Hardy cov vrnns ven 7.3 14.8 237 TN Bumphieys . ...... «sux 44.9 28.4 151 TN Crockett . ........... 7.3 12.3 237 TN HOUSION ..: cs 20:05 0 44.9 275 > Hl Brose AEYSLIRTITIE 7s is 243 TN Roane ............. 48.1 48.4 ROOM sie a a wiz wim oe : : 243 TN Rhea .............. 48.1 42.4 152 TN MaUY. «con cnvme sms 17.2 9.9 243 TN Morgan. ............ 48.1 56.9 152 TN OMS. cinnnssmane sn 72 189 246 TN Greene............. 19.7 11.0 152 TN Marshall . ........... 17.2 26.8 246 TN Cocke 19.7 35.6 152 TN. Lewis. co cuvaar evan 17.2 28 | = 000M MEY smears E : 152 TN. PBIY 210 00 viii on wm 172 47.1 255 TN Hardeman, « .u « « vss « 66.5 66.5 155 TN Hamilon. «cows saws 1.9 0.9 265 TN Haywood. ........... 33.3 33.3 io 2A Woke, win me wm 92s 3 2 270 TN Dyer .............. 16.8 11.9 CEE REE d . 5 So ; : 155 TN Marion . ............ 1.9 8.6 0 ™N Lasts 168 eas 155 GA Dade .............. 1.9 0.9 414 TX Tom®Green ...:...... 33 13 155 TN Sequatchie .......... 1.9 4.4 414 TX Reagan... ...wvswsis 3.3 20.5 157 TN Davidson. . . ....... .. 15 07 414 TX SUBON usr zsrav carson 33 6.8 414 TX Crockett. ........... 33 9.5 157 TN Sumner ............ 15 1.2 oo: 414 TX Coke uivsvssnsmusns 33 16.0 157 1 RR, 1.3 1.7 ; E 414 TX Schleicher. .......... 33 1.9 157 TN. WIBON xs mansvns ow 1.5 25 i 414 TX Hon. ..osvsassncnes 33 3.8 157 TN Roberson... «vv as suis 1.5 2.2 414 TX Sterling 3.3 5.4 157 TN. Cheatham; : saws sess 1.5 91 \ 0 TEEErrrrrrrrrry 157 TN Macon ............. 1.5 8.6 416 TX Taylor. .«:vs mus asmws 6.1 2.8 157 TN “SBN. «zon sms 0 smaiv 1.5 8.0 416 TX Eastland............ 6.1 21.4 157 TN Trousdale . .......... 1.5 0.9 416 IX Callahan. .... iu: 45 6.1 7.8 416 TX Stephens ........... 6.1 16.1 1s nN oe | SHER nse 2 J > 416 TX Shackelford. . ........ 6.1 11.0 158 IN Anderson . . 21 27 416 TX Throckmorton. . ....... 6.1 68.8 158 TN Sevier. ............. 2.1 21 421 TX Hams. .ovvninconany 1.8 0.7 158 TN Loudon. ............ 2.4 45 421 TX Galveston ........... 1.5 0.4 158 TN Scott . © cover vnc x wns 2.1 20.7 421 TX FotBend..:vwws owns 1.5 25 158 TN Union. ............. 24 25 421 TX DBrazofig zvoc nna 1.5 3.0 163 TN Washington... ....... 8.4 9.8 pir 2 oy Corry ; 3 ¥ 3 163 TN Cantor. ..o: cvsws vos 8.4 75 Goi EBL mn 163 TN Unicoi 8.4 24 421 TH AUSHA. . oo. soneina smi wis 1.5 25.1 TICE 2 8 wis 2 ve alr a 421 TX Chambers. .......... 15 14.9 173 TN Montgomery ......... 15.6 7.8 421 TX Sandacinto.......... 1.5 51.0 178 KY Christian. ........... 15.6 21.2 422 TX Wichita. . o.oo. 5.7 25 173 KY Caldwell ....co0:00:0 15.6 22.0 422 TX Montague . .......... B37 31.2 173 KY. TOO . 255 0% corm mmm 15.6 37.2 i 422 TX Clays vow ue awn 9s ses» 5.7 34 173 KY THO « vovwnvmnsr swe 15.6 12.5 173 IN St t 1 6.3 422 TX Archer ............. 85.7 10.5 BWA. . vee 58 26. 422 TX Baylor ............. 5.7 12.9 176 TN Sullivan... .. ..s sussex 15.4 14.8 425 TX Lubbock ............ 3.4 0.9 176 VA Washington GEE he 15.4 9.3 425 TX Hockley . o.oo... 34 19 176 TN Hawkins ............ 15.4 25.3 425 TX Lamb 3.4 14.1 176 YA BusSell. ...:wimi sme 15.4 19.7 425 TX Tery o.oo 34 13.7 176 VA Scoft.............. 154 14 425 TX Bailey. ............. 3.4 14.8 186 TN Hamblen. .... ss sense 32.9 18.8 425 TX Crosby. ............ 3.4 16.6 186 TN JeHerSon. : : veins ssw 329 48.1 425 TX Lymn ..ccovrvomsmevns 3.4 8.8 186 TN ‘Grainger... «. «ww vs sigs 329 53.3 425 TX Garza. auams iannnns 3.4 1.4 186 TN: Hancock nv: snsasssa 32.9 25.1 425 TX Cochrah.... uve 3.4 3.8 194 TN Rutherford. .......... 26.6 30.0 425 TX DIkens wus wamms mains 84 51.5 194 TN. Warnen. . cc cozovsn 26.6 18.3 426 TX Bexar.............. 1.6 0.7 194 TN DeKab ............ 26.6 28.8 426 TX AlasCoSA.; cvs ww smitna 1.6 1.1 194 IN Cannon ws sms soos 26.6 11.5 426 TX Medina. ............ 1.6 33 199 IN Coffee 15.9 14.2 426 TX WIISON . ou wniviws swims 1.6 16.6 426 TX Kendall............. 1.6 30.5 199 TN. Franfin .v.uvinswe sn 15.9 B.7 426 TX Kames... womomu sims 1.6 31.6 199 TN Bediord ...u0vwewnen 15.9 25.0 426 TX Bandera .: sxe» wana 1.6 31.7 199 TN Grundy... ews wimps 15.9 28.4 426 TX Mc Mullen 16 25.6 199 TR MOOIe vn sv nmin in bas 15.9 20 | T4000 TREE eee ’ i 209 TN Bradley. ............ 20.9 19.4 428 TX Midland . ss s055 0 vows 75 7.3 95 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked i ee unlinked i a area no. State County Area County area no. State County Area County 428 TX Howard. ............ 75 6.0 461 TE BIOWH- vives moms 14.4 9.4 428 TE. Mati. «ou v vn swarms 7.5 12.8 461 TX Colematy. . .ovu: sv vws 14.4 22.9 428 TX Glasscock. .......... 7.5 14.8 461 TX SanSaba ........... 14.4 34.7 428 TX Borden............. 75 60.0 461 TX Mills. .............. 14.4 19.0 430 TX Dallas. .:.ocusnnimsas 4.6 4.4 462 TX Wibarger . ..:.: 0:0 29.8 28.3 430 TX Collin, ............. 4.6 33 462 TX Hardeman. .......... 29.8 33.8 430 TX EMS... covupnunzans 4.6 9.3 462 IX Foard... :w: wy: ne os 29.8 30.4 4 T; Kaufman. ........... 4 7.2 i; ; 2 n A al. a 2 5.0 463 TX VICIOHB. civ von vs vena 4.8 1.4 Torre 463 TX Calhoun .... vs: e90% 4.8 7.2 432 TX Cameron. ........... 2.9 2.9 463 TX DWH. .: ion vosnnrss 4.8 3.7 432 TX. WIBEY « «4 vw own va 29 2.1 463 TX LBVBCA... vs swe vmswu 4.8 10.8 434 TX Travis. ............. 23 0.9 a x galkeon mwa ahs 2 152 434 TX WiIlHamson. « c+ s+ + vv 23 42 | YT TEER rns : i 434 TE Hays ...ccomsme aman 2.3 4.8 465 TX Brazos... inivssmana 10.8 3.7 434 TX Bastrop: .msws smon 23 6.2 465 TX. GUMBS . vv suis on sms wa 10.8 33.4 434 TX Caldwell . .......:::. 23 7.7 465 TX Robertson........... 10.8 16.1 434 TX Burnet ............. 2.3 8.5 465 TX Burleson. ........... 10.8 33.8 24 TX Land. vasnrwsonense 23 Ho 466 TX JONES. . vii, 41.2 50.7 435 TX Tamant....svonns mas 6.3 6.5 466 TX Haske. .:5 voi cman 41.2 23.7 435 TX Johnson. ... sssas% 6.3 4.0 466 TX Box .:.revnnes naan 41.2 32.1 435 TX Parker ............. 6.3 4.1 466 TX Stonewall ........... 41.2 34.0 435 TX HOO, os ois wmowm 3 wo ew 6.3 4.5 466 TX Rel. vuwovvinsopssy 41.2 62.5 435 TX Somervell ........... 6.3 23.5 470 TX Kerr... o.oo... 15.9 9.0 437 TX Hale. ....:00cm550 00 28.3 28.1 470 IX Gillespie. ....u 65:0 15.9 8.1 437 TX Swisher ............ 28.3 28.3 470 TX Blanch xz: ssismsmmams 15.9 59.2 437 Th FloYo ois ws swems wma 28.3 255 470 TX Kimble ............. 15.9 14.7 437 TX Briscoe. ............ 28.3 38.7 470 TX Mason ..:-voo00 mans 15.9 27.6 437 TX Motley . : voovoms mwas s 28.3 40.4 470 TX Real. .............. 15.9 53.9 439 TX Titus oo 12.5 35 470 TX Edwards. swswos mama 15.9 35.9 439 TX Morris. . ............ 125 26.2 472 TX. WaIKSK . . v.vnu vimana 22.6 15.1 439 TX “CaMP: vv vivo wen 35 50 12.5 11.8 472 TR THY. coxcvssrrosnas 22.6 36.2 439 TX Franklin ............ 12.5 20.8 472 TX Madison............ 22.6 40.5 440 TX NUBCBS, . vous smu us 3.4 1.6 473 TX Grayson ..secsvcvens 12.6 7.6 440 TX San Patricio. ......... 3.4 1.9 473 OK Bryan. ...coomrmos ns 12.6 10.2 440 TX" BE. .u.uscwsmsws un 3.4 12.8 473 TX Fapnin......vva9 ses 12.6 26.1 440 TX Aransas ............ 3.4 2.0 473 OK AMORA. +: snsasuniws 12.6 38.0 ie = YS oa fle sie of te 4 77 476 TX Childress. . .......... 22.4 5.2 rors ’ 476 T% Hall.a:ws cos mssnies 22.4 14.6 442 TX Nacogdoches. ........ 15.8 10.6 476 TX: DOBBY : «ozo sv wiv ons 22.4 56.1 442 TX Shelby. .ws:wzuosase 15.8 22.1 476 TX Collingsworth. ........ 22.4 34.7 442 TX Sabine............. 15.8 30.8 476 TX Colllg..svvvonasnana 22.4 10.3 442 TX San Augustine ........ 15.8 16.2 476 TX KAG. vow ow sme he pons 22.4 68.7 450 TX Jefferson. ........... 15.7 14.3 477 TX Mclennan ... ous 10.8 6.6 450 TX Orahgs. .u:svoswsnwns 15.7 16.2 477 TX Hl ov ovne vw 25 np me = 10.8 24.1 450 TX Hardin ............. 15.7 15.9 477 TX Limestone........... 10.8 19.6 450 TX.. JASPBY . sss vw smannsy 15.7 20.5 477 TX Falls.......... 00:00. 10.8 34.7 $0 2 Ts gama onions vhs 7 po 480 TX ECHOr +o eveeennn 6.4 5.0 Tors 480 TX Pecos. .......ocncne 6.4 9.9 453 TX Poor... so v-radnss 3.0 1.9 480 TX: Upon. sc vssimaawisns 6.4 14.8 453 TX Randall. ............ 3.0 1.4 480 TX. Crane... es wo sme ue 6.4 9.9 453 TX HUchingon . ...ws 555 3.0 25 480 IX Teme ..:vnsonnmean 6.4 49.5 453 TX Moore ............. 3.0 1.8 480 TX LOVIOG wo nn vis smn monn 6.4 50.0 ii TX od Bef cena oz 422 536 TX ValVerde ........... 48 43 SER EE BE Eh : . 536 TX Kinney............. 48 20.7 494 OK Latimer. ............ 16.7 213 y 537 TX Webb. ............. 2.8 1.5 495 TX Guadalupe. . . cvs ss vse 27.8 30.0 537 TX Zapata ............. 28 25.1 495 TX Comal ............. 27.8 26.3 495 TX Gonzales. ........... 27.8 24.6 709 UT CachB ,..ccovucoens 8.8 3.1 709 UT BOXES. . vo ovsms ss 8.8 19.9 498 TX Hut.............. 21.7 25.0 709 ID Franklin ............ 8.8 5.4 > RB Hopigas Cee 3 2 Jos 709 ID Bearlake........... 8.8 13.5 BREWS Sw em ai : : 709 ID Oneida............. 8 Fi 498 TX Dela .............. 21.7 33.7 Be Z » } 712 UY Utah sicsivmimssmins 4.0 3.8 500 TX Bowie atl 418%: woe non ne 8.6 4.8 712 UT Sanpete ............ 4.0 5.7 500 AR Miller. ............. 8.6 3.3 712 UT Millard «oo ooeeeen. 4.0 8.2 500 TX Cass 2 Bana an 8.6 24.1 712 UT Juah «ooo, 4.0 3.3 500 AR Little River. . ......... 8.6 12.7 a 71 ut Ke. . ooo 501 TX Ochilftes. .. ou s5 sais 29.5 19.9 7:5 ut Senin ¢ 2 : > 2 501 TX Hansford. ........... 29.5 48.6 715 UT Summit. ............ 3.6 18.6 502 TX Deaf Smith .......... 36.4 19.3 715 UT Wasatch. ........... 3.6 21.6 502 IX Pamer............. 36.4 64.6 729 UT Uintah ............. 11.3 5.4 502 TX Castle . cvsuasmvwinise 36.4 48.3 729 UT Duchesne . .. ........ 11.3 19.2 503 TX Erath.............. 27.3 26.8 729 UT Daggett ............ 11.3 46.3 508 TX Comanche. .......... 27.3 28.1 739 UT Washington . . ........ 36 22 506 TX Ward... cou ow mmeanns 33.9 34.1 739 BT HON. ones ows us wane a 3.6 4.0 506 TX WIDKIBE. + cms ms wmiz ais 33.9 33.6 739 UT Beaver. ..: x ssinuss 3.6 4.8 507 TX Mavercks . « . o.. ns 12.0 5.0 739 NV Linco. ...onwwsins me 3.6 18.2 507 TX Uvalde ............. 12.0 26.1 742 UT DaViseivcinmsmunmsns 11.8 18.5 507 TH ZBVEIA «cs 6 monn nmin 12.0 8.3 742 UT "WBBBE «.... com rimi sm sim» 11.8 2.0 507 TX Dimmit............. 12.0 11.3 742 WY Ui cus ene nnres vs 11.8 22.8 508 TX EIPasO. ............ 0.7 0.3 re on Morgen helt Lina Lx a 508 NM DonaAna........... 0.7 7: IN: (Oc a Ai ’ : 508 NM Luna .............. 0.7 5.2 805 UT Sevier. iu siivvunmess 27.9 23.6 508 NM Sierra. ............. 0.7 15.7 805 UT Wayne............. 27.9 52.0 508 TX Hudspeth ......wv.u:. 0.7 19.1 805 UT Pie sz smeme was wows 27.9 46.7 509 TX HIOAG0 vr ces pw mn v 3.1 3.3 825 UT Grand. ............. 11.5 11.5 509 TX Star... 3.1 13 829 UT Carbon. ............ 12.4 7.5 512 TX Washington . . ........ 25.3 12.9 829 UT Emery ............. 12.4 20.8 512 TX lee............... 253 48.5 1 VA Faifax ............. 18.3 13.4 514 TX Dontotl. v- swsn vrs mss 37.0 39.1 1 VA Adinglon., ......om044 18.3 32.1 514 TX WBS . ovis mmsraman 37.0 35.2 1 VA AlexandriaCity . . ...... 18.3 34.8 514 TX Cooke ............. 37.0 19.6 1 VA Loudoun: .ssswsmnss 18.3 7.1 516 TX Navamo.........ceco- 17.5 13.1 2 VA Henlto. ....vco ims cnn 47 0.4 516 TX Freestone ........... 17.5 29.9 2 VA Chesterfield Sm vw sma nw 47 0.5 519 TX Matagorda. . ......... 19.4 15.3 2 VA Dinwiddie... cuemess 4.7 04 519 TX Whart 19.4 24.2 2 VA Hanover... .cssm» as 4.7 0.8 BDA: wociwow = wim 3: Bi : 2 VA Prince George . . ...... 4.7 0.6 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked emesis unlinked ee area no. State County Area County area no. State County Area County 2 VA CalBling ... . oe sms ms 4.7 31.7 58 WY Morgan. . .v:..:uvswse 15.8 26.1 2 VA Brunswick. .........: 4.7 18.5 58 VA Clarke... ov wasmawsy 15.8 25.8 2 VA Westmoreland . ....... 4.7 47.7 58 WV Hardy. ............. 158 57.1 ¢ YA Greene . osu cures 42 14 68 VA Henry. ............. 16.2 14.1 2 VA NOUOWEY.. .. + «50% oomine 4.7 18.6 68 VA Patrick 16.2 27.7 2 VA Powhatan .....:.«uus 4.7 13 | 0 TER eee ee ’ : 2 VA Goochland. .......... 4.7 4.7 82 VA Rockingham. ......... 14.7 3.9 2 VA Lunenburg. ...« cus wus 4.7 43.5 82 VA Shenandoah ......... 14.7 27.9 2 VA NewKent........... 4.7 23.3 82 YA Page ....ccormsnmay 14.7 31.4 2 VA Sussex.........s.ux- 4.7 22.8 82 WV Pendleton........... 14.7 34.8 2 VA King William. . ........ 4.7 17.2 104 VA WISE «ooo 18.5 13.9 2 VA ESBX....:simepmsws 4.7 10.9 i 104 KY letcher,............ 18.5 9.1 2 VA Amelia; ...ocmooms us 4.7 5.9 5 VA Cumberland 47 51.0 104 VA Leg... .usoivsswaans 18.5 30.7 2 VA Richmond . . 0 a7 35.6 104 VA Dickenson. .......... 18.5 33.6 2 VA Charles City. ......... 47 15.3 113 VA Smyth ............. 25.6 18.6 2 VA King and Queen . ...... 4.7 30.8 113 VA Wythe ............. 25.6 33.6 8 VA Newport News . ....... 7.4 3.9 117 VA AUGUSTA ....5 sss sn 13.9 15.8 8 VA Hampton City. ........ 7.4 45 117 VA Rockbridge .......... 13.9 6.0 8 VA, Yor s smsos sos mb vm 7.4 2.8 117 VA Highland. ........... 13.9 21.2 8 VA JamesCity .......... 74 4.2 123 VA Lancaster ........... 29.8 25.1 8 VA Gloucester. .......... 74 49 123 VA Northumberland . . . . . . . 29.8 36.3 8 VA Isle of Wight. . . ....... 7.4 43.7 8 VA Mathews. . .......... 7.4 6.2 126 VA Prince Edward . ....... 38.7 38.7 8 VA Middlesex ........... 74 45.4 127 VA Pittsylvania .......... 17.0 17.0 8 VA SUIY «ovis vnmi sons 7.4 58.1 128 VA Spotsylvania ......... 34.2 25.6 10 VA Roanoke............ 4.1 3.5 128 VA Stafford. ............ 34.2 47.2 10 VA Campbell ........... 4.1 2.4 128 VA King George ......... 34.2 30.5 10 VA Bedford ............ 41 0.9 10 VA Franklin ............ 4.1 4.3 133 YA . AccomacK.. ....: sus 20.8 26.6 10 VA Amherst . ........... 4.1 43 133 VA Northampton ......... 20.8 7.7 10 VA Botetourk....... v5: 92 41 8.1 49 VT Chittenden. . . . ....... 1.9 1.4 10 VA Appomattox. ......... 4.1 3.2 49 VT Franklin ............ 1.9 0.9 10 VA Floyd. ............. 4.1 50.3 49 VT Lamoille ............ 19 77 10 VA Craig.............. 4.1 14.4 49 VT Grandlsle........... 1.9 0.4 14 VA \VirginiaBeach ........ 1.5 1.3 95 VT Rutland. ............ 16.0 9.6 14 VA Norfolk/Portsmouth . . .. . 15 13 95 VI AQdBON :onn: mina 16.0 28.3 14 VA Chesapeake City. . . . ... 1.5 1. . 14 VA Nansemond. ......... 15 33 116 VT Caledonia ........... 19.3 27.8 14 VA Southampton. ........ 1.5 5.9 116 VT Orleans. ............ 19.3 9.4 20 VA Carroll ............. 19.8 30.7 700 WA Benton............. 5.6 4.7 20 VA Grayson TE 19.8 10.8 700 WA Prankiin . .....v. 10:5 5.6 7.6 20 NC Alleghany ........... 19.8 16.6 706 WA Spokane. ........... 1.0 0.8 25 VA Montgomery ......... 14.9 17.5 706 WA Slavens .........e.s 10 0.8 25 VA Pulaski... .......... 14.9 10.8 706 WA Lincoin............. 1.0 74 25 VA Giles .............. 14.9 9.9 706 WA Pend Oreille. . ........ 1.0 5.1 706 WA Fatty «.omvicivmeme nin 1.0 3.6 32 VA Alleghany ........... 17.2 16.6 32 VA Bath. .............. 17.2 20.1 731 WA Grays Harbor. ........ 37.7 34.4 731 WA Pacific ............. 37.7 51.8 38 VA Halifax ............. 23.7 72 ) 38 VA Mecklenburg . ........ 23.7 30.8 4 WA King. .............. 4.4 5.2 38 VA Charlotte. . . ......... 23.7 52.2 741 WA Snohomish .......... 4.4 1.8 53 VA Prince William. . . . ..... 31.9 52.8 748 WA Pierce ............. 6.3 5.7 53 VA Albemarle . .......... 31.9 15 748 WA Kitsap ¥ Tei ve Mk LS 6.3 9.0 53 VA Fauquier. ........... 31.9 18.1 748 WA Thurston. ........... 6.3 6.0 53 VA Culpepper. .......... 31.9 56 748 WA Mason .....civs swan 6.3 3.7 53 VA Omange............. 31.9 13.5 754 WA Grant. ............. 21.6 21.7 53 VA LOUISE ,.vus nonin ams 31.9 35.6 754 WA Adams ............. 21.6 21.4 53 VA Buckingham. ......... 31.9 28.7 } 53 VA Nelson . ............ 31.9 323 757 WA WHIINAN ; cv suis ms nna 12.3 15.5 53 VA Fluvanna. ........... 31.9 29 757 ID Latah... «co cme wo vmaz 123 9.2 53 YA Madison ......csvus us 31.9 3.1 788 WA YaKIMa «ous viswi soo 55 5.4 53 VA -GIEBNG; 5 vw wins wa 31.9 3.1 788 WA Kittitas . ............ 8.5 5.6 53 VA Rappahannock. ....... 31.9 20.2 760 WA Clallam. . oo ooo 5.6 31 58 VA Fraderiok. . oo 5v 050 4 5 15.8 4.2 790 WA Jefferson. . .......... 5.6 14.8 58 WV Berkeley............ 15.8 127 798 WA Skagit. ............. 8.9 6.7 58 YA - Wanen., «4 wu swaisws 15.8 17.5 58 WV Hampshire. . . ........ 158 32.3 798 WA Island. ............. 8.9 10.0 798 WA SanJuan ........... 8.9 27.3 98 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 — Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked unlinked area no. State County Area County area no. State County Area County 802 WA Chelal, si: suanis tna 7.7 5.4 373 WI Rusk ..cvrsnnsaniss 16.4 12.5 802 WA -Douglasi . «+ vv www wes 7.7 129 373 WI Sawyer. ............ 16.4 42.4 804 WA. LBWIS. 5s 600 6 nui ss + 15.5 15.5 380 WI ROCK wyews mwas mses 19.9 9.9 816 WA Cowlitz. . + ooo. 10.4 10.0 380 WI Walworth. . .......... 19.9 42.6 816 WA Wahkiakum . ......... 10.4 25.3 382 WI Racine. .::vsensmavn 11.9 13.0 823 WA Okanogan. .......... 16.0 16.0 382 W Kenosha............ 11.9 10.4 387 WI JUNBBM, : vss cosmsns 48.1 46.0 227 WA. WHRIEONY ums ws mais 43 43 387 WI Adams. ............ 48.1 52.0 279 Wi DWAR eases enna 22 2s 388 WI Sheboygan .......... 8.5 7.4 279 WI Oconlo. «cvussmaanin 3.2 9.8 388 Wl Manit 85 10.0 279 WI DOOF «overeat 3.2 1.2 DEOWOC le recess dali 3 : 279 WI Kewaunee. .......... 3.2 4.2 389 WI Oneida. ..:viuscmeun 75 5.6 201 WI Miwaukee. . ......... 0.7 05 2% WAMhis Stk siass cise 7 a2 291 WI Waukesha. .......... 0.7 0.9 OFOSE «tee + way eo / : 291 WI Washington. . ........ 0.7 1.7 6 WV Raleigh, «as wv as swems 18.8 11.6 291 WI Ozaukee. ........... 0.7 21 6 WV Fayette. ............ 18.8 16.6 300 Wl LaCrosse. .......... 5.3 0.4 6 WV WYOMING. + «vs ss mms ims 18.8 39.7 300 WI Monroe. ............ 53 2.4 17 WV Mercer. ............ 7.0 3.9 300 WE VeImon:..c.essviwsss 53 27 17 VA Tazewell. .... vo: cv: 7.0 6.8 300 WI Trempealeau ......... 5.3 30.6 17 WV McDowell........... 7.0 5.3 300 MN HoUSION .. +.ow +» 0 vin ww 5.3 37 7 VA Buchanan........... 7.0 10.3 300 WI JBOKSON uaom vwisies aa 5.3 8.0 17 YA Bland: : cvs snes 7.0 41.8 301 W Dang... iossvvunvan 3.8 1.3 21 WV. Randolph ....ou gwen s 14.8 3.3 301 Wi. Columbia ...::v50:55 3.8 7.7 21 WV Upshur: ws sss sean» 14.8 20.2 301 WI Sauk .............. 3.8 4.8 21 WV Barbour ............ 14.8 15.1 301 WI Iowa ..xvvmwsvismies 3.8 13.1 21 WV Pocahomas . . sxx ein 14.8 41.5 301 WI Marquette . . ......... 3.8 375 21 WV Tucker............. 14.8 10.8 316 WI EauClaire... .v.uv4 4.9 2.8 34 WY Cabell ..ovanvrnsonnsy 9.5 2.1 316 WI Chippewa . . ......... 4.9 4.4 34 OH lawrence ........... 9.5 27 316 WIE Dun. ...o.ssswe sens 4.9 6.9 34 KY BOW... vrwenmrnns 9.5 1.4 316 WIE PBPIY. wr x os 0st ms on Bs 4.9 21.4 34 WY Wayne... svnimens 9.5 1.2 327 WI Winnebago .......... 6.6 23 3a KY SIOSOUD..vus vwnwia ss 25 17. : 34 KY Caller. wins wmaws us 9.5 30.5 327 WI Outagamie. .......... 6.6 7.0 34 WV Lincoln. ............ 9.5 30.9 327 WI Waupaca. ...: «os wuss 6.6 11.0 357 Wi Calumet 6.6 155 34 KY Lawrence ....«s 9.45 9.5 11.4 RSTO, She Rute : 34 KY Martin. ............. 9.5 39.3 342 WI FondDulac......... 15.3 13.0 ag WY LODGED «vue rnnvnnns 27.9 203 342 WI Portage. :. us: cnvws es 15.3 9.7 39 WWM 27.9 337 342 WI Waushara . .......... 15.3 30.9 NGO. weve : 342 WI Greenlake.......... 15.3 31.7 61 WV Monongalia . ......... 8.5 1.3 344 WI Wood. «eee 9.0 3.7 61 WV Marion ............. 85 1.9 61 PA. Greene. «s+»: ws vis 8.5 17.5 344 WI Cla c:mn: vs amass 9.2 18.5 61 WV Preston. ............ 8.5 14.9 344 WI Taylor. ............. 9.2 12.8 61 WV Tavl 85 26.4 344 WI PCE .. oi 9.2 13.3 BYIQL. ow ria raw sp uras o 346 WI Dodge ............. 44.7 45.0 65 WV Kanawha. ........... 9.5 1.6 346 WI Jefferson. . .......... 44.7 44.4 os WY PUI. usumamvins 35 74 ; ’ 65 WV Boone ............. 9.5 29.5 349 WI Marathon ........... 20.7 21.2 65 WV Jackson ............ 9.5 35.9 349 WI Shawano. .:::«vs5 v5» 20.7 34.0 65 WV Roane ............. 9.5 5.8 349 Wi LINCO. «ori in ve 20.7 114 65 WY Clay. J cova cons ames 9.5 23.3 Se 2 J2npjats i Se van a = 66 WV Nicholas . ........... 30.1 36.8 Cory ’ : 66 WY Webster ,. .:vsvv000 30.1 13.8 959 Wi Bok ow ynnnms aang oe 224 67 WY Wood. ............. 57 28 359 WI Washburn. .......... 19.6 20.4 " 359 WI Burnett oo 19.6 107 67 OH Washington. ......... 5.7 3.7 daha ’ 67 WY BHCHIB ; voiiss mms maim» 8.7 19.0 362 W Grant. ....ocomemenas 36.2 37.0 67 WY CalBUN +. vv ven ve va 57 25.9 362 Wi Richland... .«w:wssus 36.2 26.0 67 WV Pleasants ........... 57 10.9 362 WI Crawford. . .......... 36.2 44.5 67 WV Wit. .............. 5.7 10.0 363 WI Marinette. . .......... 14.5 20.8 73 WV Hamson ....:vs cates 17.6 13.0 363 MI Dickinson ........... 14.5 3.4 73 WV Lewis. ............. 17.6 12.8 363 MI Menominee . ......... 14.5 20.0 73 WY Braxton. .:...ossvsms 17.6 35.7 363 Ml -Jon...ovensamiwims 14.5 6.0 73 WY GIMBr ..:oocaiinsms 17.6 45.0 363 WE Florent. vuswi i moms 14.5 8.7 73 WV Doddridge. .......... 17.6 10.6 367 WI Green. ....csnssuins 32.2 27.3 93 WV Greenbrier. . ......... 32.9 23.7 367 Wi Lafayels.. «cma sn inn 32.2 40.1 93 WV Summers ........... 32.9 27.4 373 WI Barron... .......... 16.4 9.0 93 WV MOOR. «oss vin svn 32.9 62.9 99 Table Ill. List of obstetric service areas for 800-unlinked solution and percent of hospital births by residents outside areas in 1984-86 —Con. Percent of hospital births outside 800-unlinked area Percent of hospital births outside 800-unlinked area 800- by residents of — 800- by residents of — unlinked ree —————e unlinked Ei area no. State County Area County area no. State County Area County 8% WY JBTBISO0M. x « 4x vot woe 875 97.5 782 WY Washakie ........... 10.3 8.6 745 WY Natrona ...:corvrnes 4.1 27 782 WY Hot Springs... vss en 10.3 134 743 WY Converse .vsmevwsws. 41 2 786 WY Laramie ............ 10.3 9.1 752 WY ‘Campbell ...cc.onvns 7.4 4.1 786 WY Palle. ..uvis sven ra 10.3 21.8 752 WY Sheridan. cess uvmsnes 74 25 792 WY Lincoln. ............ 17.4 24.9 752 WY JOhNSON.:.:.:ss:penus 7.4 7.7 752 WY Crook 74 471 792 WY Talon. :uceesmssman 17.4 7.2 TORI: Al rerlelraz ve ade : 792 WY Sublette . ........... 17.4 12.8 753 WY Albany ...::cro:0v sss 9.6 5.3 753 WY Carbon... 96 13.1 806 WY Fremont. ..ccvosvusnam 4.0 4.0 753 CO JACKBON i soy «iva ous 9.6 53.1 820 WY Weston. ............ 97.9 17.1 765 WY ParK.: co covnimbmmemn 72 55 831 WY Sweetwater . ......... 11.6 11.6 765 WY BigHom. ....v:v3 v4 7.2 10.8 100 Table IV. Percent distribution of obstetric service areas and population according to travel for hospital births, by type of area: United States, 1984-86 [Data are for the coterminous United States] Type of area Obstetric areas Travel measure All Metro Nonmetro HCCA County Numberofareas. ................. 836 343 493 775 3073 Births outside area by residents Percent distribution of areas Lessthan25 percent. . . ........... 78.5 90.1 70.4 71.0 27.0 25-40 PICO. iv iv 3 wiv 2 5 W® KH 20.1 9.0 27.8 21.8 23.3 S50 percentormore. .............. 1.4 0.1 1.8 7.2 49.7 Percent distribution of population living in areas Lessthan 28 percent. ... .....ccuv es 94.4 96.9 80.2 96.2 65.3 25-49percent. . ......... 0000s 5.3 3.0 18.9 3.3 18.1 50 percent or more 0.3 0.2 0.9 0.6 16.6 Births inside area by nonresidents Percent distribution of areas Lessthan25percent. . ............ 95.5 98.8 93.1 81.5 60.7 25-40 POTCBINL + 5: + (0 + oo ti 9 ioe w ser i 3 4.4 1.2 6.7 16.6 31.9 BOpercentOr more. : «. vu + vs sms vi 0.1 — 0.2 1.8 7.5 Percent distribution of population living in areas Less than 25 percent. . . . .......... 99.0 99.5 96.2 97.3 68.1 25-49 percent. . c.o cvs vr r nas BE 1.0 0.5 3.8 25 27.41 S50 percentormore. .............. 0.0 — 0.1 0.2 4.9 Note: Metropolitan service areas include at least one metropolitan county. 101 Table V. Percent of hospital births outside area of residence, by type of county and type of area: United States, 1984-86 [Data are for the coterminuous United States] Type of area * Type of county of residence Obstetric area HCCA County Percent of hospital births ANCOUNIBS . , .vxx vos tm sms err ms smnmn sms 8.4 6.6 23.3 MBIOPOMAN. . 0 ov oi 8 557% 608 304 0570 a0 rw im 58 3.3 18.2 LBPGBICOIB. = voix wiv wx % rv wmv 00% ww ew 3.5 1.7 9.6 LBLABTRARGE: . vv iv SHE wis Bie #18 mew mes oom 08 10.0 4.7 37.5 MBIT. «oo vv wiv ws www yp wag ars a 5.8 3.8 17.2 BAN. swiss 5 EERE SEE PE Se bed BE 5.4 4.5 13.1 INOAINBITOPOMAIY « iv 1 v0 Cons wie 4 viel 3 wha 2 WHS 171 17.8 40.9 UBER ; vv vmrma mmr mrt hs EER Fran cam» 11.6 10.1 20.2 LESS UPBAIY «x cmv wu mis ins nme mie sms ie 95 EAS 18.9 20.3 47.1 BULB: + iv von rims smeis $mamia gsm vomuonn » 26.1 30.3 77.3 Yr U.S. GOVERNMENT PRINTING OFFICE: 1991— 3 12- 082/ 40012 102 Series 2 No. 113 Vital and Health Statistics From the CENTERS FOR DISEASE CONTROL / National Center for Health Statistics Sample Design: Third National Health and Nufrifion Examination Survey Nislolcinnl ISI MRAM rd . U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES 3 r Public Health Service 7%, he Centers for Disease Control CENTERS FOR DISEASE CONTROL ¥ Co. Vital and Health Statistics Sample Design: Third National Health and Nutrition Examination Survey eRAITY AF PALIT BESKEL BS UNIVERSITY Ui Limi, OCR £7 Series 2: Data Evaluation and Methods Research No. 113 This report presents a detailed description of the sample design for the Third National Health and Nutrition Examination Survey, 1988-94, including a brief description of research that led to the choice of the final design. The National Health and Nutrition Examination Survey (NHANES) is one of the major surveys of the National Center for Health Statistics, Centers for Disease Control. Information on the health and nutritional status of the noninstitutionalized population of the United States is collected through the NHANES household interviews and standardized physical examinations. EE ET, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control National Center for Health Statistics Hyattsville, Maryland September 1992 DHHS Publication No. (PHS) 92-1387 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Office of Research and Methodology Monroe G. Sirken, Ph.D., Associate Director Kenneth W. Harris, Special Assistant for Program Coordination and Statistical Standards Lester R. Curtin, Ph.D., Chief, Statistical Methods Staff James T. Massey, Ph.D., Chief, Survey Design Staff Andrew A. White, Ph.D., Chief, Statistical Technology Staff Division of Health Examination Statistics Robert S. Murphy, Director Kurt Maurer, Ph.D., Deputy Director Ronette R. Briefel, Dr.P.H., Coordinator for Nutrition Monitoring and Related Research Clifford Johnson, Chief, Nutrition Statistics Branch Katherine M. Flegal, Ph.D., Chief, Medical Statistics Branch Christopher Sempos, Ph.D., Chief, Longitudinal Studies Branch Vicki L. Burt, Chief, Survey Planning and Development Branch Jean Findlay, Acting Chief, Survey Operations Branch Robert Krasowski, Chief, Computer Systems and Programming Branch Contents INEEOAMCIION. +. m2 vane #56 902 0 855 15 2 500 iv wah BE 90 50 30 Bh © 00a 20 00 05 Sh 3 tn 0 0 0 0 0 A 0 os 00 0 00 0A 0 w 0 RL Design SPECIfICALIONS uous wmin’e vm amass oe weirs 3500408 5 5 5 5000 av 3 HITE 04 305 8808 EE 50 a0 58 5 0 EAE a 4 eas SUTVEY OVICCHIVEE «105055 55 5 eames abot me TIE 0 00 ELE 5 od so 3s A BA low i TUL Sot aed 8 HR A Domain and preciSion SPEOIICAIONS' . ooo wow ms sian iis 15 ala ism oman is 55 arbin ws was so l4 S84 pwn cs we tn RE Bin wim» Operafional rEQUITCINBINS «cv siva i mrs simmer unis Soran as EEE Se mas BRR EE a vem IE BIREAR ar mses Methods 188aAICH .. ov viv vr tiie tnnnnmnsme bss mses rnsnseseesseeanes Brey Bs Tt ge oe os SRE RR 7 3 BR de ed i ee SAMPLE AERITN. ov cin rd 55 BER wrens AG: BERS: wontons FE FIED tamersion wisest 00 BHF: AL isis mo ssw RIE Br 00: mi wasn 0 10 HH me SUMMATY + + vv wnr 501s wm se © owen wr SE 5g 2 0a smu ew SEE © 4 50 we 0 ws EER § See's Weme Oils Susu a Ea ae Stratification and. SeleCHON OF PSUIS: . + vos mim sm miami 50 4 58 m0 0 501 Bm is oh av £5 0 5% go A 4 oi 6 800 00 ie 0 Bg 8 SCICCUION OF SETINCIIS wo mio 515 1 ore oie Pris 51 5 6 ome vo #08 HBS 61805 vim so 8 08 Fae an 0 i 47 po 30 Ent oe Tw oy 50 8 Selection of NOUSEROINS AA PRISOME wu swim smi mame 055% 5 x wwom 0 vom 4050 3505400000 0 0001 BE 8 Mr 10 8 0% og 05 em 0 BC 4 in Selection Of SAMPIE PEISONE cx smn 5m so «20a © £0 5.8 90850 5 3% 2 518 800 HIE wed 0/0 3 #56 S081 ERLE 0000 3 I Rk ad IRE tr Spat a ESUIMZAHON PLOCEAUNES 24 55's 0.0 wv svv sw vii 51015 + wi sms im 800 w/in ors wate BG 50005 300 ites hie gt 810 10.0 0 pions 0 81D Gc im Weighting the Sample dala . .cvnrws sas soe vie imams sss sms as ose sas wees cei FREss en formas wpm ww ids wi National INTIATON WEIGHS Los vais memes £00858 uns ms mo 000 500 810 or 8 0 io 513 5 3590 5 ot 9 om Te 0 3 0603 0 oe 0 313 3 8 0 om VIEIANCC CSIIMYTITON 525 55k vv osm 4908 105 50 ir 0 vem i 5 80 0 i 00 0 0 0B 0 bs io 000 0 41 do RETCTEIOTE v2: imineees on rs isp ga 5: 28 ls vt emer fr 17S TREES mtd 0 00 1 mts st cis 75 A gt cs oe wr 7 5000 Sep os co 50 B SARF Hoare Listol detailed TABIES: ...ocum so numn » omois oes ol £05 5000508 ils BRE HE GERE BD Be 20 605 900 Ar 0 5 Su od 5000 51 0 A I mic eh BLE APPONTIX oo: crm a mrm srw nm Sma iu® fon wn #0 S00 2 0 F002 20 000020 0 00 0 vt 9 00 3 2 6 a 3 2 List of text tables A. Analytical subdomains classified by race-ethnicity and age: Third National Health and Nutrition Examination Survey. VOBBHOE on cen cog 5 508 15 mm 008 mr 5 0 5 050 A 0 0 2 13 te or 0 00 0 5 im Expected total and examined person sample sizes by race-ethnicity: Third National Health and Nutrition Examination Survey, JO8B=0 . ...vvuiomimnos su rsn ss Hum sme sw sms A005 BEELER 2 F054 BHT 11585 as SE SDF DEES oh SE pA Health and Nutrition Examination Surveys by selected sample design parameters .............................. Screening sampling rates and number of screened households, by density stratum: Third National Health and Nutrition Examination Survey. 1988-09: us unw:grens prams menus mein omen fas sew oars © BE a sm 8 6295 5088 5 w fin 9 » ow 58 Percent of Mexican-Americans in segment by density stratum number: Third National Health and Nutrition Examination Survey, TOBB-D4 oo cmitme mam mons sis Snes ms 5005 0 85008 0m im nw 3 50 30840 90 m0 5 9 v0 12 ad 0 Percent and cumulative percent of households in stop-rule subsamples: Third National Health and Nutrition Examination Survey, 1988-94 . .....auvuvsssmennmas sess enamine ss vs sm 5556 ns 6s en same 6 abe wness List of appendix tables L II. III. Target diseases and conditions: Third National Health and Nutrition Examination Survey, 1988-94 ............... Home examination components for selected age groups: Third National Health and Nutrition Examination Survey, TORB-O i smmmsin vs nes vows IES BE 5% EE BOER ES 0 HARD BETES FS EEE BEE 0 3 RET 0 2 ERE Gr oulEle Examination components for each age group: Third National Health and Nutrition Examination Survey, 1988-94 . . .. ION ~~ Oh = wo oN 0.0 Symbols Data not available Category not applicable Quantity zero Quantity more than zero but less than 0.05 Quantity more than zero but less than 500 where numbers are rounded to thousands Figure does not meet standard of reliability or precision Figure suppressed to comply with confidentiality requirements Sample Design: Third National Health and Nutrition Examination Survey by Trena M. Ezzati, James T. Massey, Office of Research and Methodology, National Center for Health Statistics; Joseph Waksberg, Adam Chu, Westat, Inc.; and Kurt R. Maurer, Division of Health Examination Statistics, National Center for Health Statistics Introduction The National Center for Health Statistics (NCHS) conducts several large-scale national health surveys. The National Health and Nutrition Examination Survey (NHANES), designed to assess the health and nutritional status of the noninstitutionalized population of the United States, consists of adult, youth, and family questionnaires followed by standardized physical ex- aminations in specially equipped mobile examination centers (MEC's). The Third National Health and Nutrition Examination Survey (NHANES III) is the seventh in a series of surveys using health examination procedures that have been conducted since 1960 by NCHS. The target populations, the sample designs, and the data collection procedures for the previous health examina- tion surveys—NHANES I, NHANES II, and a special survey of the Hispanic population (Hispanic HANES)—have been described in previous reports (1-8). The target population of NHANES III is the civilian noninstitutionalized popuiation 2 months of age and older. The survey is being conducted from 1988 through 1994 and includes a sample of approximately 40,000 persons. The household interview includes demographic, socioeconomic, dietary, and health history questions; the examination component consists of examinations by a physician, a dentist, and health techni- cians. In addition, a home examination consisting of an abbre- viated set of physical measurements is offered to persons who are unable or unwilling to travel to the MEC for a complete examination (9). The interviews and examinations are per- formed by permanent staff employed by Westat, Inc., the data collection contractor for NHANES III. As in previous NHANES, this survey has the following major goals: ® To produce national population health parameters; ® To estimate the national prevalence of selected diseases and disease risk factors; ® To investigate secular trends in selected diseases and risk factors; ® To contribute to the understanding of disease etiology; and ® To investigate the natural history of selected diseases. A list of the major target conditions of NHANES III is shown in table I and the examination components of the survey are shown in tables II and III. Previous NHANES have included only persons aged 6 months through 74 years. Because a growing proportion of the U.S. population consists of older Americans who experience greater morbidity and disability, NCHS imposed no upper age limit for the NHANES III. Also, because of the need for better information on the growth and development of young children, infants 2-5 months old are included in the NHANES III for the first time. Older persons and children are oversampled so that estimates of their health status can be made with acceptable precision. Although not excluded from the target sample, small num- bers of black and Mexican-American persons were included in previous NHANES. Therefore, reliable estimates of their health and nutritional status could not be produced for subgroups of these domains, that is, by age, sex, or other important demo- graphic or socioeconomic breakdowns. To resolve this prob- lem, NHANES III was designed to include a large sample of both the black and the Mexican-American populations so that reliable estimates of health and nutritional characteristics can be produced for these two largest minority groups of the U.S. population. This report describes the broad design requirements for NHANES III, the research undertaken to develop the sample design specifications, the major research, the estimation proce- dures, and the methods used for estimating sampling errors. Documentation of the survey content, procedures, and methods to assess nonsampling errors are reported in another publication (9). Much of this report is based on survey requirement docu- ments and the final sample design report prepared by Westat, Inc., as part of an NHANES III methods research contract (10). Design specifications Survey objectives A desirable first step in designing a survey is to define the analytical objectives. As in the previous NHANES, a primary purpose of NHANES III is to produce a broad range of descrip- tive health and nutrition statistics for sex, race, ethnic, and age subdomains of the population. These data can then be used to measure and monitor the health and nutritional status of the noninstitutionalized population. Because NHANES III was designed to produce cross-sectional data and because respon- dents will be followed over time for future interviews and/or examinations, a set of cross-sectional and longitudinal objec- tives was developed. These objectives are: ® To produce estimates of means and proportions with a reasonable level of precision for a broad range of health and nutritional variables by sex, race, ethnic, and age sub- groups of the civilian noninstitutionalized U.S. population; ® Todetermine differences between subgroup estimates with specified type I and type II errors; ® To monitor secular trends in health and nutritional status for subgroups of interest; and ® To investigate the etiology and natural history of selected diseases through followup of a cohort of initial respondents. Domain and precision specifications A primary interest of NHANES III is to estimate with acceptable precision the health and nutritional status of sub- groups of the population. The subdomains for which separate analyses are expected to be carried out in NHANES III are ‘shown in table A. The set of subdomains consists of sex-age groups for the largest race-ethnic subgroups in the U.S. popu- lation. The analytical domains in table A consist of the age groups shown separately for males and females. Therefore, 52 subdomains—twice the number of age groups are shown. The sample for NHANES III was designed with specified precision for each of the major subdomains in table A. The goal is to have approximately equal precision for each of the analyti- cal domains. The minimum precision requirements are: ® A prevalence statistic of 10 percent should have a relative standard error (RSE) less than 30 percent; and ® Differences of at least 10 percent in health or nutrition statistics between any two subdomains should be detected with a type I error of no more than 0.05 and a type II error of no more than 0.10. Table A. Analytical subdomains classified by race-ethnicity and age: Third National Health and Nutrition Examination Survey, 1988-94 Black White and all other Mexican-American’ 2-35 months 2-11 months 2-35 months 12-35 months 3-5 years 3-5 years 3-5 years 6-11 years 6-11 years 6-11 years 12-19 years 12-19 years 12-19 years 20-39 years 20-29 years 20-39 years 30-39 years 40-59 years 40-49 years 40-59 years 50-59 years 60 years and over 60-69 years 60 years and over 70-79 years 80 years and over "Mexican-Americans can be any race. NOTE: The analytical domains are for males and females separately. To meet the predesignated precision requirements, the sample size for each of the defined subdomains for all groups was determined to be 560 or greater. For the subdomains with considerable oversampling in Mexican-American density strata, the minimum of 560 was increased to compensate for the design effect introduced by the variability in sampling rates among density strata. In addition, the total examined sample size for both black and Mexican-American persons is required to be 9,000, with 12,000 for white and all other persons (table B). The examined sample sizes were inflated by about one-third to account for expected nonresponse to the examination portion of the survey to determine the total sample sizes for the survey. The total sample sizes by race and ethnicity are also shown in table B. Superimposed on the minimum cell size of 560 was the requirement that the total examined number of infants and young children (separately for males and females) in the 3 age groups—under 1 year, 1-2 years, and 3-5 years—needed to be at least 1,000 to provide adequate sample sizes for updating the growth charts for NHANES III. Table B. Expected total and examined person sample sizes by race- ethnicity: Third National Health and Nutrition Examination Survey, 1988-94 Race-ethnicity Total Examined persons Tolal .. i. comms 40,000 30,000 / Blagk +. ..5ue. or ceiiammmmens 12,000 9,000 White and all other ............ 16,000 12,000 Mexican-American’ ........... 12,000 9,000 "Mexican-Americans can be any race. Operational requirements NHANES is unique because the examination component of the survey involves a number of logistical and cost concerns that also have to be considered in the design of the survey. Therefore, the following operational requirements were im- posed upon the design of the survey: ® The number of sample persons selected at each stand (survey location) should be between 300 and 600, with an average of approximately 450, yielding an expected 340 examined persons. Research has shown this number to be the approximate optimum number to give as many primary sampling units (PSU’s) as possible, while keeping the sample size in each area large enough to justify the costs associated with moving and setting up the mobile examina- tion centers (MEC’s); ® The minimum time to complete field work at any stand is 4 weeks; and ® The data collection period should be 6 years. However, because it is not desirable to wait the full 6 years to have updated data since the last national survey (NHANES II, 1976-80), the first-3-year period (phase 1) should contain a national probability sample of the eligible population so that some estimates of health and nutrition can be produced when the first 3 years of data collection are completed. (The subdomain estimates produced from phase 1 will not be as detailed as those for the full 6-year survey because the analytical data requirements apply only to the full 6-year data collection period.) In addition to these operational requirements, another fac- tor considered in the final design was how to select as large an average number of sample persons per household as possible, thereby maximizing the response rates and reducing screening costs. This was also an important design consideration because previous experience with both NHANES and Hispanic HANES indicated that response rates increase when a large sample of persons are selected within households. One of the factors thought to be responsible for the increased response rates in multiple-sample person households is that each person is given a remuneration for his or her time and participation. Methods research In planning for NHANES III, new design features for the survey and a broad reexamination of the basic methodology for sampling and respondent contact were evaluated. First, the use of NHANES III as a baseline for a long-term longitudinal study of adults and children as well as a cross-sectional study created a broad range of analytical possibilities for NHANES and also created several design and cost issues. Second, because previ- ous NHANES focused on statistics for the total U.S. population, relatively small numbers of black and Mexican-American per- sons were included. Therefore, estimates of their health and nutritional status for separate age-sex subdomains were often inadequate due to their unreliability. NCHS recognized the need for information on subgroups of the population when the special Hispanic HANES was conducted in 1982-84 (8). Including adequate numbers of black and Mexican-American persons in NHANES III serves to update the Hispanic HANES and enhances comparison of the health of minorities with that of white and all other persons, thus allowing the investigation of factors related to differential health status. Therefore, the primary focus of the sample design research was on the development of efficient procedures to oversample selected subgroups of the U.S. population so that reliable health and nutritional statistics are available. To evaluate these design issues, an 18-month methodologi- cal contract was awarded to Westat, Inc., in May 1986. The development of an efficient sample design for NHANES III was one of the major tasks carried out as part of this contract. A more detailed discussion of the other methodological issues exam- ined as part of this research contract is described in another renort (11). The original design specifications required separate esti- mates for black, Mexican-American, and Puerto Rican popula- tions in addition to white and all other populations. However, research indicated that it would be costly to locate a probability sample of the Puerto Rican population of sufficient size to provide useful data. Because of cost constraints and the limit on the number of persons that could be examined in the 6-year time period, separate estimates for Puerto Rican persons were not included in the final design specifications. Puerto Rican per- sons, if selected for the sample, are included in the “white and all other” estimation cell. Similarly, the cost and feasibility implications of produc- ing separate estimates for the increasing numbers of Asian- Americans in the United States were also evaluated as part of the design research. As in the case of Puerto Ricans, the cost appeared to be very high and oversampling of Asian-Americans was not considered in the final design. 4 The three major options (including combinations of them) evaluated as part of the design research were: ®* An independently selected sample of the civilian noninstitutionalized population; ® A sample of interviewed and examined persons from prior NHANES and a supplementary sample of the current civilian noninstitutionalized population; and ® A sample of either persons or addresses selected for the National Health Interview Survey (NHIS), another major survey conducted by the NCHS, supplemented by an inde- pendent sample of the current civilian noninstitutionalized population. As described earlier, the sample design was required to produce statistically reliable estimates of means and propor- tions for a broad range of health and nutritional variables by age, sex, race, and ethnicity. To accomplish this objective, a number of related issues were addressed. The following key issues were examined: ® Thesubdomains that would need to be oversampled to meet the precision requirements and a determination of the extent of oversampling required; ® The data collection costs associated with each of the three design options; ® The treatment of prior nonrespondents and persons who died or moved if a linked NHANES or NHIS design were employed; ® The approximately optimum design of a supplementary sample, if such a sample were needed; ® The major features of an independent sample design including: — the number of PSU’s to be selected; — the approximate optimum segment size for the area sample; and — the within-PSU selection procedures that would efficiently produce the desired sample sizes. The major conclusions drawn from the research formed the basis of the final NHANES III sample design. They are: 1. The NHANES or NHIS samples for prior years are not efficient methods of establishing the sample for NHANES III. An independent single sampling frame, consisting of an area sample, should be used. Area samples should be used both in the selection of PSU’s and of households. An area sample supplement to the National Health Interview Sur- vey (NHIS) would be necessary because a single year of NHIS is not large enough to provide the needed sample for most of the subdomains, especially if only part of the sample located in about 89 stands is used. (Only about 89 stands can be included in NHANES III if the total target sample size is about 40,000 persons and if average sample size per PSU is to be 450 persons.) Similarly, an area supplement would be needed if an earlier NHANES was used as the sampling frame because of the inadequate sample size of black and Mexican-American persons and some of the age-sex subdomains for the “white and all other” category. Therefore, the cost and complexity of working with multiple frames appeared to negate the value of the alternative frames. . The Mexican-American sample should be viewed as a supplementary sample in the moderate to high dense Mexi- can areas, superimposed on a basic self-weighting sample that is large enough to satisfy the sample size requirements for all the black and white and other sex-age subdomains. . The supplementary sample should use geographical strati- fication with strata defined on the basis of the proportion of Mexican-American persons in the block group-enumera- tion districts in the 1980 census. Optimum allocation of the supplementary sample to the geographic strata should be used, taking costs and variances into account. . To establish the sampling and subsampling rates, sex-age- race-ethnicity subdomains that need approximately the same number of screened households to obtain the target sample should be grouped with a single sampling rate used for all subdomains within a group. This will significantly reduce the number of separate sampling rates that have to be applied. . The optimum allocation sampling rates for the various Mexican-American strata should be set to provide the number of households to be screened as required for the rarest sex-age subdomain for Mexican-Americans. The subsamples of the screened households needed for the less rare subdomains are to be obtained by progressively cutting back the rates in the most dense strata to equal the rate in the next remaining stratum. This will minimize the variation in weights for each subdomain, within the limits of what can be done without increasing the level of screening. . The procedure for subsampling persons and households from the screening sample should be carried out by subdi- viding the screening sample into random, although un- equal, subsets. In one random group, all persons in the screening sample are designated as sample persons. The sampling rate reflected by this random group is the rate applicable to the largest subdomain in each race/ethnicity group. In a second random group, all persons except those 10. in the largest subdomain are designated as sample persons. In the third random group, all except the two most common subdomains are designated, and so forth. The number of random subsets and their sizes are calculated to ensure getting the desired sample size. This procedure maximizes the average number of sample persons per household while adhering to the sampling rates required for each subdomain. The measures of size used in the selection of both PSU’s and segments will be XP, Myr, where Pp, is the proportion of the population of PSU j (or segment j), in the ith subdomain, M, is the total population of PSU j (or segment J) and ris the sampling rate in subdomain i. The summation isover the subdomains. For measures of size, the subdomains will be collapsed to the three race-ethnic groups. The reason for this is that the race-ethnicity composition of an area is fairly stable over time but the age-sex distributions can vary greatly, particularly because of the aging of the population since the most recent decennial census. . A two-PSU-per-stratum sample design should be used for noncertainty PSU’s in NHANES III. The two PSU’s should be selected with probabilities proportional to size (PPS) and without replacement. The measures of size should be the ones described above. After the selection of the noncertainty PSU’s, one PSU should be assigned to the first 3-year period (phase 1) and the other one to the second 3- year period (phase 2). No special action will be taken to supplement the sample of low-income persons since oversampling poverty areas on the basis of geography is not very efficient and any other method is expensive. However, it is likely that even with- out oversampling for persons below the poverty level, there will be enough black and Mexican-American persons and total persons in the sample below the poverty level to permit a reasonable amount of analysis for collapsed sex- age subdomains. The sample design should be reviewed at the end of the first year of field operations, and periodically thereafter, to ascertain whether modifications are necessary. It is par- ticularly important that this be done in 1991, when the 1990 census data become available. In addition to these major design decisions, the research also provided much of the data needed for the detailed specifi- cation of the sample design, for example, stratification vari- ables for the selection of PSU's, the optimum segment size, the method of calculating measures of size for both PSU’s and segments, the sampling rates to be used in the various density strata, and the level of screening required. Sample design Summary The general structure of the NHANES III sample design is the same as that of the previous National Health and Nutrition Examination Surveys (4-7). Each of these surveys used a stratified multistage probability design. A summary of the major design parameters for the two previous NHANES and the special Hispanic HANES, as well as NHANES III, is given in table C. The NHANES III sample has been designed to be self- weighting within a PSU for subdomains and almost self-weight- ing nationally for each of the subdomain groups (but not for the total population). The NHANES III sample represents the total civilian noninstitutionalized population 2 months of age or older in the 50 States of the United States. The first stage of the design consists of selecting a sample of 81 primary sampling units (PSU’s), which are mostly individual counties. In some cases, adjacent counties were combined to keep PSU’s above a mini- mum size. The PSU’s were stratified and selected with prob- ability proportional to size (PPS). Thirteen large counties (strata) were chosen with certainty (probability of one). For logistical and operational reasons, these 13 certainty PSU’s were divided into 21 stands (survey locations). After the 13 certainty strata were designated, the remaining PSU’s were grouped into 34 strata, and two noncertainty PSU’s were selected per stratum. The selection was done with PPS and without replacement. Each noncertainty PSU is also referred to as a “stand.” The NHANES III can thus be considered as consisting of 81 PSU’s or 89 stands. The 89 stands in the sample were randomly subdivided into two sets, one consisting of 44 stands and the other consisting of 45 stands. One set of stands was allocated to the first 3-year survey period (1988-91) and the other set allocated to the second 3-year period (1991-94). Therefore, unbiased estimates (from the point of view of sample selection) of health and nutritional characteristics can be independently produced for both phase 1 and phase 2. For most of the sample, the second stage of the design consists of area segments comprised of city or suburban blocks, combinations of blocks, or other area segments in places where block statistics were not produced in the 1980 census. In the first phase of NHANES III, the area segments are used only for a sample of persons who lived in housing units built before 1980. For units builtin 1980 and later, the second stage consists of sets Table C. Health and Nutrition Examination Surveys by selected sample design parameters Parameter NHANES | NHANES I Hispanic HANES NHANES III Age of civilian noninstitutionalized target population ............ 1-74 years 6 months—74 years 6 months—74 years 2 months and over Geographical areas .......... United States (excluding Alaska and Hawaii) Average number of sample persons Per household . ...vwewwnsive Number of survey locations .... 100 64 United States (including Alaska and Hawaii) Southwest for Mexican-Americans; NY, NJ, CT for Puerto Ricans; Dade County, FL for Cubans United States (including Alaska and Hawaii) 2-3 2-3 17 in Southwest; 89 9in NY, NJ, CT; 4 in Dade County Children 2 months— Domains for oversampling Sample size ........... Examined sample size ... Yearscovered ......... Low income; children ages 1-5 years; women ages 20-44 years; persons ages 65 years and over Low income; children ages 6 months-5 years; persons ages 60-74 years Dade County: persons ages 6 months—19 years and 45-74 years; Southwest and NY, NJ, and CT: persons ages 6 months—19 years 5 years; persons 60 years and over; Mexican-American and black persons and 45-74 years 28,000 16,000 40,000 (expected) 20,000 12,000 30,000 (expected) 1976-80 1982-84 1988-94 NOTE: NHANES is the National Health and Nutrition Examination Survey. of addresses selected from building permits issued in 1980 or later. These are referred to as new construction segments. In the second phase, 1990 census data and maps are being used to define the area segments. Because the second phase follows within a few years of the 1990 census, new construction will not account for a significant part of the sample and the entire sample comes from the area segments. The third stage of sample selection consists of households and certain types of group quarters, such as dormitories. All households and eligible group quarters in the sample segments are listed and a subsample is designated for screening to identify potential sample persons. The subsampling rates enable pro- duction of a national, approximately equal, probability sample of households in most of the United States, with higher rates for the geographic strata with high Mexican-American concentra- tions. Within each geographic stratum, there is an approximate equal probability sample of households across all 89 stands. The screening rate in each stratum is designed to produce the desired number of sample persons for the rarest age-sex domain in the race-ethnic group defining the geographic stratum. Persons within the sample of households or group quarters are the fourth stage of sample selection. All eligible members within a household are listed, and a subsample of individuals is selected based on sex, age, and race-ethnicity. The definitions of the sex, age, race-ethnic classes, subsampling rates, and designation of potential sample persons within screened house- holds are developed to provide approximately self-weighting samples for each subdomain within geographic strata and simultaneously to maximize the average number of sample persons per sample household. Experience in previous NHANES indicated that this increased the overall participation rate. Although the exact sample sizes will not be known until data collection is completed, estimates have been made. A summary of the expected sample sizes at each stage of the design is shown in the following table: Numberof PSU's ..............iiiiiian.. 81 Number of stands (survey locations) ............ 89 Number of SBgMEeNMS ....:...coansnumnpsss ces 2,138 Number of households to be screened .......... 106,000 Number of households with persons ............ 20,000 NUMBOT Of PEISONS: ; 1 « & » wanismimmmintiamames ss 52 40,600 Number of interviewed sample persons ......... 35,000 Number of examined sample persons ........... 30,100 Stratification and selection of PSU’s The sampling frame for NHANES III was composed of all of the counties, parishes, and independent cities in the United States (including Alaska and Hawaii), all of which are referred to as “counties” for convenience. From these counties, primary sampling units (PSU’s) were formed. Most PSU’s consist of single counties although a few were made up of small groups of contiguous counties. The PSU’s for NHANES III are defined as individual counties. This definition reduces the amount of travel necessary to visit the mobile examination center (MEC) for the examination component of the survey and achieves as high a response rate as possible. Combinations of counties were used only where counties were so small, in terms of population, that their probabilities of selection would have been lower than what is required for some of the domains. If selected for the sample, they would introduce considerable variability in the sampling weights. Consequently, these small counties were combined with one or more adjacent counties to form more efficient sampling units. For the same reason, the independent cities in Virginia were combined with one or more nearby counties to define the PSU’s for sampling. Of the approximately 3,100 counties and county equivalents in the United States, 2,812 PSU’s (most of which consisted of individual counties) were defined for NHANES III. After stratification, the sample of PSU’s was selected from the 2,812 PSU's. The sampling frame, measures of size, and stratification variables used 1980 census data. The frame of PSU’s was constructed by merging information from two data sources. The first of these was the Bureau of the Census STF-1C file, which contained information for each county or county-equivalent. The information used for the PSU sampling was: (a) region; (b) metropolitan status; (c) 1980 population; and (d) 1980 popula- tion for Mexican-American and black persons. The second source was the Bureau of the Census file “Population Estimates by County with Components of Change, 1981-1985.” Data from the latter sburce were used to update the 1980 population figures in the STF-1C file as discussed in the next section on calculation of PSU measures of size. This information was useful for stratification of PSU’s and for determining the probabilities of selecting the sample of PSU's. Calculation of measures of size Multistage area samples selected with probability propor- tional to size (PPS) generally use a single variable as the measure of size, for example, total population or total housing units. This was the practice followed in NHANES I and II. However, when subdomains are to be sampled at different rates, but a self-weighting sample is desired for each subdomain, such measures of size can result in highly variable workloads. This applies to the sample selection at both the PSU and the segment level. For example, assume that there are L subdomains that are to be sampled at rates r, r,, . . ., r,. In the jth PSU, let the proportions of the population in the various subdomains be denotedby P, .. ., P,,whereP, +P, +...+P, =1 foreach J. If the measures of size of the PSU were N,, where N, is the total population of the PSU, then the populations in the subdomains in the jth PSU are P, N,, ... P,, N,, respectively. With these measures, the probability of selection of a PSU will be kN,. For self-weighting samples, the within- PSU rates for subdomain / need to be proportional to r, /kN,. Thus, the number of sample persons in the ith subdomain is (P,N)(r,[kN,)=rP, [k. The total sample size in the PSU will be as follows: Srp, lk The total sample size in a PSU will thus depend on P, , the percentage distribution of the subdomains within the PSU. Unless the P, is approximately the same in all PSU’s, the workloads will vary. In the case of the particular subdomains established for NHANES III, the distributions by sex and age will probably not create any problems. Age and sex distribu- tions are reasonably consistent in large areas such as counties. However, race and ethnicity vary considerably. The percent- age, for example, of Mexican-Americans and black persons will range from close to zero in some counties to almost 100 percent in others. For logistical and operational reasons, workloads per noncertainty PSU must be kept within a fairly narrow range. Using total population or households to establish the PSU probabilities of selection would create large differences be- tween the desired uniform workloads and the actual sample sizes. Subsampling or sample supplements in each PSU would be required to bring the sample size in each PSU in conformance with the desired numbers. This would add substantially to the variances. Furthermore, it would introduce uncertainty into the total sample sizes by subdomain. Using the quantity M, = Y=, Ny, as the sampling measure of size for PSU / avoids these com- plications. Under PPS sampling, the probability of selecting a PSU is then kXP, N,r. Within a PSU, the sample size in the ith subdomain will be (P, N,r, )/(kXP, N,r.). The total sample size in the PSU over all subdomains is then iP, N,r, W(k2P,. N,r.) = 1/k and is thus independent of both the N, and the P,. The formula for the PSU measures of size given above was somewhat oversimplified. For example, the NHANES III de- sign included provisions for stratifying areas by concentration of Mexican-Americans, and sampling households within strata atvarying rates depending on the race-ethnic distribution within the stratum. Consequently, the sampling rates will actually depend on the density stratum, as well as on the sex-age and race-ethnicity subdomains. Specifically, in terms of the differ- ent subdomains and minority strata, the measure of size as- signed to PSU A was: M, = au Ni h = PSU i = Mexican-American density stratum (as de- fined in the next section) where k = race-ethnicity subdomain = sex-age subdomain N,, = estimated 1990 U.S. population in the (k,/)th subdomain in density stratum i in PSU A = U.S. sampling rate in density stratum i for the (k,D)th race-ethnicity-sex-age subdomain. To The exact values of the N, ,, were not known at the time the PSU’s were selected. Hence the required population counts were estimated from 1980 population figures and adjusted by the most recent Bureau of the Census projections of the county’s population. Thus, for NHANES III, the measure of size as- signed to PSU Ah was: Cc C M, = bE Nou dt 22 > A h Cc , 1980 population for PSU 4 C’,, = 1980 population count for race-ethnicity subdomain in PSU A C, = mostrecent (1985) population count for PSU h C'.. = 1980 population count for race-ethnicity subdomain k in minority stratum i in PSU A where C7, C ,, = projected 1990 total U.S. population count for race-ethnicity-sex-age subdomain (k,/) C , = projected 1990 total population count for race-ethnicity subdomain &. A further simplification was necessary because the values of C’,, were not available. Assuming that the distribution of a minority among geographical strata is identical among PSUs, then C’, = C’, P,, where P, is the U.S. proportion in the ith stratum for the kth race-ethnicity subdomain. Hence, the meas- ure of size for PSU h was calculated as follows: C M, = Po pYe he h where C o kl A, = Yr, i ei il C , The derivation of the r,, used to compute the measures of size is described in the section on selection of segments. Minimum measures of size The selection probability of a PSU determined the maxi- mum rate at which persons residing in that particular PSU could be selected for NHANES III. If the measure of size of a PSU was too small, the required sampling rates for some subdomains could not be achieved. Consequently, special weighting proce- dures would be required for these PSU’s, and the resulting variability in weights would increase sampling errors. To ensure that all required sampling rates could be achieved, the measure of size M, of a PSU had to satisfy the following inequality: M, 2 Mr, [2 where h denotes the PSU, i denotes the Mexican-American density stratum, k denotes race-ethnicity, / denotes the sex-age subdomain, M is the measure of size of the stratum in which the PSU is located, and r,, is the sampling rate for the (k,))th subdomain in density stratum i. The factor of 2 in the right-hand side of the above inequality reflects the fact that two PSU’s were selected per stratum. Counties for which M, < Mr, /2 were combined with a neighboring county to form PSU’s satisfying the minimum size requirement. The procedures used to construct the frame required estab- lishing the set of PSU’s prior to stratification, thus the value of M was not known at the time decisions on combining PSU’s were made. Therefore some approximations were nec- essary. Also, the values of r,, depend on the density strata, and many PSU’s do not have high-density strata. For these PSU's, it was unnecessary to have criteria that guarded against situations that will not occur. The data files used to select the PSU’s did not provide direct information on whether all den- sity strata occurred in a county. or whether some strata were not present and thus could be ignored in the inequality above. Because the strata sizes were to be made as close to equal as possible, a good approximation to M was obtained by assuming they were equal. Let m_represent the number of certainty stands in the certainty PSU’s. The number of non- certainty PSUs is then 89 -m . (Note: m_ is the number of stands (survey locations), in the certainty PSU’s, not the number of PSU’s. Some certainty PSU’s contain multiple stands.) tM, is the total measure of size of the noncertainty PSU’s, the value of M used to check the PSU measures of size was: M=2M, [89 -m) The maximum values of r, (screening sampling rate) for the six density strata (defined by proportion of Mexican-Ameri- can population) are referred to as F, and are shown in table D. The highest value is 6/930, for stratum 6 (the highest density Mexican-American stratum). The next highest value is 5/930 for stratum 5, and so forth. The goal was to avoid combining counties unnecessarily. Increasing the geographic area of a PSU was likely to make a high participation rate more difficult to achieve due to the distance to the mobile examination center (MEC). The ap- proach taken was to keep individual counties as PSU’s when they would not require weighting, or, at most, only a small amount. However, combinations were made when the alterna- tive was to impose large weights. The procedure avoided using the high rates required by the minority density strata for the minimum measures unless there was a reasonable chance that the density strata would actually occur. The value of r, that was used for the minimum measure of a county depended on the proportion of the Mexican-American Table D. Screening sampling rates and number of screened households, by density stratum: Third National Health and Nutrition Examination Survey, 1988-94 Number of screened households Screening sampling Including ~~ Excluding Density stratum rate (f) reserve reserve Total wimcinsben ngs a 158,927 105,950 Yen porn 1/930 82,770 55,180 DL rE ren wk sr 1.8/930 11,569 7.713 8 eesnessvseassie 3/930 22,319 14,880 Boni a EE 2 4/930 14,558 9,705 B imme m5 5/930 14,287 9,524 8 nmmeEaE ERE 6/930 13,424 8,949 population in the county. The value used for a county is shown in the following table: Percent of Mexican-American persons in county Value of r, Lessthan1 ............... 1/930 1-28 simi vnvrnnrrnnnnnnt 2.5/930 8 snemernmmEenrissrraases 5/930 As a result of these rules, the minimum was smaller than it should be for a few counties. Using these rules, about 300 counties were found to have measures of size that were too small and thus were combined with neighboring counties. The effect on the sample was minimal. Selection of certainty PSU’s After assigning measures of size to each PSU, the 13 largest counties (in terms of the measure of size) were included in the sample with certainty, that is, they were designated as self- representing. The cutoff used to identify the certainty PSU’s represented approximately three-eighths of the average stratum size for the noncertainty PSU’s. (Since two PSU’s were selected from each noncertainty stratum, a PSU can be considered as representing half a stratum. A certainty PSU should thus be close to half a stratum size or larger. The three-eighths is equivalent to three-fourths of half a stratum.) For operational purposes, the largest certainty PSU’s were subdivided so that each part would have approximately the same workload (450 sample persons) as in the noncertainty PSU’s. The 13 certainty PSU’s were thus converted to 21 stands. Most of the certainty PSU’s consisted of single stands, as was the case for noncertainty PSU’s. The additional eight stands came from three large PSU's. The 13 certainty counties selected for NHANES III are shown in table 1, along with the number of stands (survey locations) designated for the county and the measures of size (expected sample size). Most of the certainty counties are in California or Texas, reflecting the substantial oversampling of Mexican-Americans. Stratification and selection of noncertainty PSU’s Because there were 13 certainty PSU’s designated, an additional 68 noncertainty PSU’s were necessary to produce an 81-PSU sample. As indicated earlier, a two-PSU-per-stratum sample design was planned. This implied the creation of 34 strata. After selecting the certainty PSU’s, the remaining noncertainty PSU’s were stratified by region, within region by metropolitan status [Standard Metropolitan Statistical Area (SMSA) vs. (non-SMSA)], within metropolitan status by race- ethnicity and finally by income. Within these groups of PSU’s, a total of 34 detailed strata of approximately equal aggregate size were created. The definitions of the 34 noncertainty strata are shown in table 2. Also listed in this table are the sample PSU’s selected from each stratum. To facilitate variance estimation, the noncertainty PSU’s were selected by using the Durbin procedure (12) and involved the following steps: (1) Initially, one PSU was sampled from each stratum follow- ing the usual PPS procedures. These PSU’s were excluded from the master file (frame) of the PSU’s before selecting the second round of PSU's. The PSU’s selected in the first round are denoted by the subscript j in the notation below. (2) Next, within a particular stratum, Durbin probabilities P were computed for each PSU in the frame (other than those previously selected) as Pm =p,[(1-2p,)" + (1-2p,)" where p, = probability of selecting the ith PSU in the stratum (i#)) p, = corresponding probability of the jth previ- ously-selected PSU For those PSU’s that were selected in the first round, Py was set equal to 0. (3) The restricted frame of PSU’s was then sorted by stratum and one PSU was selected with probability proportionate 0p, The two PSU’s sampled in each stratum are identified in table 2. The total of 81 sample PSU’s for NHANES III is shown in the map included with this report. Allocation of PSU’s to time periods To permit separate analyses for the two 3-year periods (1988-91 and 1991-94), as well as for the entire 6-year survey period, the sample of PSU’s was randomly allocated to the two 3-year periods shown in tables 1 and 2. The allocation to the two periods was made in a way that retained as much of the original stratification as possible. Because two noncertainty PSU’s were selected per stratum, one of the PSU’s was randomly assigned to the first time period (phase 1), and the other was assigned to the second period (phase 2). In making the assignments to periods for the certainty PSU’s, the PSU’s were sequenced in a way that brought similar PSU’s together (that is, by region, race-ethnicity-income class, using ascending-descending se- quences), and then the PSU’s were alternately assigned to the two time periods. In the three multiple-stand PSU’s, half the st nds ineach PSU were assigned to each phase; when there was an odd number of stands, one of the phases was chosen at random to have an additional stand. Selection of segments The second stage of the design involved stratification within each of the 81 sample PSU’s and selecting a sample of segments (clusters of housing units). The within-PSU sampling procedures were designed to achieve the target number of sample persons by age-sex-race-ethnicity. The sample sizes shown in table 3 are the desired numbers of examined persons. A much larger sample has to be identified and contacted because some sample persons refuse the examination portion of the survey. Table 4 shows the expected number of sample persons to be identified to produce the examined sample sizes 10 in table 3. Table 4 was prepared by inflating the sample sizes in table 3 by the reciprocals of the expected response rates. Individual race-ethnicity, sex, and age response rates were used. The Mexican-American response rates were based on the Hispanic HANES experience. The other subdomains used NHANES II response rates. Two sampling frames were used to select the sample of housing units within each of the sample PSU’s. For phase one of the survey, the larger area segment frame is based on the 1980 census of the population and is only for a sample of persons who lived in housing units built before 1980. For units built in 1980 and later, the second stage consists of sets of addresses selected from building permits issued in 1980 or later. These are referred to as new construction segments. For phase two of the survey, only 1990 census information is being used. Since phase two will be carried out in late 1991 through 1994, the 1990 census data will be current during the entire interview period, and anew construction frame is not necessary. Stratification within PSU’s The sample size for rare subdomains can be increased by differential sampling within PSU’s. In NHANES III, to reduce the high cost of screening necessary to locate the desired Mexican-Americans for the sample, area segments consist- ing of census block groups (BG’s) and enumeration dist- ricts (ED’s) will be stratified by the percent of the population that is Mexican-American, with a higher rate of selection used in strata containing 3 percent or greater Mexican-American population. Households will also be sampled at variable rates depending on the concentration of Mexican-Americans within the stratum. Essentially, the procedure will involve a basic sample and a supplemental sample as follows: (a) The basic sample will be a national self-weighting sample that is large enough to provide a self-weighting sample for all sex-age subdomains of both black and white and all other persons (that is, non-black and non-Mexican-Ameri- can persons), and for a few sex-age subdomains for Mexi- can-American persons. (b) The supplementation necessary for the increased sample in most sex-age groups for Mexican-Americans will be re- stricted to those BG-ED’s with high Mexican-American populations. (¢) The supplementation will introduce variations in sampling rates that increase design effects. There will be an increase in the number of sample Mexican-Americans for those sex- age cells that were at the 560 level in the original plan to keep the precision at the specified level. This increase will be compensated for by a reduction in sample size in those Mexican-American estimation cells that exceed the 560 level. Therefore, the total sample size will remain at the level specified. (d) The amount of supplementation in the high-density strata will be large enough to supply the number of sample persons needed for the rarest cell, Mexican-American males 60 years of age and over. Other sex-age domains will be subsampled to minimize the variability in sampling rates. This will be done by reducing the sampling rates in the highest density minority stratum until it is the same as the rate in the next highest stratum. For subdomains requir- ing further reduction, the rates in the two highest strata will be reduced, and so on. A detailed discussion of the analysis leading to the ap- proach summarized in items (a)—(d) above is described in another report (11). The principal findings of that analysis are summarized below. Amount of screening without stratification Without differential sampling, the screening costs for the NHANES III would be extremely high. Table 5 shows the screening levels that would be necessary to locate the required numbers of examined sample persons if there were no stratifi- cation of BG-ED’s. For example, to obtain the desired number of female Mexican-American infants 2—11 months old, slightly over 172,000 households would need to be screened under a self-weighting design. For female black infants 2—11 months old, the corresponding screening level is about 67,500. Further, research showed that at the level of 67,500 households, the screening sample would be large enough to provide the desired sample sizes for all age-sex classifications for both black and white and other persons. Optimizing sample size among strata In table 6 there are projections of the 1990 distribution of Mexican-American persons according to the degree of concen- tration in BG-ED’s, adjusted to reflect the changes in the distribution expected to occur in the 1980-90 decade. These distributions were derived from tabulations of the 1980 Bureau of the Census Master Area Reference File (MARF). The MARF tape does not identify Mexican-American persons separately from other Hispanic persons. Hence, the Hispanic distribution in the United States excluding New York, New Jersey, Con- necticut, and Florida was used to approximate the Mexican- American distribution. The Hispanic community in the four excluded States is predominantly Puerto Rican and Cuban. The data were used only to develop the general sample design strategy. The detailed 1989 Bureau of the Census STF-1 tapes were used for sample selection; they contained separate counts of Mexican-American persons, as well as other Hispanic sub- groups. The data in table 6 indicate that most Mexican-American persons live in areas with a high concentration of Mexican- Americans. For example, almost 50 percent of all Mexican- Americans live in areas (BG-ED’s) that are more than 25 percent Mexican-American. The fact that Mexican-Americans tend to be concentrated in certain areas indicates that it is possible to achieve worthwhile reductions in screening by oversampling the more highly concentrated areas. The opti- mum allocation of the sample among the geographical strata shown in table 6 will be restricted to the high-density Mexican- American BG-ED'’s, with the sample in the less dense strata restricted to the part of the basic 67,500 self-weighting sample located in the less dense strata. To facilitate the allocation process, two subsets of strata will be defined. Subset “a” will consist of the least concentrated areas (strata 1 and 2), and subset “b” will consist of the more highly concentrated areas (strata 3-10). The following terms are defined: n, = expected number of sample persons for a given sex-age-race group in subset a from the basic (67,500) screening sample n, = corresponding number to be selected from subset b to meet specified precision requirements. For a particular age-sex-race group, the precision require- ment was expressed in terms of the variance of an estimated proportion, p’. In the subsequent analysis, V is the desired variance of an estimated proportion based on a self-weighting sample of size n, where n is the desired number of sample persons necessary to meet the sample size targets. These sample sizes are shown in table 7 for two selected subgroups. The subgroups given in this table represent two of the rarest groups. They were selected to give an indication of the amount of supplementation necessary with stratification. In the analysis leading up to the optimum allocation, a number of different subgroups were considered in the analysis. However, the deri- vations of the final sampling rates were based on the optimum allocation for elderly Mexican-American males. The sample sizes include the inflation factors necessary to compensate for nonresponse. The value of n fora particular subgroup is determined from the 67,500-household screening sample. In table 7 the n, is shown in the column headed “Expected number of sample persons in low density strata from basic sample. The value of n, can then be determined as follows. First, however, it should be noted that the precision requirement on p’ is: op) =WioX(p',) + WicX(p',)=V where W_ = proportion of the population of interest in subset a W, = proportion of the population of interest in subset b p’, = estimated proportion in subset a p’, = estimated proportion in subset b The required values of V for a 10-percent item are shown in the next-to-last column of table 7. (A 10-percent item was used in variance calculations. However, the optimum allocation holds for all estimated proportions provided that the proportion does not vary widely from stratum to stratum.) It should be noted that 6°(p’)) is fixed because it is determined by the number of cases supplied by the basic screening sample. Thus, the overall precision requirement can be expressed as a require- ment on the variance of p’,, that is, ,, _V-Wioipl) op’) TT we me =V, bh 11 Using these values of V, (shown in the last column of table 7), it follows that to minimize the cost in subset b, the optimum allocation of the sample to stratum 4 (in subset b) is given by the usual formula (13). w,sc, 7, = hn, No Sws NC 7 l J J where W is the proportion of the group of interest in stratum j (in subset b), C; is the corresponding relative unit cost, Sis the standard deviation of the item being estimated (assumed to be the same in all strata), and n, is the total sample to be allocated to subset b. To meet the specified precision requirement, n, was computed from the following formula: (LW, S,NCH2W 5 NC) n= J ) J ) b Vv In table 8 there is a summarization of the n, for two selected subdomains, and the corresponding expected reduction in screen- ing levels with geographical stratification. The analysis discussed above was carried through using a number of different assumptions concerning the overall level of screening. The actual within-stratum sampling rates used in NHANES III have been derived under the assumption that the screening will be at a sufficiently high level to produce the required numbers of Mexican-American males aged 80 years or more, the group requiring the most screening. Stratification of segments Area segments consist of city or suburban blocks (as defined in the most recent census), combinations of two or more blocks, or other area segments in places where block statistics were not produced in the 1980 census. (In PSU’s with examina- tions scheduled for 1991 or later, 1990 census data are being used.) Most of the United States was blocked for the 1980 census. In areas where block statistics were available, segments are single blocks when the measure of size (MOS) of the blocks exceeds a certain minimum. Blocks that are below the minimum are combined with other blocks that are in close geographical pioximity. The combinations are carried out as a computer operation. Within each PSU, the blocks reported on the 1980 census STF-1B file in each minority and density stratum are sorted by tract, block group (BG), and block number. Blocks with MOS below the minimum are combined with succeeding blocks until the desired MOS is achieved. The combinations are kept to the same BG. When the combinations approach the end of a BG without reaching the minimum, earlier blocks within the same BG are added. Consequently, the combinations consist of blocks in close geographical proximity, and in most cases, they are adjacent blocks. As a result of the method of combina- tion, some large blocks that could have been segments by themselves are combined with small blocks. In the nonblocked part of the United States (mostly rural areas), 1980 census enumeration districts (ED’s) generally comprise the segments. They are always the first stage of selection, with small ED’s combined in the same way as small 12 blocks. Where ED’s are unusually large and it appears that they create unreasonable workloads for the person performing the listing operation, they are “chunked.” In a few cases the chunking can be done as an office operation, from information available on maps. More often, however, a field visit is necessary to subdivide the ED into a number of smaller geographical areas. One such area is selected at random. This random selection is taken into account in recording the probability of selection of the segment. Maps of all segments are prepared and they define the areas that are subsequently listed. The new construction sample utilizes a three-stage sample design: (1) PSU’s; (2) clusters of building permits issued during one or several adjoining months by a building permit office; (3) and housing units within the clusters. The sampling rates at the various stages are arranged to provide a self-weighting sample of new construction. All new construction is classified into the nondensity stratum and uses the sampling rates for that stratum. The measures of size are based on the assumption that all residents are in the “white and all other” category and that the housing units are occupied by average-size households. The source of the data used to establish measures of size for the building permit offices and, within each office, the measures for each month or year starting with 1980 is the Bureau of the Census C-40 reports, “Construction Reports—Housing Autho- rized by Building Permits and Public Contracts.” The selection of offices and time periods is performed as an office operation by Westat, Inc. A segment is defined as all residential permits issued per month in a building permit office reporting monthly to the Bureau of the Census or in a year for annual reporters. Where the monthly or annual permits are below a predesignated minimum, consecutive time periods are combined. Within each PSU, the places are listed in sequence, and within place there is a listing of the total segments for each month or year (a segment is based on the number of housing units authorized). With a random start, a systematic sample of segments is selected. To sample specific housing units and obtain the addresses of the selected units, field visits to the sample building permit offices are necessary. Interviewers visit the offices, list all of the permits for the months specified, subsample permits to obtain the equivalent of a single measure following instructions pro- vided, and obtain the addresses of the sample units. When the sample units are located in large apartment houses, the entire building is subsequently listed and subsampled. The same procedures for listing and subsampling are used for the building permit sample and the area sample. The procedures for selecting the segment sample involve both explicit and implicit modes of stratification. The PSU and the minority-density geographical stratacomprise explicit strati- fication. The six density strata within the PSU’s are shown in table D. To keep combined blocks within a single BG, the stratifi- _ cation is done on the basis of the characteristics of the BG or ED in which the segments are located rather than on the specific block or blocks in a segment. This stratification is only applied to area segments. The new construction segments are included -in stratum 1, the nondensity stratum. Within the geographical strata, there is implicit stratification created by sorting the area segments by tract number, BG or ED number within tract, and segment number within BG or ED, and selecting a systematic sample with PPS. The new construction segments are sorted by month and year the permits were issued. The sort order gener- ally introduces a partial effect of stratifying by socio-economic level. Measure of size of segments To describe the procedure for creating measures of size for segments, the following notation is used: The subscripts for the indexes are: h = PSU Mexican-American density stratum ~. Il = segment race-ethnicity subdomains ~~. 1 = sex-age subdomains The parameters used for the computations are: Now = estimated population in the segment N, , = total population in k, /th subdomain in PSU across all segments N ,, = total in U.S. population in k, /th subdomain Rw = sample size in segment Fi = Ru IN, kl = U.S. sampling rate in stratum i for k, /th race- ethnicity, sex-age subdomain M, = measure of size of hth PSU ,; = measure of size of A, i, jth segment M = total measure of stratum in which PSU is located The measure of size of a segment is calculated in the same way as for PSU’s. Research on intraclass correlations and unit costs indicated that an average of 14 examined sample persons per segment is reasonably close to an optimum for most statis- tics in NHANES. Also, as noted earlier, operational require- ments make it necessary to have a fairly constant number of examined sample persons per stand, about 340 in most cases. This implies having about 24 segments per stand. There were 24 segments selected in each noncertainty PSU. The number in certainty stands varied a little from this number depending on the measure of size of the stand. M, is denoted as the measure of size of a segment where M,.. = > Nom The conditional probability of selection of a segment (j) in stra- tum (i) within noncertainty PSU (4) is therefore (24) M,. IM, The sampling rates within a segment are r, M/48M hij. The sample size in the segment is then M Xr 48M. « hij = M/48 hijkl As M/2 is approximately 340, the average segment sample size is 14. Defining the measures of size as indicated above produces an approximately constant number of sample persons per segment. A similar strategy was followed in the certainty PSU’s, designed to produce about the same sample size per segment and the same ultimate probability in each subdomain as in the noncertainty PSU’s. As indicated at the beginning of this section on measure of size of segments, N, is defined as the number of persons in a segment in each race-ethnicity-sex-age subdomain, and N, , is the total number in the subdomain in the PSU. The current population of the PSU or segment is not known in such detail; the sample selection therefore uses 1980 data, except for the census updates of the total county population. The Bureau of Census’s estimate of the current population of a county is denoted by C, and the 1980 count by C’,. (The 1985 census updates were used for the values of C,.) Similarly, the current U.S. population in a subdomain is denoted by C , and the 1980 count by C’ | . The race-ethnicity totals are C and C’ ,. The estimates of the current population of a PSU by race-ethnicity are estimated by: For purposes of sample selection, itis assumed that the age- sex distribution within a race-ethnicity group in a segment conforms to the current U.S. distribution rather than resembling the 1980 distribution in the segment. Consequently, C 4S ’ N = hijkl ’ hijk. c «nail c h The following simplifications can be made in the compu- tations of the measures of size. M hij 2 ikl Nw ©, C2, z # ' Tk / Cc, | | & ™M a where A, ll t oy The values of A, used in calculating the measures of size are shown in table 9. The measures of size also used the sampling rates r,, which differ in the various high-density Mexican-American areas (table 10). 13 Departures from self-weighting sample The development above makes a number of assumptions in demonstrating that the measures of size will provide a self- weighting sample with equal size samples in all PSU’s and segments. The assumptions, of course, do not apply exactly. Deviations from exact sample sizes in segments are permitted to retain the self-weighting features of the sample, except for a few unusual outliers. The number of sample persons per PSU, however, is fixed in advance and can not be changed. To retain the preassigned workload, some variation among PSU’s in sampling rates was necessary, primarily caused by the follow- ing factors: ® There is an assumption of equal size strata, equal to M*/66, where M* is the total measure of size for all PSU’s in the United States. In practice there is some variability in stratum sizes. For a self-weighting sample, the variable stratum sizes should be reflected in variable sample sizes per PSU, and thus, per segment. ® For equality in workloads, it is necessary for the current proportion of the population in each race-ethnic group in a PSU and in a segment to be the same as in 1980. ® The proportion of each race-ethnic group living in high- density Mexican-American areas is estimated by using 1980 data with some attrition based on earlier experience. ® The measures of size treat the age-sex distribution for a race-ethnic group as being identical in all segments and PSU’s. ® Assumptions are made about the nonresponse and cover- age rate for each subdomain, and it is assumed these rates apply in all PSU’s. These assumptions may not hold ex- actly. Number of segments and probability of selection The discussion above indicates that there are 24 segments in most stands, and that the within segment rates provide a uniform sampling rate across all PSU’s if the sizes of the strata used in the PSU selection are equal. Although the strata sizes can not be made equal, the range is fairly low. The sample selection is based on 24 segments per stand. Because the measures of size of the certainty stands were not equal to half the nc ncertainty strata, all of them did not have 24 segments. The sample for the two phases consists of 2,138 segments in all 89 stands in the sample. The actual probability of selection of a segment depends on the measure of size of the segment, the measure of the PSU, and the total measure of the stratum from which the PSU is selected. The following terms apply in determining the probability of selection: measure of size of a segment in the Ath PSU in stratum a ahij P, = measure of the PSU that was used in sample selection P = stratum size The probability of selection of the PSU was 2P , /P . The probability of a segment within the PSU is: 14 24M - ahij XM if ahij The overall probability of selection of a segment is: M,. 2. ABC yh) 2M i ahij a It can be noted that P , is approximately equal to M 7 ahij The segment measures of size implicitly include provision for the required oversampling in minority density strata. For example, if the sampling rates in one density stratum are twice those in another, then the measures are twice as large. Minimum measure of size of segments One of the goals of the sample design is to create equal probabilities of selection for each domain, within each density stratum, within a PSU. To create equal probabilities, the within- segment sampling rate for a domain in noncertainty PSU's should be: dM, P 7 ahij a A. = i.) 48M ahij ah In certainty areas it is: >M 1 Su - (r.) nr 24M ahij ahij To avoid creating special weights, the within-segment sam- pling rates need to be ahij a < 1 hijkl i 48 M ahij ah and PIM > Sali M,., = 48 P ah ahij = a In certainty areas, the requirement is y 2M, Mu 2 i 24 In some of the certainty counties, the 24 in the denominator is replaced by the number of segments designated for the PSU, in most cases 48 for two-stand counties and 72 for three-stand counties. Controlling sample size per PSU To implement the sample segment selection, the minimum measure is made 50 percent greater to permit a reserve 50- percent sample to be selected. The procedure for controlling the sample sizes in the PSU’s and calculating the weighting factors is described below. The sample size in each PSU that will result from a self- weighting sample in each domain within each density stratum is derived. This number is based on several assumptions that are expected to hold only approximately. However, once calcula- tions are prepared of the sample sizes, they are treated as quotas and the number of sample persons in each PSU must adhere to the quota. The calculation of the sample size goals within a PSU is described below. The probability of selecting anoncertainty PSU is 2M ,/M where M , is the measure of size of the 4th PSU in stratum a. For any domain-density stratum, a constant sampling rate in the United States is desired. This rate is denoted by r, . Within the sample PSU, the sampling rateis 2r, ,M /M ,. The total number of sample persons in a PSU is hI No M, 12M ikl 2 where N, is the total population in the PSU in class 7, k, [. On the assumption that the population distribution will be approxi- mately the same as in 1980, M , = Xr N, and the sample size will be M_/2. In certainty PSU's, the sample size is approxi- mately equal to Xr, N, = M,. The quotas (goals) assigned to the PSU's are thus proportionate to the measures of size of the strata from which the PSU is selected. If M is the total measure of size of all PSU’s in the United States, the quotas for the PSU’s are calculated as shown below. (The total quota shown in the formula below, 40,561, is taken from table 4.) For noncertainty PSU’s, the quota is M (40,561) 2M For certainty PSU’s, the quota is M, —"_ (40,561) M The stratification established for the PSU selection keeps the values of M, within fairly narrow bounds. Thus, in noncertainty PSU’s the quotas do not vary substantially from the average of 456, which is 40,561 divided by the 89 total stands in the sample. There will be greater variation among the certainty PSU's. As there is a constant number of segments per PSU (24) in noncertainty areas, the variation in quotas per PSU is also reflected in segment sample sizes. In addition, since 1980 the changes of the population distribution among segments is likely to be greater than among PSU's. Thus, more variation can be expected in the average segment size than in PSU’s, but even this should be within a moderate range. The approximate equality that exists in sample sizes per PSU and segment does not occur in the screening sample. Considerable variation can be expected. The amount of screen- ing per segment varies considerably among the density strata. About half of the screening will be in the minority density strata; therefore, the amount of screening in a PSU is partially based on what part of the population lives in high-density strata. The number of sample persons in a generated PSU depends upon several factors that include the race-ethnicity breakdown in the PSU, the age distributions, and the proportion of Mexi- can-American persons living in the various density strata. A set of assumptions is made about these factors to permit the sampling operations to proceed. However, there is no way of knowing in advance whether the assigned quota for a particular PSU is lower or higher than what would arise from self- weighting samples within the various domains and density strata. Consequently, it is necessary to have a sample selection procedure that can produce samples either somewhat larger or smaller than those arising from the application of the self- weighting sampling rates. Initially a screening sample that uses sampling rates 50 percent larger than those for self-weighting samples is desig- nated in each PSU. This screening sample in each PSU is then divided into a group of subsamples, referred to as “stop-rule” groups. Each subsample is a systematic subsample of the screeners, with the screeners sequenced prior to subsampling in the following order: density stratum, segment number, and household number. Each subsample thus cuts across all seg- ments and density strata. The stop-rule groups and the percent of households in each subsample are shown in table E. The 50-percent subsample is released first to the interview- ers. When the initial assignment is about 75 percent complete, the resulting yield is analyzed and used to project estimates of the total number of sample persons expected from the initial assignment. Based on these estimates, additional subsamples Table E. Percent of Mexican-Americans in segment by density stratum number: Third National Health and Nutrition Examination Survey, 1988-94 Percent of Mexican- Density stratum number Americans in segment Nondensity, less than 3 D ouicnmenan es RE RE § eh ee EAA 34.9 B iunanzanne nnn nai 5 Esha mamma 5-9.9 Borin BRE HE 5 xh 55 Rs 25S BEE RETR 10-19.9 Bi sit vm Sn Ea BH a § RR ES 20-49.9 81.112 cpmsmsrmsmuommnis 3am vm Br Bi a 2 eons hase SAE A 50 or'more 15 are released. As additional households are screened, the deci- sion on the number of subsamples required is reevaluated to ascertain if more households will be necessary to achieve the target number of sample persons. If so, additional subsamples are released. The reevaluation is done continuously. A count is kept of the number and size of subsamples used. This will provide the information necessary to calculate the PSU sample weights that will reflect the deviations in sample sizes from self-weighting samples used within the PSU's. Selection of households and persons The third stage of sample selection consists of households and certain types of group quarters. All households in the sample segments are listed, and a subsample of households and group quarters is designated for screening to identify potential sample persons for interviews and examinations. The subsampling rates are designed to produce a national, approxi- mately equal, probability sample of households in most of the United States, and higher rates for the geographical strata with high minority concentrations. Within each geographical stra- tum, there is an approximately equal probability sample of households across all 81 PSUs. A constant sampling rate for the screened households is desired within each density stratum (subject to the stop-rule - modification). The screening sample in each density stratum must equal the highest rate among all subdomains for the screening sample to yield the desired number of sample persons in all subdomains. These screening rates denoted by r, are shown in table D. Applying these sampling rates to the expected number of occupied housing units in the various density strata provides an estimate of the number of households to be screened. They are shown in the last column of table D. For domains with the maximum sampling rates in a density stratum, subsampling is not required. For other domains, how- ever, subsampling reduces the screening sample to the rates shown in table 10. The subsampling rates are the ratios of the sampling rates for a domain divided by 7 These are shown in table 11. W. hin-segment sampling rates To achieve equal probability of selection within a density stratum, the subsampling rate within a segment must be /prob Si == Fin hij where §, is the selection probability for the , /th subdomain in the jth segment in the ith density stratum in the 4th PSU and r,, is the overall probability of selection, as reported in table 10. Prob, is the probability of selection of the jth segment. As stated earlier, the probability of a segment in a noncertainty PSU is M, P, = 48 ahij ah hij - Sm LJ Prob ahij a 16 where P , is the measure of the PSU and P is the measure of the stratum from which the PSU was selected. In certainty PSU’s, the probability was hij 24 hi Sm hij Parameters used in computing measures of size One of the goals in computing measures of size is to create approximately equal workloads among PSU's and segments. Except for differences arising from variation in the size of the strata used for PSU selection, which were fairly small, equality is achieved by the following measures of size: For PSU's M, = Xe. NF, where M, is the measure of the 4th PSU; P, is the proportion of the population in the PSU that is in the ith combination of subdomains and density strata; and r, is the sampling rate for that subdomain-density stratum. For segments M,, = ry No As noted earlier, the values of P,, N,, and N, must be approximated and use the following formulas: and C where A, = Y-. ot and with = current population of PSU A = 1980 population of PSU A = 1980 PSU population in kth race-ethnicity subdomain = sampling rate in the ith density stratum for &, Ith subdomain P, = U.S. proportion in the ith density stratum for the kth race-ethnicity subdomain C,, = projected 1990 U.S. population in the , /th subdomain C , = projected 1990 U.S. population in the kth race-ethnicity subdomain. These approximations are used because it is likely that race-ethnicity composition in 1980 is an accurate projection of the 1990 data, but the 1980 age-sex breakdowns may be poor projections of the 1990 data. The assumption that the national age-sex breakdowns within a race-ethnicity group apply to each PSU and segment seems more reasonable than that the 1980 age-sex distributions are retained in 1990. The values of the parameters used to calculate A, and A, are shown in tables 9 and 12, respectively. Selection of sample persons After the sample of screened households are identified, a sample of persons to be interviewed and examined from indi- vidual households is selected. All eligible members (persons 2 months of age and older) within a household are listed and a subsample of individuals is selected based on sex, age, and race- ethnicity. Sample persons are selected at rates established to ensure that the target sample sizes by subdomain will be achieved. This means that young persons, elderly persons, black persons, and Mexican-Americans are oversampled. The sample is also selected to maximize the average number of sample persons per household because it appeared to increase the overall participation rate in previous surveys. The 52 analytical subdomains were collapsed into 16 groups with acommon sampling rate used for each group. Table 10 shows the sampling rates used for the 16 groups of subdomains in the six density strata. These sampling rates are designed to provide a 50-percent reserve sample, as well as a provision for the expected nonresponse in each subdomain. Sampling rates were calculated for the subdomain in each race-ethnicity group that requires the highest sampling rate to achieve the desired sample size. The calculation is based on the optimum allocation method described in the “Selection of segments” section. These subdomains are in the collapsed classes assigned to the lowest domain numbers for each race- ethnicity group. (The collapsed classes were numbered in descending order of sampling rates, and thus the one with the highest sampling rate appears firstin table 10.) These maximum rates determine the screening sample. In each density stratum, a sample of households to be screened is selected at the highest one of these rates that appeared for that density stratum. All screened persons in the subdomain used for optimum allocation are retained in the sample. The screened persons in other subdomains are subsampled to bring the samples down to the desired levels. The subsampling rates were designed to mini- mize the variability in sampling rates among strata, but still achieve the desired sample sizes, and thus the required preci- sion. This was accomplished by progressively reducing the sampling rates in the highest density domains to equal the ones in the lower density domains to the extent this could be done and the desired sample sizes still be attained. There was considerable subsampling needed to reduce the screened sample of 106,000 households (which con- tain about 285,000 persons) to about 41,000 sample persons. If independent random or systematic selections had been made for the subdomains, in most cases only one person in a household would have been selected and the average sample size per household would have been quite low, not much above one. Experience with recent NHANES and Hispanic HANES indicates that response rates improve when larger sample sizes within households are used. Therefore, amethod of subsampling was developed to increase the average sample size per house- hold. The sampling procedure, described later in this report, appears to maximize the number of sample persons per house- hold. (Conversely, it minimizes the number of households containing sample persons.) Assuming that a screening sample has been designated and persons are to be subsampled, the persons are classified into L subdomains with subsampling rates r, . . . r,. The subdomains are ordered by subsampling rate so that WwW pT Wa i=1 where N, is the census estimate of the civilian, noninstitutional population in domain k, and the summation in the denominator is over all sample persons in domain &. The final examination weight (w) for each sample person (i) is the product of the above three component weights: W, = waw,w, (5) The adjusted weights will be smoothed even further to ensure that there are no extremely small or large sampling weights. For each sample person, a final examination weight reflect- ing the unequal probabilities of selection, adjustments for nonresponse, and poststratification will be included in the public-use data tapes. Similarly, an interview weight and any special subsample weights will also be included in the data tapes. Most of the commonly used computer software packages have an option for incorporating sample weights in cross- tabulations and statistical analyses. Variance estimation The NHANES III is based on a complex sample design. The assumption of simple random sampling for estimating variances is not appropriate because it would result in estimates of variances for most items that are lower than those actually present. Design effects are often used to gauge the effects of the various sampling techniques, such as clustering and stratifica- tion, and they provide an indication of the success of the complex sample in controlling the variances of the estimates compared with simple random samples. A design effect is defined as the ratio of the actual variance of an estimate from a complex sample to the expected variance of the same estimate, if the sample were drawn from a simple random sample. When the design effect is close to 1.0, the complex sample design is determined to have little effect on the variances. Analysts could consider assuming simple random sampling for data analysis. The use of average design effects is discussed in the last section on variances for subdomains. Because of adjustments for nonresponse and post- stratification in NHANES III, precise formulas for com- puting sampling variances from the complex survey are not available. However, there are several methods that can provide good approximations for the sampling variance. For a variance approximation to be satisfactory, the variance estimates must reflect all the major features of the sample design used in the survey, including the weighting of the sample. data. The three methods generally used for variance estimation .with complex samples are Taylor-linearization, balanced repeated replication (BRR), and jackknife (17,18). Generally, the differ- ent approximations give similar estimates for sampling vari- ances. No approximation method is substantially better, in all circumstances, than any other method. Software available for linearization methods may not provide for poststratification. This can lead to serious overstatements of the variances for some statistics. In addition, linearization methods often do not take into account the variance effects of nonresponse adjustments because of the difficulty in expressing the adjustment methods in algebraic form. BRR and jackknife can handle poststratification and nonresponse adjustments more easily, although complete reweighting must be done for replicates. For NHANES III, two PSU’s were selected from each noncertainty stratum, making it possible to compute virtually unbiased estimates of variances. NHANES III analytical data tapes will include variance strata and replicate weights that can be used for either BRR or linear approximation, thereby allowing users choices in variance estimation procedures. Com- puter software packages that compute appropriate standard errors of estimates from surveys with a complex sample design are available. Previously, NCHS used the NCHS BRR program (19) to calculate variances for data collected in NHANES. Recently, analysts of NHANES data have also used two SAS procedures, SESUDAAN and SURREGR (20,21), which depend on a Taylor-series approximation. Other soft- ware packages available include SUPER CARP (22), a pro- gram developed at Iowa State University that depends on a Taylor-series approximation; Wesvar (23), which was devel- oped by Westat, Inc., that can be used for BRR or jackknife; and OSIRIS, developed by the University of Michigan (24), which contains procedures for either linearization, BRR, or jackknife. A new software package, “Software for SUrvey DAta ANalysis” (SUDAAN), has been developed by the Research Triangle Institute in cooperation with NCHS and other agencies of the Public Health Service (25). SUDAAN uses the first-order Taylor series approximation to determine estimates of standard errors for means and proportions (and differences in means and proportions) with appropriate corrections for complex survey designs, including poststratification. One advantage that this program has over other linearization software pack- ages is that it allows analysts to incorporate the actual complex sample design of the survey in the calculation of standard errors, for example, the joint probabilities of selec- tion for each pair of PSU’s and whether the sample was selec- ted with or without replacement. In addition, the software is available for personal computers as well as for mainframe computers. Variance estimates for each phase In the allocation of noncertainty PSU’s to the two phases of NHANESIIII, one PSU in each stratum (selected at random) was assigned to phase 1 and the other one to phase 2. Thus, the sample design for each phase was a one-PSU-per-stratum selection. There is no completely unbiased method of estimat- ing variances for such a design. A common approximation is to pair strata that are similar and to compute variances as if the two selected strata from each pair had been selected from a single stratum, with replacement. These pairings will also be indicated on the data tapes. Variances for subdomains For some subdomain analyses in NHANES III, estimates may be based on small sample sizes or come from a small number of PSU's. The variance estimates for these statistics are likely to be unstable, that is, the estimates of variances may themselves be subject to high variability. In this situation, the approach often used is to compute an average design effect to correct estimates of variances based on the assumption of simple random sampling. This was the recommended proce- dure for analysis of data collected in Hispanic HANES, which was a survey of three special Hispanic subgroups in selected areas of the United States (26, 27). This strategy may also be advisable for some subdomain estimates in NHANES III. 21 References 10. 11. 12. 13. 14. 22 . National Center for Health Statistics. Plan and initial program of the Health Examination Survey. National Center for Health Statistics. Vital Health Stat 1(4). 1965. National Center for Health Statistics. Plan, operation, and re- sponse results of a program of children’s examinations. National Center for Health Statistics. Vital Health Stat 1(5). 1968. National Center for Health Statistics. Plan and operation of a health examination survey of U.S. youths 12-17 years of age. National Center for Health Statistics. Vital Health Stat 1(8). 1969. Miller HW. Plan and operation of the Health and Nutrition Examination Survey, United States, 1971-73. National Center for Health Statistics. Vital Health Stat 1(10a). 1978. . National Center for Health Statistics. Plan and operation of the Health and Nutrition Examination Survey, United States, 1971— 73. National Center for Health Statistics. Vital Health Stat 1(10b). 1977. Engel A, Murphy RS, Maurer K, Collins E. Plan and operation of the HANES I Augmentation Survey of Adults 25-74 Years, United States, 1974-75. National Center for Health Statistics. Vital Health Stat 1(14). 1978. . McDowell A, Engel A, Massey JT, Maurer K. Plan and operation of the second National Health and Nutrition Examination Sur- vey, 1976-80. National Center for Health Statistics. Vital Health Stat 1(15). 1981. . Maurer KR. Plan and operation of the Hispanic Health and Nutrition Examination Survey, 1982-84. National Center for Health Statistics. Vital Health Stat 1(19). 1985. . National Center for Health Statistics. Plan and operation of the third National Health and Nutrition Examination Survey, 1988— 94. Vital Health Stat. To be published. Chu A, Waksberg J. NHANES III Sample Design—Final Re- port. Contract No. 282-86-0042. Rockville, Md., Westat, Inc., 1988. Chu A, Lannom L, Morganstein D, Waksberg J. NHANES III Methods Research—Final Report. Contract No. 282-86-0042. Rockville, Md., Westat, Inc., 1989. Durbin J: Design of multistage surveys for estimation of sam- pling errors. Applied Stat 16:152-64. 1967. Hansen MH, Hurwitz WN, Madow WG. Sample Survey Meth- ods and Theory. New York. John Wiley and Sons, Inc. 1953. Landis JR, Lepkowski JM, Eklund SA, Stehouwer SA. A statis- tical methodology for analyzing data from a complex survey, the 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. first National Health and Nutrition Examination Survey. Na- tional Center for Health Statistics. Vital Health Stat 2(92). 1982. . Forthofer RN. Investigation of nonresponse bias in NHANES II. Amer J of Epidemiology 117:507-15. 1983. Ford, BL. An Overview of Hot-Deck Procedures. In: Madow WG, Olkin I, and Rubin D, eds. Incomplete Data in Sample Surveys, Volume 2: Theory and Bibliographies. New York. Academic Press, Inc., 185-207. 1983. Rust K. Variance estimation for complex estimators in sample surveys. J Official Stat 1: 381-97, 1985. Wolter, KM. Introduction to variance estimation. New York: Springer-Verlag. 1985. Jones G. The NCHS BRR Program. Unpublished document. Shah, BV. SESUDAAN: Standard Errors Program for Comput- ing of Standardized Rates from Sample Survey Data. Research Triangle Institute, Research Triangle Park, NC. 1981. Holt, MM. SURREGR: Standard Errors of Regression Coeffi- cients from Sample Survey Data. Research Triangle Institute, Research Triangle Park, NC. 1977. (Revised April 1982 by B.V. Shah.) Hidiroglou MA, Fuller WA, and Hickman RD. SUPER CARP., Sixth Ed. Survey Section, Statistical Laboratory. Iowa State University. Ames, Oct., 1980. Flyer P, Rust K, and Morganstein D. Complex survey variance estimation and contingency table analysis using replication. 1989 Proceedings of the Survey Research Methods Section of the American Statistical Association, 110-19. 1990. Survey Research Center Computer Support Group: OSIRIS IV User’s Manual, Institute for Social Research. University of Michigan, Ann Arbor. 1979. SUDAAN: Software for Survey Data Analysis. Research Tri- angle Institute, Research Triangle Park, NC. 1989. Kovar MG, Johnson C. Design effects from the Mexican-Ameri- can portion of the Hispanic Health and Nutrition Examination Survey. 1986 Proceedings of the Survey Research Methods Section of the American Statistical Association, 396-9. 1987. Lago J and others. Evaluation of design effects for the Mexican- American portion of the Hispanic Health and Nutrition Examina- tion Survey. 1987 Proceedings of the Survey Research Methods Section of the American Statistical Association, 595-600. 1988. List of detailed tables Number of stands (survey locations) and PSU measures of size by certainty counties: Third National Health and Nutri- tion Examination Survey, 1988-94 Noncertainty stratum measure of size, phase, county, and state: Third National Health and Nutrition Examination Survey, 1988-04. ....c.cuvuvnmernsnsommssize ssn se Number of examined persons, by race-ethnicity, sex, and age: Third National Health and Nutrition Examination Sur- VEY, 1988-04 . cv ivmnmnvnn sins vimmnns vrs suns sss Number of sample persons, by race-ethnicity, sex, and age: Third National Health and Nutrition Examination Survey, VOBBOG .oniis vinnimin ais monn ares ws 0 5% ww» wo 3059 Hp 815 51 Number of households to be screened without stratification to meet minimum sample size requirement per consolidated class after allowance for nonresponse, by race-ethnicity, sex, and age: Third National Health and Nutrition Examina- tion Survey, 1988-94 Projections of the 1990 U.S. Mexican-American population by stratum number, according to the percent in block group or enumeration district, number in thousands, and percent distribution: Third National Health and Nutrition Examin- ation Survey, 1988-94 Number of sample persons for self-weighting sample, level of screening, proportion of subgroup in low-density stratum, expected sample sizes, desired variances, and re- quired variances in high-density stratum, by selected 24 25 27 28 29 29 12. subdomains: Third National Health and Nutrition Examination Survey, 1988-94 ............c.vvnnnnn. Sample sizes and screening levels for self-weighting and stratified samples, by selected subdomains: Third National Health and Nutrition Examination Survey, 1988-94 Values of A, used in calculating measures of size, by race- ethnicity and density stratum: Third National Health and Nutrition Examination Survey, 1988-94 .............. Sampling rates and expected sample sizes, by Mexican- American density stratum, race-ethnicity, and sex-age do- mains: Third National Health and Nutrition Examination Survey, 1988-94 ...... . Domain subsampling rates within screened sample, by Mexican-American density stratum, race-ethnicity, and sex-age domains: Third National Health and Nutrition Examination Survey, 1988-94 ...................... Values of parameters used to estimate A, for primary sam- pling unit measure of size, by race-ethnicity domain: Third National Health and Nutrition Examination Survey, V9BB-I: | wee iien mani sis ums 50003 300 38 5 5 # eae » oom ow 0 00 . Proportion of households with each sampling message label describing which household member to include in the sam- ple, by density stratum and race-ethnicity domains: Third National Health and Nutrition Examination Survey, T9802... conn pmumsmes tne sosmers ens 55 EwEsbmnmm 30 30 30 31 32 33 23 Table 1. Number of stands (survey locations) and PSU measures of size by certainty counties: Third National Health and Nutrition Examination Survey, 1988-94 2 Number of stands (survey locations) PaLr fond stratum) measure Certainty counties . Both phases Phase 1 Phase 2 of size? TOA 5 : 55 amr mmmnmmnuna u § 58 £ 18 & 7 & Hessen mimes ie dg 5.5 70 4 fm vote enka 21 10 11 14,452 Maricopa, AZ APNOBIIXY +ucunsviisisrrives sasvvssmesssss es sea ayaesas vr is 1 1 638 LOS ABGBIES, CA. + mrsssiminior nn s.0 on sans TURRET 1-5 esos ss 3 gh HE SRB SH 0) 6 3 3 4,546 OTANNG, CA cvs vonpmmmmnin bs rminitia dn» wm #00 0 i ns i onsen v5 0 0 1 1 726 San DISGO,/CA . i »oilitimmpn ares vd 8 5 03 54% 8 44 DESERET aay frist 4 5302 508 55 1 - 1 781 SAMA CIARA, CA . «on vim ss 68 dv 5s 5858 5500 WRmmmlaios 54 EET s§ 3155 1% 1 1 540 COOK, IL. (CHICAGO) » =nrmmnsamoniei sein 5 BEE 37 255 0m = nm im meres BRA 53 H 5 25% #2 3 m5 ww 3 1 2 1,851 Wayne, MI (Detroit) ...........coiiiii 1 1 - 707 KINGS, NY (BIOORIVNY. ...cssacss sons bas a8 8 onder dT Emami ns « 5u 5 soda #56 556 1 - 1 673 Philadelphia, PA. .vicivensmeme en shen 20 5 barssssismems be vweest s SEES $5 bimwai 1 - 524 Bexar, TX (San Antonio) . .. o.oo 1 1 - 976 DEBEE, TX ivi eminem op ow abies oot iio so GE fim sees 3 1 1 650 EIPaso, TH :iiissosnmmamumsennirais 6506008 5susmmhme esas dif i 1 - 1 596 HAs, TX (HOUSIONY wuimw ssn sk 4 5 5 55 75 3.3% 5 5 9 esses iysisy £595 + 4% 3 2 1 1 1,244 "Primary sampling unit. Expected sample size in PSU. 24 Table 2. Noncertainty stratum measure of size, phase, county, and state: Third National Health and Nutrition Examination Survey, 1988-94 Metropolitan status Stratum and substratum description measure of size Phase County State Region 1 Less than $8,500 per capita: NONSMBA. cmseroistimns vos £5 05 £5 BEARERS § 1165 1 Oswego NY NONSMBA 5st sve viv £5 5 28 vrmmmen spain ses 2 Androscoggin ME $8,500 or more per capita: NOBBMBR. ceive opt Bia Sten neimssmmmis teRor 1333 : 1 Providence RI NONSMEBAL «coos s wire a EE HRB we es 2 Northampton PA 18 percent or more black persons: TRE 1406 ! 1 Queens NY BMBA iciiimes #0 055 54 % 5 8 SH ERRRRERREE 1 $25 2 New York NY 10.2-18 percent black persons: . MBA! cerrasein 12 5 BBE 2 5 2 ale 2 0 he wie esmemnie sna 7 £ 5 3 3 1214 1 Allegheny PA BUSA, ovis v5 255 32 55585 5 emmaeinmamsinsnnten 5% 3 2 Westchester NY Less than $11,000 per capita: SMSA Loi 1247 : 1 Erie NY BISA: 0 505s TEER: 0 mo wi ois HH PIRES 2 Delaware PA $11,000 or more per capita: SMSA ©. 1398 : 1 Middlesex MA BMBA. isn svt unineme bd hyn 2 Nassau NY Region 2 Less than $8,000 per capita: NONSMSA ete eee eee 1250 { 1 Douglas MN NONSMSA 2 Caldwell MO $8,000 or more per capita: NONSMSA ieee eee 1255 1 Vermilion IL NONSMSA 2 Crawford IL 17.5 percent or more black persons: SMSA 1357 1 Hamilton OH BMSBA. ...ccooiirmstimiv pn pairs vom mesma SEE 1 2 Cuyahoga OH 10.6-17.5 percent black persons: BISA: imino v's £3 7 47 2 BHA ARE 1310 1 St. Louis MO BMBA. sonnmrimisnans vv vv 3 55 3% 3 SPEIER Sa 1 2 Lucus OH 6—10.5 percent black persons: BISA remit eA RD 8s 2 ne ie emma 1222 1 Ingham Mi 2 2 Allen OH Less than $10,000 per capita: BMBA. covets i Ee memsmeneen = i iin wen os FEES 1219 1 Clermont OH BMSA: conn ivmn sen ERITDIER Ras inns hee lam inn 2 Minnehaha SD $10,000 or more per capita: BMS ooccsiem i tether mss Sea Sik 1332 1 Oakland MI BMBA. coordi teh SHE 3B swe esr oe mens meses 2 Kane IL Region 3 40 percent or more black persons: NOBSMBA .. ...comicitemmtin 124251 « 22 = 00s mE 1306 1 Granville NC NONSMBA “: io « + mnimmsimmsmminnin s 5 52 3 2 £2 § 8205 © insmisio 2 Montgomery MS 26.540 percent DIBCK PBISONS. ......« «us i+ 323s swniminns 4 SP 1364 1 Chester SC NONSMSA . .. 0.0 5 csininns wns Easiness ss 3 36 ny Benes 2 Spalding GA 9-26.4 percent black persons: NOTES. come sma 2:4 258 hm nm tvsamrie merase 1433 : 1 Alcorn MS WORSMSA. cvs Sie 5 48 5S Bisse dntuns mshi io asm 2 Gibson TN Other: NONSMSA «iia 1376 1 Owen KY NONSMBA ..ccovmrnnnna@u i F285 E55 4 #7 8 2ansmweivne 2 Warren KY 42 percent or more black persons: BMA. . «+ 2 5 5 5B BREET mon + m+n 0 ws 2 5 2 RB ES 1246 ! 1 Hinds MS SMEAR on eis nnn mnt r bliss #55 3 5% 55 3 4 2 samawnni 2 Twiggs GA 34.5-42 percent black persons: BMBA. . ves ve ndibiscre wise mniam rma eis sss ss subs 1450 ! 1 Henrico Area VA BOMBA. cova e vii 5780000 EB fF Bela patmaasmnss 58 Ri 2 W. Baton Rouge LA 24.5-34.4 percent black persons: SMSA 1507 { 1 Duval FL BMSA: «cnr mt Sig TRB SA 11 8 ees Bema RR 2 Mecklenburg NC 17.5-20.4 percent black persons: ...................... SMBA. . : ; 5 sania» 5» 25% 8 § Elbe 1378 : 1 Forsyth NC SMBA: : . oo 5 wisrommT EE & & wn +0 8% «35% 5 METRES 2 Jackson MS Table 2. Noncertainty stratum measure of size, phase, county, and state: Third National Health and Nutrition Examination Survey, 1988-94—Con. Metropolitan status Stratum and substratum description measure of size Phase County State 13.4-17.4 percent black persons: SMSA oo 1541 1 Palm Beach FL BMSA. +. «+ co ian BEET 4 8 Hie 5 2.5 5 5% 5mm 2 Dade FL $11,000 per capita: SMSA 1387 1 Yadkin NC BMBA. oc cvimcmimmnsomieinin ls 5.5 55 5 § & #Fiemes-aonrasemmomsminiamasn 2 Clayton GA $11,000 OF MOI Br CAPRA: «us vs n ins ms mii snmianss SMSA eee 1484 1 Fairfax Area VA SMBA. (515 5125 ne sragprasstaiis mess STEERER 3 3 ape wtamsesnmnmes 2 Cobb GA Region 4 29.5 percent or more Mexican-American persons: NONSMSBA .. . .«ommissininisi 5555550 + 50 20 0 wgriowiwimomsminen 1282 1 Luna NM NOASMSA . ..:ooswommanassssasasssass os raparems 2 LaSalle TX 6-29.4 percent Mexican-American persons: NBAEMISE. o.oo ore it £250 5 2125 Spain steamers in 1271 1 Merced CA NONSMSA. . civsviimrnimiew on «36 5 4 fastens SRR ERLE 2 Jackson TX Other persons: INOHSMBA. . . scored i Traits srg emesis 1186 1 Hamilton ™ NONSMBA . i: 55 eussems umes ns 4 5o0 tps es 2 Hopkins TX 33 percent or more Mexican-American persons: ..................ouuiniinn NI 1416 1 Hidalgo TX SMBA. + is wus BeOS 5 HSU 5x ek RENE 2 Cameron TX 20.4-33 percent Mexican-American persons: SMEBA: . ovr rn GATHEAA SE HA EE 558 0 EER BE MAAR 1568 : 1 Monterey CA SMSA. . cippnmipmmaamenis 51 4 § 2s § 2 8 3 REEWRE vw 2 Fresno CA 17.3-20.3 percent Mexican-American persons: SMSA © 1511 1 San Joaquin CA BMEBA, oo nmmmirse mane i wi wig aT IES ER 30] mms wa 2 Ventura CA 9.5-17.2 percent Mexican-American persons: SMSA ©. 1626 : 1 San Bernardino CA BMBA. . . in vcevisimsnB EET 2 5 2 355 28 8 +b mmmmnn a 2 Travis TX Less than $10,000 per capita: SMSA 1415 : 1 Bell yi BMSA. . . commmmmmmnrs nn ik 24833 15% 3 # summers 2 Sonoma CA $10-11,000 per capita: SMBA, ennmemareimns 20s 52s 52 5 22 5 ARERR 1521 : 1 King WA BMSA civics aan ee 2 Tarrant TX $11,000 or more per capita: BMBA. oo wins Ew. #5 Airis EEE 1524 1 Alameda CA SMEA: 1.555 mim wmimsm am wa sss ats 11 4 8 57% # www eine 2 Anchorage AK NOTE: SMSA is Standard Metropolitan Statistical Area. 26 Table 3. Number of examined persons, by race-ethnicity, sex, and age: Third National Health and Nutrition Examination Survey, 1988-94 Number of examined persons White Mexican- Sex and age Total Black and all other American’ BOHVEBREE. ..ovevr on 5.50005 0080.5 FE eo hpasait nde 3.0. ATE TM HE SH RATE 30,094 8,971 12,123 9,000 Male BITTER 0 oa FEET St onto 30.9410 1 ARREST FT A EE RR 14,919 4,335 6,050 4,534 DAY TIONIIE . « ¢ «sometimes wn wn ww x 0 TEASER 1 B&R ror 1,000 r 170 ] 660 170 ] 12-35 TAOS . . . « voremm tata nnd 5 5 vv vx 50 1 #4 sein tenis EER Y RE 1,278 394 490 394 F-BIYOOIE inn 700555 omnes ey § S55 PLP 28% 8 Burtielmiaierns dg abs E283 1,610 560 490 560 BoA YTOID 505 2 8 5 5 5 = inmates so 0 no 6 wx 45% E83 BETRTEITRALE Eo 0 sv ws vn 1,610 560 490 560 VI AOYBRIE os «= 5 5% 10 oh FEET sr wo 0 2 mk nn 5 3 0 0» 2ST BRD 8 43 1,610 560 490 560 SOZOYBAIG .. \ «vviviv inion nde m as 6555 BE 05 5 545 50S SURREAL YE TAF 4 1,580 511 1 490 579 ] BO-3DYOAIS vo 50 vs bois SEERA 0 2 SHS TEE BF BRE saRel ye 31 v0 1,529 460 490 579 QOBOYBAIT +. 055 bh rrr minimis mninis om x a 5 0 2mm i rea wv 0 1,193 [ 350 490 353 ] BO~BOYOAIS: rvs. 0 toni ibe a i IH bamennscsts vv 3 518 Ai ne ARM SRT TF 920 210 490 220 BOBO YEAS: cn v0.0 0 join minis mime ima on SAR RTRIETER 4% Wo 0 0 00 lr rs Sm ae 1,137 300 490 347 TO-TDYOOIS rain ss v's 5 aicss ras wis 558 SERVES RRR W) E008 RRR 9 wei wt 827 [ 186 ] 490 | 151 ] BOYBAIE ANDOVEE vv uys 21% 5 Lt fmm mm wisi sowie sanssimome tA RR S30 30 RAR bo 625 74 490 61 Female ALBOBS 155155550 0s mmmmmmnmmm a en «8 xk 85 2% &REEBRET GY € E x 15,175 4,636 6,073 4,466 OEY ITONING + 555s «0 wm mens Bu 5 6% 6 95% 5% 35 3 5 3 BEIEPEIA IE HS AY eas ah 995 141 683 171 12-3BMONINS ..ccvvnnmimnmmrmnns rc rrr erst rsnns amr ret rere 1,278 395 490 L 593 ] DBYOAIS 0 von ns mins in ib HEHIEHR A 2 45 56 0% #2. b 4 Saisie EE BEER TY 1,610 560 490 560 B=TTYOAIS i: oisvsas pate hivereres Cai uEs F586 RUPE TENnBEs ers 234 bes 1,610 560 490 560 VDHOYOBIT 50 v2 12 5% vw ovum arms sm wn 31 3 0 Afi vc ERT rman) 7 50 1,610 560 490 560 DO=2TNBAIE: cine 153 B58 Fh itm amens i 31 1 5 4% so ERRATA 7 8) 3 1,628 [ 599 490 539 BO=SOYBEIS. ormvon mins iw imi bimla sete RTEIIETE HA 4, 040 WS 0 0 Be me Hn 1,614 594 490 530 ] BOBO YORE ivncinin on oh BIER EA SBR AEELAR 9 RD Bin wi 1 0 Ew ent 1,246 [29 490 360 ] BOBO YORE: -oscomumistn as sidision shasta on wisi noe sess mwsms minis sim sit wt tt 0 om manmens 994 271 490 233 BO-BIYBAIE. .. commision dR 2 ris 4 4 st sata dessa RAR ee AAA iF A 0 6 WE AOE 1,082 279 490 313 TO-TOYBAIE : : : : svi #5 5957 EFF 8H BREE SR RESC ET EERE s 1% EW 845 | 189 ] 490 | 166 BOYORISANI OVE iivisimiinsssminins #4 4 ¥%1% 5 4% ts TarREEmEREL io awww % % io 0 0 663 92 490 81 NOTE: Consolidated age classes for analyses are indicated by C 7] . Sample sizes assume 560 examined persons per consolidated class, and 1,000 examined persons in the 2-11 months age class. For white and all other persons, the minimum sample size is 490 in order to satisfy the total sample size requirement. "Mexican-Americans can be any race. 27 Table 4. Number of sample persons, by race-ethnicity, sex, and age: Third National Health and Nutrition Examination Survey, 1988-94 Number of sample persons White Mexican- Sex and age Total Black and all other American’ BOIVSEXEE! :.:56.0:05 0h mmvimsesntons + wm nms girs dh bHEGIREIRETE 5 #54 voids 40,561 11,882 16,781 11,898 Male ARAGBE .. cons rmnmmmmmieesssssszrs orn woe ERE R DDR TFET 88 8 5 me 20,041 5,612 8,236 6,193 D1 TBOIHNG «oc mmrmdiinrst am 6 ov v5 wv ox wx vo Wimmera 5 6 BE « & 3 § 85 ® w Gm 1,231 198 ] 835 [ 198 ] $2-BE MODINE somemermmnnt cnn nen rons nmi a EEA AE E5242 250» ns nom 1,499 443 598 458 B-BYOAIS ; sommnmmudtebna ds sas ished Anas nlieenanes saa fas tas sss sane 1,900 651 598 651 BTV YBBIS cin ivinivnina das 54 £3 5 3 213 3% Harman smite 4 4 £3 2% 3 § QBRTuES 1,925 667 583 675 F2-TOVBAIG Jv sioiviitiommnis 05 wrbin mwas oma sem zngarsimmsnm ma dn 5 45 G58 3 $58 Ssmamerne 1,991 675 598 718 DO-2G YBAU tssrsrarsiaannibinss o 0: 91mw Hid chs shires SAA A AAA RE 5% 0 4 Fear sammsneit an 2,149 [ 623 662 [ 864 BO-BOYORIB' uo omivivinmins alsin ve AEE 6 ERIE RE TE HOH SEER Shanta rtd 2,193 657 671 865 QO-ADYOAIS . «ovum cna ovmm os soe sms sven wnsin sees sa eet ssse buss 1,861 [ 603 731 [2 BOSBOYBAIS 5.5.2.4505 255.05 fois wontseminsons Commarea Bll oo ETRE + over 1 1,389 309 731 349 BO-BOYORIE 5 us svn vim itn denied 58 3 D5 AES SERS a ATT eerie asks 2 4 1,707 435 721 551 TO-TOYRBIT + ov rusia sms asa ERES oi 5 HT 3 55 DOS ili bs 2 50 20 ; 1,245 [= ] 754 E ] BOYyearS and Over .............c.vivinunnnnnnunnnnnnnsnasvnnnsnnnnnnn 951 100 754 97 Female BUBGBE ... x ciumimmmia ss 5 a8 7 E588 %3 8 #5 SARA Eh SESE IEEE 4S 20,520 6,270 8,545 5,705 Z=1T MOMNS .vnmmmuss e230 25% 5 5 4 FR smusomEsRva E525 E25 gE 20405 4 1,243 [ 191 843 Fo» V2-35 MONING' uivivinins sinnvivs vs out» x vb wrvsionommmimns msm a & 5 x §% & % 5% #400 1,551 459 613 479 B-BYOHIS my vamammmanminsans Sus & hon wessneion WEAR SHA to 5 4 bah 8 # ane 1,955 667 605 683 B11 YBAIS i oiviniviscnininhisie 5.8 2.6 $5 BAENTV RIG BEARS HAEEIE E208 3 ¥ & Subrtntch 1,932 683 598 651 YDNOYBEIS: «v.cinon shiv srwmsmd von $515 woes amos sop awaits 44 Sa 0 3 9 BRIERE 2,012 683 620 709 POHDGYEIE 5 2.50 5.55 asin cogs cw wremosemsivmens. ie mists do bE 8 le § Teese sens, 2,127 [ 799 645 [ 683 BOIO YORE 1160 20mm snin min oi Seti fie 6 AAA RB BR 04 SAR Hin et 4 2.177 836 671 670 AOD YOAIS ou vivevn mnie Sunaivalemr ss Hl 3 5 05 PE a9 sem eins da esg 4 81 1,757 591 ] 710 [2s BOBO YEarS .........iuitiri iii aaa aan 1,495 411 742 342 BOBOIYOBIE 15 i's vs vind Frm masn ns mrs & nd bBo Hoa SIE SATE 5 + 0s 1,703 465 778 460 TO=7OYRAIS . 1 : i» uimwomnvavins oti TEES 33 5 DRT A CERES ars 58 ands 1,430 [sz ] 860 E ] BO YEArS ANA OVET «ooo i tite eee eee es 1,138 159 860 119 NOTE: Consolidated age classes for analyses are indicated by C ] . Sample sizes assume 560 examined persons per consolidated class, and 1,000 examined persons in the 2-11 months age class. For white and all other persons, the minimum sample size is 490 in order to satisfy the total sample size requirement. 'Mexican-Americans can be any race. 28 Table 5. Number of households to be screened without stratification to meet minimum sample size requirement per consolidated class after allowance for nonresponse, by race-ethnicity, sex, and age: Third National Health and Nutrition Examination Survey, 1988-94 Number of households White Mexican- Sex and age Black and all other American’ Male DAT NTTOIVIG. 2.0.0.0 all on garmsnanmiens srsoec Si BP BR ec ERR A hr ns 67,500 63,536 [ima V2-BBOAMAG -o00.5 595.3 5 6 5 5 baci fn RE RA 6 CY SERRE BE RE [ 63,881 18,800 146,834 B-BYORIS 5 vioivs vos 5055 554 5 FREHERD IES LEER EYE §EE EDR VEY 63,648 12,563 144,710 BY TYROS! 0mm #5 15010 0 sms es A BA 8 4 7 00 0 wc SHEE AT 35,652 6,400 86,145 VDENOYBIIBD! vsti i 2 v1 immo eis opeime Eder rd & BBE 8 2 # i eeomamaieniionsio 32,821 5,320 79,402 DO DOVOAIS: viv vst 45785 035 22 5k 6 simi ras #65 1 1.05 5 08 1 # & % Bbw [ 27,706 4,208 [ 79,258 BOBO YORE: viv ues vs ass ns Edad hE SIRE 4 5% ESTES eS SEE 32,456 3,830 79,258 HOADYBOIS cocniers 55 5m mm v0 ww 0 00 n in 8 BTATEA € 2 % 9 % % % # £4 6% % 0 1s wi [ 43,441 5,398 [ 79,258 BOBO YOBIS: wins v2 5 0 nn es ennesisnmsmmhnmmrasts EEF ER 15008 buns LP 37,075 7,888 84,290 BO-BOYOAIS .vivmvns 1 £55 55472553 sd Pane One erns 7 ¥ 7428 £4 54s § Buhvae 60,532 8,463 194,468 TOTO YBAIS iicivivivursimas 515 9% 753 5 F 5555505500000 ki Bs 33057 % 05% % 5 ow a0 | 56,257 ] 14,488 [ 94,695 ] BOYOAIS QI OVBI oc nummnmsinuson ww mm rie 3 Bie he sss: 4 0 nse 56,763 39,839 194,118 Female DTV ITIONMNNG olson o0560 0 58 2% 1 ET 000 BABES 6 51 FE HH [275 67,500 [2% V2eBE IONE. ovo v0 sw + 0 0 0 50 amis TERS 0 im 1, i SR 67,941 20,251 167,881 B=BIYBHIS: mmm rn pcs © mrt oie in 4 1 TRA At 8 BET fe wt svsaamoa 66,686 13,362 165,851 B-ITYOAIS ivcrton sina s HERES 640 8810 Rieiersammiivdn ds v5 5 8 #8 7 REE 2 3 5 Rhu 37,240 6,882 90,497 2-ADYOMIS soimmin 5552 555520325 5284 WAST iiaians nua vs eye esse wes 33,726 5,788 84,764 PODOYBAIG cirommmmnin vivo unin s 5525553 § § BE ERREERRSRA © 1 £ 1% £55 % 2 0 1 3 STS [ 30,292 4,119 [ 67,219 BOBO YBAIE 5.6 temlikhisesi vom os 1 1 1 wm mm 200 wie Brin 3 Bugs RARE A 8 5B 5 ot 4 0 20 0 smn 31,999 3,798 67,219 40-4D.YORIS . ..cinwvisriniind 5 R57 F325 533 2b RAGOEVIRRTIEY 5 5 8 885455 ewe [ 33,909 5,032 [ 67,219 BOBO YOAIS cio iwimammsrsmaivetn ¥ 4 5 3 0.55% 85°38 eines w/b eam smn v5 45 4 03% HEB 34,423 7,502 78,092 BOBO YOBNS | 1:10pm 30 0 ris 308, rte afc sas meg sed ge es 69 ch 43,800 7,799 145,862 VO-TOYBBEG' 502 ens wanmimemormaas ami rch hci os ro ERR msn 5 ar vn [5200 ] 11,285 [rissa ] BOYORITANIOVBY.. - x 5c 1s waiiriti ins & an min aie Semis Ew 4m i 3 POA © 0 wi 45,292 22,728 143,899 NOTE: Consolidated age classes for analysis are indicated by [ ] . "Mexican-Americans can be any race. Table 6. Projections of the 1990 U.S. Mexican-American population by stratum number, according to the percent in block group or enumeration district, number in thousands, and percent distribution: Third National Health and Nutrition Examination Survey, 1988-94 Percent of Mexican-Americans Stratum number in BG/ED " stratum Mexican-American population in thousands Percent distribution of Mexican-Americans less than 1 1-29 3-4.9 5-7.4 7.5-9.9 10-14.9 15-19.9 20-24.9 25-49.9 50+ 11,603 580 1,044 696 696 696 928 696 696 1,857 3,714 100 DOOD OW, CO) = nN "BG/ED is block group or enumeration district. SOURCE: Tabulation of 1980 Master Area Reference File. The counts of Mexican-Americans are derived from the Hispanic counts for the entire United States excluding New York, New Jersey, Connecticut, and Florida. 29 Table 7. Number of sample persons for self-weighting sample, level of screening, proportion of subgroup in low-density stratum, expected sample sizes, desired variances, and required variances in high-density stratum, by selected subdomains: Third National Health and Nutrition Examination Survey, 1988-94 Number of Expected number of sample persons sample persons in that meet Screening Proportion low-density strata Required precision target necessary of subgroup from basic sample 2 Desired variance of with self- to meet population in low- (adjusted for variance estimate in high- Subdomain weighting sample precision target density strata’ nonresponse) of estimate density strata’ Mexican-American males, ages 80 yearsandover ............ 97 194,000 0.14 4 0.00093 0.00067 Mexican-American males, ages2-11months ................ 198 150,000 0.14 11 0.00045 0.00039 ' Low-density and high-density strata are referred to as subset “a” and “b,” respectively, in the section of the text “Optimizing sample size among strata.” For this table, low density for Mexican- American has been defined as strata 1 and 2 in table 6, and high density is comprised of strata 3-10. 2 Basic sample is assumed to be based on a screening sample of 67,500 households. Table 8. Sample sizes and screening levels for self-weighting and stratified samples, by selected subdomains: Third National Health and Nutrition Examination Survey, 1988-94 Level of Screening (Self-weighting sample) Level of Screening (Stratified sample with optimum allocation) Number of Number of identified sample identified sample persons to meet persons to meet precision target precision target (adjusted for Basic (adjusted for Basic Subdomain nonresponse) Total sample ' Supplement nonresponse) Total sample’ Supplement Mexican-American males ages 80yearsandover ............. 141 194,000 67,500 126,500 141 106,000 67,500 38,500 Mexican-American males ages2-11months ................. 228 150,000 67,500 82,500 228 82,000 67,500 14,500 ' Basic sample is assumed to be based on a screening sample of 67,500 households. Table 9. Values of A,' used in calculating measures of size, by race-ethnicity and density stratum: Third National Health and Nutrition Examination Survey, 1988-94 Density stratum (=i) White and all other (k=1) Black (k=2) Mexican-American ? (k=3) YP mcmas sb & 3 0R STR GR SEAR BRE.E § 4 88 1 5 aetna A ER 5 8 0.119 0.583 0.844 ET I 0.119 0.583 1.177 Bits vie © Soar mam se ns LER REBAR wR NR HR AT 8 0.119 0.583 1.418 Bh tonto anmmtest nn ws s B R G 6 ATRaRTRRr w) E mmm B 0.119 0.583 1.502 B! uit tvs tit 5 BE 8D RRA SR EG Des RR ERA PRE Be 0.119 0.583 1.539 B coi 1% Rn Be RR RISA B88 8S ee 0.119 0.583 1.576 c d NOTE: 4, =r, 2 "Numbers are the numerators of A, ; denominators are 930. See section of text entitled “Measure of size of segments” for a description of how A, is used in the calculation of the segment measure of size. 2Mexican-Americans can be any race. 30 ‘paje|dwod "a0el AUB 8q UBD SUBOLISWY-UBDIXS, ‘sieak 6G—9 OJewWwa} 10 Sew S| 9} *JOAO pu sleak 09 ajews) ‘sieak 9 Jepun sjew si G| 'sieak g Japun ajewsy si y| "JOA PUB S1eak 0g ejew si gl 's1eak 6G o|ews) ‘sieak ge—z| ojew si ZL ‘sia 6G—0p Bjew ‘sieak | |—Q sje) 10 jew si || *JOAO pUE S1eak 0g ajewsy si 0} "JOAO PUB S1eak 09 ‘G—| ojew si 6 's1eak g Jepun a[ews} ‘1eak | Jopun sew si 8 'sieak 6E—0z Jews} Jo sew si 2 'sieak 6p—0p ‘Sieak §1—9 sjewas) Jo djew S| 9 ‘sieak 69—0G o|ews) 10 dew Si G 's1eah g/—0/ ‘sieak G—¢ e[ewsy 10 slew si “JOAO pUE SIE 0g ojews) ‘sieak g—| 8jews) Jo Sew S| § *19A0 PUB SIeaA 0g 8jew si 2 "Jed | Jopun jews) 10 jew si | :suonjulep urewoq, SINVHN Usym apew aq ||Im a)is AeAIns yoes uj pasn sales ay} jo Arewwns vy "sajes Bujidwes ay} uj spew uay} a1em sabuey) ‘siaquinu palisep ay} Wwolj ajeinep Alqeqoid pinom surewopqns awos ui sezis ajdwes ay} Jey} palealpul Ajjeaipoued Ino pales pial ajduwes ay) JO MaIABY "PaLElS ||| SINVHN UBUM pasn sale. ay) aie 8say| "eAIasel Juadiad-0G au} Jo} SMOJ[E UOIUM ‘OEE Je Siojeulwouap ay) ‘sales Bulidwes ay) Jo siojesewnu ay} 818 UMOYS siequinu ay] :310N 19G°0¥ G95°09 910'9€2 0ce0 0ceo ovo ocho 0900 orto ~~ wnjeJ}s ul sueduswy -ueoixa Jo uoiodoid 60S‘, 18Y'6 1.0'6 6.,°0 000°} 000° 000° 000°} 000} 0080 "9 0€L'2 6.9 GLE} 8110 0082 008°C 008¢ 008¢ 008} 000°} 1 WEL ¥8G'C WLLL 9900 00,°€ 00L'€ 00.'€ 000°€ 008} 000°} Tvl 888 vet oy LE0°0 0009 000'S 000'v 000°€ 008} 000°) "Torreon el 868°L 1 169°Lt €59°L 1 et - = . v= = reer |ejoL ,UBOLIBWY-UBDIXS)\ G/2'S 9/8, 685°GL €65°0 0v'0 0L¥'0 0L¥'0 0L¥'0 0.¥'0 [074 200 cl 2922 00€‘e ¥8S‘G 8610 0650 0850 0SS0 0850 0SS0 0ss'0 rors LE 056 ryt 800°C 1200 0490 090 0490 0290 0490 090 rrr ol 088‘L 18.2 188° colo 0060 0060 0060 0060 0060 0060 Crore 6 SIS glee 8LL'e SL00 000°} 000° + 000° 000k 000° 000k Terror 8 288‘LL CLL} 081'82 e660 =O Tye ke me re i re [elo yoe|g 6¥9°C £86'C 6v.29 0ce0 6500 6G0°0 6500 6500 6500 6600 Crores Z ov8‘e 2y8's L1S'€9 ace 9800 9800 9800 9800 980°0 9800 rrr 9 262 LIEV 08¥'GE 2810 vLLEO vHL0 viLO 141% vO PEED rrrrrrrrrereccaiehetiel S L182 8ve'y 628'02 9010 0640 0640 0610 0610 0610 o6L'o rrr v 1202 LoL'e 8016 6¥0°0 00€°0 00€0 00€°0 00€0 00€°0 [07013 c ¥SL CL} €8L°} 6000 0650 0650 0650 0650 0650 08g "rrr c 8/91 LESS Live 2¢io0 000} 000°} 000} 000° 000° L000 I 18°} 202'se £81961 oo0’v TTT TTT TT ren nnn [elo 18yo ||e pue alum 82s 9jdwes aniesel (spuesnoy) [eio1 jo (a1ow 10 05) 9 (6-02) § (61-01) ¥ (6-6) ¢ (rele (€>)1 \urewop pu Ajjoluyjs-soey B10} yim azis ajewijse uojuodoid pajoadxy g|dwes Ill SINVHN (wnjess ul uedLBLWY-UBIIXBY JuSI8d) WNBAS AJISUSP UBILIBWY-UBJIXON uoyeindod 0661 $6—8861 ‘A9AINS uoneUIWEXT UOHIINN PUB YleaH [euoneN piiy] :sulewop abe-xas pue ‘Ajoiuyle-adel ‘wnjesls Alisuap ueouawy-uedixapy Aq sazis ajdwes pajoadxa pue sajes buydwes "gL ajqeL 31 "80BJ AUB BQ UBD SUBDLIBWY-UBDIXS|, ‘s1eak 6G—9 jews) Jo afew si 9} "JOAO pue sieak (9 jews) ‘sieak g Jepun afew si G| *s1eak 9 Japun ajewsy si | *18A0 pue sieak 0g slew si g| ‘sieak 6G—z | olews) ‘sieak ge—z| dew si Z| ‘sieak 6G—0p ojew ‘sieak | |—Q ojeway Jo Sew si || *18A0 pue sieak 09 8jewsy si 0} *JOA0 pue sieak 09 ‘G-| djew si ‘sieak 9 Japun ajewsa) ‘Jeak | Jepun afew si g ‘sieak 6E—0Z o|ewa) Jo Sew si / ‘s1eak 6p—0v ‘sieak 61-9 ajewsay Jo dew S| 9 ‘sieak 69—0G o|ews) Jo dew SI G ‘sieak 6/—0/ ‘sieak G—¢ a|ewsy 10 dew Si "JOO PUB S1eak 0g B|ews) ‘sieak g—| ejews) Jo dew S| § *JOAO pUE S1eak Og ajew si g "Jeak | Jopun ajewsy 10 dew Si | ‘suoniulep urewoq, 0009 000°'S 000% 000°€ 008+ 000'L rnin | Wnjess 0} dAje|a) ales wnwixew 0.10 0020 0520 0€E0 0950 008'0 ~~ ooorrsmreseeseeeees gl 0L¥°0 095°0 0040 0€6°0 000+ OG © rRiEeemeseemedeeseghelisth seis Gl 0290 0vL0 0€6°0 000° + 000+ 000°) | rrerrrerseesBbiveisie sess ce sees vi 000° + 000} 000'L 000} 000+ O00) | erEeeehemaReesieesles rise die einsh el ZUBOLIBWY/-UBDIXO\ 0800 0600 0zL'0 0910 0920 Qi = EEEESReReae.esaopecednn hme zl 0600 0LL0 ovo 08L°0 0L€0 [ole1= + I Lb 0LLO 0EL0 0LL0 0220 0.€0 OLG0 2 UrveTesdsssssexvestsian esha ot 0S+0 0810 0€2°0 00€°0 005°0 OOGIGT © RPrEesremesliee ho Re 6 0L+0 0020 0520 0€E0 0950 OE © Trermremesscscsd bilities deen 8 Yoelg 0100 2100 GL00 0200 €€0°0 [SIT 0 HA Z 7100 LI00 2200 620°0 2100 GAOD ~~ vereeseeensen sas snssistssnpen 9 6100 €20°0 6200 8€0°0 €90°0 BELG rehreeemeesmsscsiiszicesssiamesias S 0€0°0 000 0500 0900 0LL0 BIO verereesdisteisiesescieacnsennaaii v 0500 0900 0800 0040 010 vr SE i € 0010 0zL'0 0510 002°0 0€€°0 08810 | Tereiemeessessrses heehee 2 0L+0 0020 0520 0€€0 095°0 (0100 J EAR L Jaylo ||e pue slum (810w 10 05) 9 (66702) § (6'61-01) ¥ (66-5) ¢ (6v-¢€)e (e>) (ssuep jou) | \urewop pue Ajoluyje-eoey (wes uj ueoLBLY-UBIIXSYY JUBI8d) WMHS AlISusp UBILBWY-UBJIXa) $6-8861 ‘A9AINg uonBUIWEXT UOIIIINN PUB Y}|eaH [RUOHEN PJIY] :Sutewop abe-xas pue ‘Ajoluyle-aoe. ‘wnjells Alisuap uesuswy-uedixapy Aq ‘ajdwes pausalos uiyum sajes Buijdwesqgns urewoq || a|qel A [se] Table 12. Values of parameters used to estimate A, for primary sampling unit measure of size, by race-ethnicity domain: Third National Health and Nutrition Examination Survey, 1988-94 Mexican-American y C. C IC . density stratum (i) P, White and all other J LL LT LL TT rr DE Spe: 0.012 Zen BE TE REE sty © SERIES § 5 BEE 00 8 neh GER 0.009 BD © hf etme #2 © 7 5 3 F585 § BYE Gmina « % 5 05 2 58 £ 6 55 8 phmmmueammein 0.049 B50 RAR #2 5 EW 5% 5 BES SHOEI 2 4 5% RF 8 30 & 55 RAMEE 0.106 TD TTI IIIT m—— 0.182 6 «2 bomnrdnn ribet 530 Bers 5 guises ene Ratio 53 5a 4 oe 0.322 Time Era EER A ER LA thon Marr mremeThna ed 4 4, 2S RT sen 0.320 Black 8) cons mrmnat B EEE EE eR RE EE 2 TRE D6 § Sree ve 0.075 Brrr nnn #8 EE bh SER mame + wg SH 59S 5 2 & BOER nal 0.102 HO + ov mem 5 Ea ETE FR SR fa 45 20% 4 % 8 EA ESR 0.071 HE Gs mm ER EB 5 Y 6 ses RRERTE A 5 8 59 E106 #8 * Rb SURE 0.198 $2 30 cn emm—E 5 BE BER § 8 BEB wwsiemmpri & #5 28 £8 8 54 BB 8 sw mmmmmmne 0.553 Mexican-American’ UB) riven rans ST Fira i RR I RT aaa So 8 ww 0.037 1 0.140 0.118 WD oni ress soromimmmstonsiesn Coit To SEE 10 oni coco seh STRAT nto rion 0.066 2 0.060 0.071 1. I PT 0.118 3 0.120 0.170 1 | J eT TTL rT TT TT mm 0.276 4 0.140 0.210 7 on om m SBE Ef 5 TE EE ARTES 35 EE 5 4 Smee 0.340 5 0.220 0.338 IB 24 or x wonmamsmmmihn 8 2 £5 5.5 2 § § § Sonmsnmmmmerneien dt 45 2 2 7 3 208 & 8 § £ abstr 0.163 6 0.320 0.504 White and all other: A, = 0.119/930 Black: A, = 0.583/930 Mexican-American: A, = 1.412/930 Cc A] Where A = S Pel Zz. ok 'Mexican-Americans can be any race. NOTE: Mexican-Americans are listed as having six age-sex domains although other tables show only four domains. At the time the PSU's were selected it was expected that six domains would be used, and the parameters for the PSU selections were selected on that basis. The number of domains were collapsed to four prior to selection of households and persons. ‘90.1 AUB 8q UBD SUBOLIBLIY-UBDIXSIN, - Jano pue sieak 09 ale €8€°0 0920 G00 2 - = E33EEEssrasteaeRmsuiEEaeee el 18A0 puke sieak 09 slew ‘sieak G Jepun sjewa 0SL°0 0810 6220 7900 “ wo ERR came RR — 19AO pue s1eak 09 sew ‘g Japun s[ews) Jo S[e 00€°0 09€°0 0St°0 009°0 y°0 0020 rere Si—¢h II L910 0020 0520 €€E0 955°0 0080 Corie 8lI—¢l 000+ 000+ 000+ 000+ 000+ GORE tps RR EE Ti ne [elo Jueduswy-uedixap suosied a|duwes oN £€8°0 008°0 05.0 £990 laad) mt en 8 Be EE ae BUON sieak 9 Japun ajewsy ‘Jeak | Jepun sey L100 0200 6200 €€0°0 9500 OO¥iQ FR rReremabuibismeiisi ia fg ar 8 18A0 pue sieaf 09 sew ‘sieak 9 Japun jews) 10 sje 8€0°0 900 8500 L100 8210 QEZp Te sEREmsEetessssassnsies 6-8 18A0 pue sieak 09 jews) Jo slew ‘sieak g spun sews) Jo S[e| 0200 ¥20°0 0€0°0 0v0°0 990°0 (7A I 01-8 18A0 pue sieak 09 8|ewsa) ‘18A0 pue sieak Op ajew ‘sieak g| Japun ajews) 10 sje 7100 9100 0200 £200 Sv0°0 O80) Ctcrrtisakeskekssanmse 11-8 \ 8.00 ¥60°0 8110 L510 192°0 (074720 IE 2-8 000+ 000+ 000+ 000+ 000+ GOOF] STs smisSEes «se [elo] Yoelg suosied ejdwes oN £680 0080 1520 1990 v0 ws eR See ane auoN sieak g Jopun jews) 10 sje 690°0 2800 2010 9€L'0 822°0 (01520 I b 18A0 pue sieak 0g ejew ‘Z Jopun ojews) Jo 8[epy 8v0°'0 850°0 £00 1600 1910 QBZ0 “CC rirUrrerereressesesenes 7-1 18A0 pue sieek 0g ‘py Jopun sews) Jo sje 8100 2200 1200 1€0°0 190°0 OID PFE tRsissknssamRsnaviuei e—1 Jano pue siesk (7 ‘9 Jepun ojews) J0 [ep £100 S100 6100 5200 £70°0 O00 77 irrrvsssesesen®ysmpseves —1 180 pue s1eak 0g ‘9 Japun sews} 10 je S000 9000 L000 6000 S100 B20 cr rechetREEEGETeeseoss G1 Jano pue sieah Op ‘0g Jepun jews) Jo ae $000 S000 1000 6000 5100 LEO CT rrremsessessessenessis 9-1 I 0400 2100 GL0'0 0200 €€0°0 6G0°0 Creer i 000°} 000+ 000+ 000°} 000+ 1p RE LL [elo] 18410 |[e pue 8lyM uoniuyep uewop aAleINWNY (stow 10 05) 9 (664-02) S (661-01) ¥ (66-5) ¢ (6v-¢€)e (e>) (asuap jou) | \urewop pue Aoluyje-eoey (wes ul uBoLBLLY-URIIXSYY JUSSI) WBNS AIISUBP UBILIBWY-UBDIXS) v6—8861 ‘AeAing uopeujwexy uonuINN pue yjjeaH |euonieN pay :sutewop Ajoluyie-aoe. pue wnjess Aisuap Aq ‘ejdwes ay} ui apnjoul 0} Jaquiaw pjoyasnoy yaiym Buiquosap |aqe| abessaw Bujdwes yoea yum spjoyasnoy jo uopodoid “gl ajqeL 24 Appendix Table I. Target diseases and conditions: Third National Health and Nutrition Examination Survey, 1988-94 Table Il. Home examination components for selected age groups: Third National Health and Nutrition Examination Survey, 1988-94 Allergy Arthritis Cancer Cardiovascular disease Chronic obstructive pulmonary disease Dental health Diabetes 2-11 months 20 years and over Recumbent length Height Weight Weight Mid-arm circumference Mid-arm circumference Triceps skinfold Triceps skinfold Head circumference Food frequency (proxy) Gallbladder disease Hearing Infectious diseases Kidney disease Mental health Osteoporosis Cognitive function’ Physical function’ Venipuncture Interview questions Spirometry '60 years and over. NOTE: Home examinations are not offered to sample persons 1-19 years. Table lll. Examination components as conducted in the mobile examination center for each age group: Third National Health and Nutrition Examination Survey, 1988-94 2 months-5 years 6-19 years 20-39 years 40-59 years 60-74 years 75 years and over Physician exam Venipuncture' Body measurements 24-hour dietary recall Dental exam’ Proxy interview Physician exam Venipuncture Urine specimen Body measurements 24-hour dietary recall Food frequency? Bioelectrical impedance? Spirometry* Dental exam Allergy Audiometry and tympanometry Interview Cognitive test® Physician exam Venipuncture Urine specimen Body measurements 24-hour dietary recall. Bioelectrical impedance Spirometry Dental exam Bone density Gallbladder ultrasound Allergy® Interview Central nervous system test® Physician exam Venipuncture Glucose tolerance test Urine specimen Body measurements 24-hour dietary recall Eye fundus photograph ECG Bioelectrical impedance Spirometry Dental exam Bone density Gallbladder ultrasound Allergy® Interview Central nervous system test® Physician exam Venipuncture Glucose tolerance test Urine specimen Body measurements 24-hour dietary recall Eye fundus photograph ECG Bioelectrical impedance Spirometry Dental exam Bone density Hand-knee x ray Gallbladder ultrasound Physical function Interview Physician exam Venipuncture Urine specimen Body measurements 24-hour dietary recall Eye fundus photograph ECG Bioelectrical impedance Spirometry Dental exam Bone density Hand-knee x ray Physical function Interview '1 year or more. 212-16 years. 312-19 years. “8-19 years. SProcedure is for half-sample only. 66-16 years. 35 Bs rin Primary Sampling Units in the Third National Health and Nutrition Examination Survey: 1988-94 ® CENTERS FOR DISEASE CONTROL © % A “3 Anchorage, AK (Phase 2) not shown Br Q \7 Ca County PSU’s DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service fl Phase 1 ( 988-91) Centers for Disease Control National Center for Health Statistics [] Phase 2 (1 991 -94) Office of Research and Methodology Statistical Technology Staff BE Phases 1 and 2 SOURCE: Vital and Health Statistics, Series 2 No. 113. lute C79 AVY, tes 2 NATIONAL CENTER FOR SNA o JBL Vital and Health Statistics Covariances for Estimated Totals When Comparing Between Years BAe (REM CENTERS FOR DISEASE CONTROL Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested Citation Bean JA, Hoffman KL. Covariances for estimated totals when comparing between years. National Center for Health Statistics. Vital Health Stat 2(114). 1992. Library of Congress Cataloging-in-Publication Data Bean, Judy A. Covariances for estimated totals when comparing between years. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 114) (DHHS publication no (PHS) 92-1388) “Results of a study to measure the degree of correlation between yearly estimates of totals for selected characteristics from the National Hospital Discharge Survey and the National Health Interview Survey." By Judy A. Bean and Keith L. Hoffman. Includes bibliographical references. ISBN 084.6.4548 1. Medical Care — United States — Utilization — Statistics. 2. National Hospital Discharge Survey (U.S.) — Evaluation. 3. National Health Interview Survey (U.S.) — Evaluation. 4. United States — Statistics, Medical. 5. Analysis of covariance. |. Hoffman, Keith L. Il. National Center for Health Statistics (U.S.) Ill. National Hospital Discharge Survey (U.S) IV. National Health Interview Survey (U.S.) V. Title. VI. Series. VII. Series: DHHS publication no (PHS) 92-1388. VIII. Series: DHHS publications ; no. (PHS) 92-1388. [DNLM: 1. Analysis of Variance. 2. Health Status — United States — statistics. 3. Health Surveys — United States. 4. Patient Discharge — United States — statistics. W2 A N148vb no. 114] RA409.U45 no. 114 [RA 407.3] 362.1'0723 s—dc20 [362.1'0973'021] DNLM/DLC for Library of Congress 91-36074 CIP Vital and Health Statistics Covariances for Estimated Totals When Comparing Between Years Series 2: Data Evaluation and Methods Research No. 114 This report presents the results of a study to measure the degree of correlation between yearly estimates of totals for selected characteristics from the National Hospital Discharge Survey and from the National Health Interview Survey. The report also gives an idea of the magnitude of the covariances of estimated totals and the effect of assuming independence when making year-to-year comparisons of totals. ET sp my, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control National Center for Health Statistics Hyattsville, Maryland April 1992 DHHS Publication No. (PHS) 92-1388 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Office of Research and Methodology Monroe G. Sirken, Ph.D., Associate Director Kenneth W. Harris, Special Assistant for Program Coordination and Statistical Standards Lester R. Curtin, Ph.D., Chief, Statistical Methods Staff James T. Massey, Ph.D., Chief, Survey Design Staff Andrew A. White, Ph.D., Chief, Statistical Technology Staff Contents I IOAUCHON 1 wv mrin msm sum sin ms nw rm amb SR AEB BRE SAAR EAE ABA AHEREAN SE Es RE ARE RAR ERGs MRE SAR 1 Description. of NHDS and NHS sonar sermins annie vt insms mimuea nr ips mn sn msm cet sweden bam wens 2 National Hospital Discharge Survey. oo cnimeisaisiwes moms ine mmem soins isle osm meats nm bo ssl s 4040 42 or 00 00 008 00 0 aie 2 National Health JNCTVIEW SUTVEY «ou h uw oo ois 008 500 000 tr 10 005 50 80m 505 60 0d 55 2.50500 600050900 06 90 0 00008 00 3 060 2 STUY TEEIIYw + 05 50 5010.05 15 405 13 52 10% 5 05 9 4 0 4 90 0 00 9 00 00.00 0 90 09 3 09 0 48 00 1 4 0 4 Data. TOM NHDS . .. vvonsprinmsbommsns sus apa @ emi Bema sbi dni dbs bids oni EI EIB Rik Mrmr hR bE S 4 Data Hom NHS: vio vr nium imei 0mm mm mr Basham ni Re £m mse fs Oe 15 fa Ie 0s B30 oa Bon 0 8 0 4 0 4 NN AIANCE COMBO: 30 2403 05 157 SB E3802 B30 5280318 15 5 8 A308 bt Fah 0 033 600 0 5 50,2 508 50 0 55 0 4 0 4 2 03.00 0 0 6 00 4 LOGISHICS OF TIE TIWEBLIGATION is wis ox 010 m5 50 500 i 6 0500 ino 5 07 504 0 0 0 4 0 0 4 5 4 6 00 00 9.00 00 14 5 Verification of the NHDS variants CSUMAIOT « sos nesemat main sam satis same es es sass asians amas esse 7 RESUS + vsnminsntins eminem i mie meni eri RI DIR rE BIR FE HL 0 ER OR FER HIE SIR EET IMF ARR FL BE ERB 0 LE AE 8 NHDS (IO8R08CS rm ns thin sims nimi ami mime iw mms ime smi desea Danese ml 1@ sem sam om sheen 1m soy 8 NEIDS ays Of GAC. vron0 2m ens ro 80a 01080 5 15 5A 02000 © 55850 50% 15 59058 00 30 0% 50 a 5 610 ck 50 0 5 E60 5, 4 8050 7% 37 3 00 50 04.90 i 0 ck 9 NHIDS SULGIOAT PIOCCRUICE «5 10 615 0 5 100 5 5. 51% in 8.5017 00002 50005 5009 98 2 5 09 2 508 3 5 008 0 dn 9 HIS acute and chronic CONAIIONS oie xn mrininrinmpansmntbsmbinsobomsailshnssals imams usuasssenioseh 10 NHIS physician visits and short-stay hospital episodes .............ooiiiiiiiiii ieee 11 NHIS restricted-activity days, bed days, and work-10ss days... iin 12 TIVIDACE OF TNTCTONOC: «50 w swsiu mms £5 505 580 508080 8 wwe 00H E R08 HE #108 408 5 0090800 0 05 0 1 04 00 0 5 id 04 13 COIMPOTIEIIES: OF VATIANGE.. : iv 0 x 650 0 9181003055 8 50% 408 50595 390408 0% 90800 0% 08 8 04009 4 06 0 0 0 0 Ho 04 0 4d 4 4 14 IDISRUSSION. 0.5 5 gir 015 00090000009 007 0 08 000 0 0 5100 00 00 050 0 1 F000 0 05 0 00 4 0 0 4 16 TRCICTCII0E 1 cm 10% sm 000 0 30 0 4 002 53 0 00 tn 00 00 0 00 300 0 B03 0 30 A P20 4 0 SI EB 17 At Of AetailEd BADICE, «vw tin vs mamma mn mons one sins bin 38 050 Boum wd 80 2 400 5 29 8000 5 Hrs 0 2 05 0 2 2 4 58 2 9 RE 19 List of text tables A. Correlation coefficients and effect of independence assumption on variances for number of patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 .............................. 8 B. Correlation coefficients and effect of independence assumption on variances for number of days of care for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 ............. 9 C. Correlation coefficients and effect of independence assumption on variances for number of surgical procedures for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84.......... 10 D. Correlation coeflicients and effect of independence assumption on variances for number of acute and chronic conditions, by selected characteristics: United States, 1982-84 ........... iii 10 E. Correlation coefficients and effect of independence assumption on variances for number of physician visits and short-stay hospital episodes, by selected characteristics: United States, 1982-84 ..................ccvvinnnn. 11 F. Correlation coefficients and effect of independence assumption on variances for number of restricted-activity days, bed days, and work-loss days, by selected characteristics: United States, 1982-84...................... 12 G. Effects of independence assumption on confidence intervals using statistics from the National Hospital Discharge Survey and the National Health Interview Survey, by selected characteristics: 1982-84. ............ 13 H. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of patients discharged from short-stay hospitals: United States, 1982-84 ........................... 14 J. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of days of care for patients discharged from short-stay hospitals: United States, 1982-84. ............ 15 K. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of surgical procedures for patients discharged from short-stay hospitals: United States, 1982-84 ...... 15 Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of acute and chronic conditions: United States, 1982-84... ... iii, 15 Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of physician visits and short-stay hospital episodes: United States, 1982-84. ........................ 15 Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of restricted-activity days, bed days, and work-loss days: United States, 1982-84 .................... 15 Covariances for Estimated Totals When Comparing Between Years by Judy A. Bean, Ph.D., Department of Epidemiology and Public Health, University of Miami, and Keith L. Hoffman, M.S., Office of Research and Methodology Introduction The National Center for Health Statistics (NCHS) conducts several large-scale surveys whereby data are obtained through complex designs including sampling in two or three stages, clustering, and stratification. If the design is multistage the first stage consists of drawing a sample of primary sampling units (PSU’s). These PSU’s can be hospitals, counties, or other units, depending upon the specific survey. In several surveys sponsored by NCHS the primary sampling units employed are often sampled on more than one occasion. Two examples are the National Hospital Discharge Survey (NHDS) and the National Health Interview Survey (NHIS). In NHDS, the design used from 1965 to 1987 was a two-stage highly stratified one. The PSU’s selected in the first stage were hospitals, and within the chosen hospitals, discharges were sampled. Although some hospitals have closed, new ones have been added, and a few do not participate, the data collected annually from 1965 to 1987 came from the same hospitals originally selected. NHIS is an annual survey of households located in the 50 States and the District of Columbia. Each year interviewers contact approximately 41,000 new households; however, these households are selected from the same primary sampling units, which are groups of counties and metro- politan areas. From one year to the next the households are chosen from adjacent segments of housing units within PSU's. When sampling inquiries are conducted at successive intervals of time on a continuing basis and covering the same universe, such as in NHDS and NHIS, three dif- ferent types of parameters for totals may be estimated. These estimators are the change in a total from one time period to the next, the average total for all the time periods, and the average total for each time period. The best design is not the same for these estimators and depends upon the correlation of the estimators from one time period to another. This report focuses only on the effect of the correlation on estimated changes between time periods. This procedure of gathering data in a survey from the same basic units from one year to the next means that the estimates produced each year are not independent. This lack of independence is an important feature that, for a variety of reasons, is often ignored by analysts. Statistics produced from NHDS and NHIS are pub- lished annually in reports prepared by NCHS. In the appendix to reports presenting the NHIS statistics, the reader can find relative variances curves from which vari- ances can be calculated. If the reader wishes to compute the variance of a difference between two yearly estimates, a formula is provided. This variance formula is the sum of the variances of the two estimates. In using this formula one is assuming that the covariance between the two estimates is negligible. Similarly, the variances of NHDS estimates are given in reports, but nothing is explicitly stated about the covariances of estimated totals when comparing between years. This report is intended to give NCHS users an idea of the magnitude of the covariances of estimated totals and the effect of assuming independ- ence when making year-to-year comparisons of totals. To determine the degree of correlation between the estimates of totals, an empirical investigation was con- ducted using data from two NCHS multistage, complex surveys, NHDS and NHIS. The main objective of this investigation was to measure the degree of correlation between yearly esti- mates of totals for selected characteristics produced from the National Hospital Discharge Survey and the National Health Interview Survey. To achieve this overall objective the specific aims were to: ® Determine a method for applying the balanced re- peated replication (BRR) technique to the NHDS design. ® Estimate variances and covariances of totals for se- lected variables from NHDS and NHIS using the BRR method. ® Compute the variances of the year-to-year differences for totals using the BRR technique. ® (Calculate the year-to-year correlations between the statistics using the BRR variances and covariances. This report shows the results of the study. Specifically, correlations of estimated totals for selected variables and a measure of the effect of independence are examined. Description of NHDS and NHIS In order to understand the methods and results of this study it is crucial that a person have an understanding of the sample designs for the two surveys. The two designs are briefly described here. National Hospital Discharge Survey This survey is a source of data on the short-stay hospital experience of the civilian noninstitutionalized population. The data are obtained through a probability sample of all discharges, both alive and deceased, from short-stay non-Federal hospitals. Estimates produced from the survey include the number, rate, and average length of stay of patients discharged from short-stay hos- pitals by selected patient and hospital characteristics. The NHDS sample plan for 1965-87 was a two-stage stratified design. Each hospital with 1,000 or more beds in the universe of short-stay hospitals was considered to be a stratum; the other hospitals, called noncertainty hospitals, were stratified by bed size and geographic region. The noncertainty hospitals were classified into 24 size-by- region primary strata. If a stratum contains only one hospital, that hospital came into the sample with a proba- bility of one. Controlled-selection techniques were em- ployed to choose hospitals within each of the 24 size-by- region classes (1). Primarily because of cost, the sample hospitals were allocated to a series of subsamples, re- ferred to as panels, and the subsamples were implemented in the field over time. The second stage of selection consisted of sampling discharges from within each of the hospitals. The daily listing of discharges was the sampling frame. The within- hospital sampling ratio for choosing the discharges varied inversely with the probability of selection of the hospital. In 1988 the sampling plan changed from a two-stage design to a three-stage design. The first stage is geographic area, with hospitals selected from sampled areas in the second stage. Discharges are then sampled from the selected hospitals. A more indepth description is given by Shimizu (2). After the abstracted data are extensively edited, an estimating procedure is used to produce national estimates from the survey. The procedure has three components: Inflation by reciprocals of the probabilities of sample selection, adjustment for nonresponse, and ratio adjustment. National Health Interview Survey This survey is designed to produce national estimates for the civilian noninstitutionalized population residing in the United States. Members of the Armed Forces are not included in the target population. Each year data are collected on personal and demographic characteristics, illnesses, injuries, impairments, chronic conditions, and other health topics. Estimates produced from this survey include morbidity rates and measures of utilization of health services. The sampling design produces a weekly national prob- ability sample of households that are representative of the target population. The weekly sampling permits continu- ous measurement of characteristics and, through consoli- dation of data over time, detailed analysis of rare health- related items. The first stage in the design used from 1973 to 1984 consisted of drawing a sample of 376 primary sampling units (PSU’s) from approximately 1,900 geographically designed PSU’s. (A PSU is either a county or a group of contiguous counties, metropolitan areas, or minor civil divisions.) Then the housing units within each sample PSU were geographically clustered into segments, and each year a cluster of four expected households was selected. Prior to 1985, the annual NHIS sample consisted of approximately 41,000 eligible occupied households, which yielded a sample of about 106,000 individuals (3). For several reasons a redesign of NHIS was imple- mented in 1985. Besides reducing the number of PSU’s from 376 to 198 and increasing the number of interviewed households to 49,000, four other significant changes were made. Instead of using a list sampling frame, an area sampling frame was employed. The number of PSU’s selected in certain strata was changed from one to two. Furthermore, national subsamples of PSU’s were created. The fourth change was to oversample the black population (4). In order to produce estimates the basic data are extensively manipulated. The four steps are: Inflation by the reciprocals of the probability of selection, household nonresponse adjustment, first-stage ratio adjustment, and poststratification by age, sex, and race. Statistical theory indicates that a ratio estimator for a statistic generally has smaller variance than an inflation estimator has. The first-stage ratio adjustment and poststratification are car- ried out for this reason. These adjustments also bring the sample data into close conformity with U.S. Bureau of the Census population totals. Study design Data from NHDS Data collected on sampled discharges of inpatients from short-stay hospitals during the years 1982, 1983, and 1984 were used in the study. The original frame of the survey consisted of the hospitals listed in the 1963 Na- tional Master Facility Inventory. To reflect the universe of short-stay hospitals for a given year, hospitals that have come into existence since 1965 were also sampled. These discharges represent three years of data collection in NHDS, which has been conducted since 1965. Data used in the study were on diagnostic and surgical procedures for 1982, 1983, and 1984. There was one record for each discharge. The variables chosen for this investigation were a subset of the variables NCHS employed for computing variances for NHDS statistics. Data from NHIS As described previously, NHIS uses an annual proba- bility sample of households throughout the 50 States and the District of Columbia. Data collected each year are available on five different types of record formats. In this study, tapes for the years 1982, 1983, and 1984 containing a record per individual or a record per condition were used. The 60 statistics examined in this study were used in generating the relative variance curves formed in NCHS Current Estimates Series 10 reports. Variance estimator The variances of NHDS sample statistics are calcu- lated directly using the estimating equations discussed by Simmons and Schnack (1). In NHIS the balanced re- peated replication (BRR) procedure for general variance estimation is employed. Further discussion of the vari- ances calculated for the survey estimates is presented in an appendix to NCHS Current Estimates Series 10 re- ports, which give NHIS estimates, and in Series 13 reports on NHDS estimates. For ease of computation the BRR method was used to estimate variances and covariances in this study for both data sets. The basic premise of BRR is that the variability of a statistic based on the entire sample data set can be estimated by the variability of that statistic based on subsamples that reproduce the complex design of the total 4 sample. Several investigators — McCarthy (5,6), Kish and Frankel (7,8), Frankel (9), Koch and Lemeshow (10), and Bean (11) —have studied this procedure extensively. The basic technique is described below. Consider a finite population of N primary units classi- fied into L strata, each containing N,, units (h = 1, 2, .. ,, L), with Ls N= SN, h=1 The sample design is a stratified simple random sample with two units chosen from each stratum with replace- ment. An unbiased estimator of the population mean Y is where W, = weight for stratum (h = 1,2,...,L) = N,/N V7, = the sample mean for stratum A The usual estimator of the variance of jy is L V(r) = > Wihsi2 h=1 where 5 is the sample variance for stratum A. One general method for estimating variances is ran- dom groups, but for this situation there are only two independent random groups. An alternative approach is to use half-samples comprised of one unit from each stratum. This results in half-samples that have overlapping units, which means that they are correlated. This estima- tor can be expressed as K Ve (7) = D0: — TUK i =1 where K = 2% L Tr= Sw, Gmiym + 8 puy nm) h=1 1 if unit (h,1) is selected for the ith half-sample, Oni 0 otherwise Buz = 1-8, There are 2- possible half-samples. For this case of a linear statistic, Vz (VV) = V(y). When L is large the computation of 2- half-samples is not feasible. McCarthy (5) derived a method for choosing a subset of the half- samples in a specific fashion so that V5 (7) is reproduced. The set of replicates generated is said to be orthogonally balanced. The subsets of replicates are formed by deleting one unit in each stratum. The set of 4 half-samples is defined by an A xX L matrix with elements a;,, where +1 if unit 1 in the Ath stratum is in the ith half-sample where i = 1, 2, .. ., 4, i= —1 if unit 2 in the Ath stratum is in the ith half- sample. This set of A replicates satisfies the property A > Ai Ai = 0 i=1 forallh (6 -6)24 i=1 Krewski and Rao (12) show that both linear and nonlinear statistics are asymptotically normal and that the BRR variance estimators are consistent under certain conditions. If more than two units are selected from a stratum but n, is an even integer, investigators suggested forming pairs within the stratum randomly (8,13). However, research indicates that as the number of units within a stratum increases, the BRR variance estimator becomes increas- ingly unstable (14). This finding played a role in determin- ing the methodology used in this study. Besides being useful for estimating the variances of ratio and nonlinear estimators, BRR is applicable in the area of inferential statistics. Several studies (15-19) have demonstrated that the BRR approach could be used to estimate covariances as well as variances. Logistics of the investigation The various steps taken in conducting this investiga- tion are as follows. Step 1: Creating a working file For the NHDS data the working file consisted of one record per hospital from the more than 500,000 discharges for the study. A computer program using the statistical package SAS (20) was written to select the variables needed from each record; then records within a hospital were combined. For a diagnosis the output consisted of panel number (denoting the particular subsample the hospital was in), hospital number, bed size, region, and 19 variables. For each surgical tape, the output consisted of 23 surgical variables, along with the panel number, hospi- tal number, bed size, and region. When the statistic was the number of diagnoses or surgical discharges for se- lected patient characteristics, the value for a hospital equaled the sum of the final weights across all the individ- ual records having the specified diagnosis-related group (DRG) code and the selected characteristic. To compute a hospital’s total days of care for patients by selected char- acteristics, the number of days of care was multiplied by the final weight and then accumulated across all the records having the specific DRG code and characteristic. (Discharges for newborn infants were excluded from the diagnostic, days-of-care, and surgical statistics.) A total of 61 statistics were produced for NHDS. This same general procedure was used to prepare the data from NHIS. For this survey the working file consisted of one record per primary sampling unit for a specific year. The statistical package SAS was utilized to select the variables needed from each record and then to combine the records within a PSU. The output consisted of PSU number, demographic variables, and the variables of inter- est. The value of a variable for a PSU equaled the sum of the final weights across all individual records having the selected characteristics and the specific variable. A total of 60 statistics were produced for NHIS. Step 2: Certainty strata Because 17 strata in NHDS consist of only one hospi- tal, these hospitals enter the sample with a probability of one. Each of these hospitals was treated as a pseudostra- tum by randomly dividing the records into two groups. (These hospitals are in panel 0.) Using the random num- ber function available in SAS, the program used to create the hospital records was modified to accomplish this step. NCHS has a procedure for NHIS whereby PSU’s consisting of an entire stratum are allocated to a pseudo- stratum. Accordingly, these pseudostrata were employed. 5 Step 3: Nonresponse problem One program in SAS was written to perform all the remaining tasks, but these various tasks will be explained in steps. In NHDS not all of the sampled noncertainty hospitals in the size-by-region classes responded for all three years—1982, 1983, and 1984. The hospitals not responding for all three years were dropped from the data file used in estimating the variances and covariances. Because the number of hospitals not responding for all three years was small —making up less than 5 percent of the total —this nonresponse is not thought to invalidate the results. NHIS does not have this type of nonresponse, because a household is included in only one year of sampling. Step 4: Noncertainty strata As discussed earlier in the section “Variance estimator,” research indicates that having two units per stratum is the most efficient design in terms of BRR variance estimator stability. However, if the NHDS non- certainty hospitals were simply paired, the number of pseudostrata would be more than 200, which means that an extremely large orthogonal matrix would be required for variance estimation. To avoid the problem of con- structing a large orthogonal matrix, a pseudostratum was defined to be four hospitals. The pseudostrata formed were independent of the sampled hospitals. After the original sample of hospitals was selected, other short-stay hospitals came into existence. The NHDS survey design was altered so that a sample of these noncertainty hospitals was chosen; these sample units are designated as panel B. The original noncertainty hospitals sampled belong to panels 1-6. The hospitals in panels 1-6 were placed into one group and collapsed into strata. However, because weighting is different within panel B, the hospitals within it were collapsed to form strata. Again a function of the statistical package SAS was used to form the strata and pairs within them. The hospitals within panel B were sorted by region, bed size, and hospital. This was done so that hospitals with similar characteristics, especially size, were paired together. Each group of four was considered a stratum. Because there were only three hospitals for the last stratum, the records for the final hospital were duplicated. This procedure was also followed for the hospitals in panels 1-6. However, the last pseudostratum had only two hospitals, which meant that the records for two hospitals were duplicated. Prior research into the behavior of BRR estimators indicates that if the number of units sampled from a stratum is even, pairs within the stratum should be formed randomly (8,13). Because each NHDS pseudostratum con- tained four noncertainty hospitals, random pairs were formed. As stated above, the pairing used by NCHS to pro- duce BRR variances for NHIS statistics was employed in this study. Because there were noncertainty PSU’s, pseu- dostrata were formed. 6 The creation of pseudostrata raises questions con- cerning the bias introduced in the estimates of variances through this procedure. Stanek (21) investigated the effect on variance estimates of pairing strata into pseudostrata. The findings indicate that, as the pairing became hetero- geneous, the bias of the BRR variance estimator in- creased. In addition, subjective pairing was studied using data from the National Health Examination Survey (an- other NCHS survey). The results show that the BRR estimates of variance of mean weights and heights of children were not highly dependent on the arrangement of pseudostrata. For NHDS noncertainty hospitals, the pairing for this investigation was done by combining adjacent geographic- size classes. The pairing utilized in NHIS is subjective. Although no estimates of bias were made, the variance estimates for NHDS estimates should be only slightly overestimated. Step 5: Estimation of variances, correlation coefficients, and effect of independence assumption The next step was to generate an estimate of the statistic utilizing the entire sample. SAS data statements were employed to write a program to construct a 120 Xx 111 orthogonal matrix consisting of 120 half-samples and 111 pseudostrata for NHDS. For NHIS, a 160 x 149 matrix was constructed. The final component was calculation of the half- samples of the statistic. There were 120 half-sample esti- mates in NHDS and 160 in NHIS. This step was accomplished by taking advantage of SAS features. One example is the use of a pointer to read both of the records in a stratum at one time. This process also allows the location in the data file where the program is reading to be maintained. The theory of BRR indicates that data from a half- sample should be subjected to the same estimation proce- dure as data from the total sample are. This would mean, for example, recomputing the first-stage ratio adjustment and poststratification for each half-sample for NHIS data—a task that would require considerable work. Inves- tigations by Simmons and Baird (22) and Kish and Frankel (7,8) found that the adjustment factors based on the parent sample can be applied without seriously biasing the estimate. Other studies (11,23) indicate that these adjustment factors should be calculated for each specific half-sample. For this study, the decision was made to use the factors based on all the survey data. After the half-sample estimates were produced, the variance was calculated. This procedure was repeated for each of the 38 diagnostic and days-of-care statistics and each of the 23 surgical statistics from NHDS and for each of the 60 statistics from NHIS. The basic program was modified to yield covariances between the estimates across years. The BRR formula for the covariances computed is A Cv (Ys, Yas) = > O's — Y's2) ("ss = Y's3)lA where ol A = the number of half-sample estimates produced y's2; = the estimate of Yj, utilizing half of the total sample y"s, = the estimate of Yj, based on the entire sample y gs; and y "gy are defined similarly In addition to the variances and covariances, the variance estimates of the year-to-year differences were directly computed. For example, the estimator for the years 1982 and 1983 is A V,(Ys2 — Ya3) = > [0's — Y's3:) = 00 "s2 = ¥ "s3)]24 i=1 All the output from the programs was passed to a SAS data file so that additional statistics could be computed. Two other statistics were also calculated. ® Correlation coefficients for the estimates — Correlation coefficients were calculated for the estimates of 1982 and 1983 and for the estimates of 1983 and 1984. The formula is: p= CV, (YoY) VV, (Yi) V, (Yy3) ® Effect of independence assumption —To determine the effect that making the assumption of no correlation has on the variance of the difference between two variables, the ratio E=[V,(Y) + VW, (Y, - 1) was computed for all the statistics. Verification of the NHDS variance estimator Table 1 gives the number of patients discharged, standard error, and standard error of the differences for 1982, 1983, and 1984 from NHDS. (The variables in the tables are shown in the order of magnitude of the number of patients with the characteristic discharged in 1982.) Examination of the relative standard errors (RSE’s) for several characteristics indicates that the BRR estimates are reasonable. (The standard error divided by the esti- mate = RSE.) For example, consider these relative standard errors: Females: RSE = (635/21,554) x 100 = 2.9 percent Persons 15-44 years: RSE = (520/14,515) x 100 = 3.6 percent Atherosclerotic heart disease: RSE = (25/465) x 100 = 5.4 percent Malignant neoplasm of lung, male: RSE = (11/188) x 100 = 5.9 percent As the size of the statistic decreases, the RSE increases. Charts providing general relative standard errors for a wide variety of estimates are available in the publications concerning the utilization of short-stay hospitals —for ex- ample, (24). Comparing the RSE’s produced by the BRR method with the values of RSE’s from the figures in the publications shows that the BRR estimates are larger. The ratios of the BRR estimate of RSE divided by the value derived from the figures are: Females—1.16; persons 15-44 years— 1.44; atherosclerotic heart disease —1.35; and malignant neoplasm of lung, male —1.18. Note that the denominators were obtained from the figures by rounding the estimates. Because empirical evidence (8,9,11) shows that in general the BRR method overesti- mates the variance, this is the result one would expect. The standard errors are consistent across the years. Associated with each diagnosis is a days-of-care esti- mate. These values are examined next. The estimates for days of care are presented in table 2. Most of the patterns found in table 1 are seen again in table 2. The BRR relative standard errors are usually larger than the values obtained from the chart of general RSE’s for NHDS estimates shown in NCHS publications. The estimates for the surgical procedures and their standard errors are presented in table 3. The size of the estimates varies from 33,519,000 to 55,000 for 1982 data. The estimates for many variables increase from 1982 to 1983 but decrease in 1984. The RSE’s for four variables in 1982 are: All procedures: Total: RSE = (1,186/33,519) x 100 = 3.5 percent Persons 15-44 years: RSE = (647/14,281) x 100 = 4.5 percent Bilateral occlusion of fallopian tube, 15-44 years: RSE = (54/558) x 100 = 9.7 percent Hysterectomy, 65 years and over: RSE = (4/55) x 100 = 7.9 percent To compare these RSE’s with the ones obtained from the published NCHS charts, the ratio of the BRR RSE divided by the published RSE was calculated. The ratios are: All procedures, total —1.09; all procedures, persons 15-44 years—1.12; bilateral occlusion of fallopian tube, 15-44 years—1.61; and hysterectomy, 65 years and over —0.88. Again the BRR estimates are generally larger, but in some cases the two RSE’s are essentially the same. These comparisons indicate that, in general, the BRR estimate of variance is larger than the estimate produced by the design variance formula. Based on BRR research findings, this overestimation is to be expected. Although no one has compared BRR estimates of covariance with exact covariances, it is reasonable (using the results from variance comparisons) to assume that covariances are also overestimated. However, assuming that the magnitude of the overestimation is the same for variances and covari- ances, the correlation coefficients calculated will be ap- proximately the same as those calculated using the correct design formulas. Results NHDS diagnoses The correlation coefficients in table A for the number of discharges for the 1982 and 1983 statistics range from 0.96 for two statistics—the number of discharges for mental disorders for persons ages 15-44 years and the number of discharges for persons under 15 years of age—to a low of 0.45 for the number of discharges for malignant neoplasm of the lung among males. (The vari- ables in the tables are shown in the order of the magni- tude of the estimates in 1982.) Of the 19 variables, only 7 have values below the value of 0.80 and none has a negative value. The average value of the correlations is 0.80. No clear pattern related to the magnitude of the estimate was observed. If the statistic has an estimate of more than 800,000, the correlation is always larger than 0.88. The other statistics have smaller correlations, with one exception: The variable congenital anomaly for per- sons under 15 years of age, which has the smallest esti- mate, has a correlation of 0.95 for 1982-83. The three malignant neoplasm diagnoses and inguinal hernia for males have the smallest correlations, ranging from 0.45 to 0.58. For the correlations for 1983 and 1984 (table A), the values vary from a high of 0.98 to a low of 0.41, with an average of 0.81. Again the statistics with estimates larger than 800,000 have the higher correlations; they are all bigger than 0.93. The rest of the variables —with the one exception, congenital anomaly for persons under 15 years of age —range from 0.41 to 0.83. The estimated number of discharges for the age group under 15 years has a correla- tion of 0.98, whereas malignant neoplasm of the lung among males has the value of 0.41. Although 11 of the 19 variables have larger correlation coefficients for 1983-84 than for 1982-83, generally the correlations do not differ greatly from one year to the next. Table A also shows the ratios (E = independence effect) showing the effect on the variances when the assumption is made that the correlation between esti- mated totals across years is zero. (The formula for E is presented earlier in the section “Logistics of the investigation.”) For 1982 and 1983 the ratios range from a minimum of 1.8 to a maximum of 22.4. For example, for persons ages 15-44 years with mental disorders, the vari- ance of the difference (assuming the correlation is zero) is 22.4 times larger than the variance taking into consider- ation the correlation. The values for 1983 and 1984 range from 1.7 to 40.4. The mean value is 9.6 for 1982-83 and 13.3 for 1983-84. These statistics indicate that making the assumption that the correlation is zero results in the variance of the difference for diagnoses always being an overestimate. Table A. Correlation coefficients and effect of independence assumption on variances for number of patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 FEMaIR «ov nvm sai os SHEEE TNS 08 REE EE BEE 8.8 20s 0.91 0.96 10.8 28.2 IBAA YBBAIS . . . oot nnvn omnes nn ne ee a 0.92 0.96 12.2 22.4 VIBE... ivvsa: rv vim oon wi 0, conse: 0 0 0 oo cpa gp damn cao 0 0.91 0.95 11.2 19.9 BB YSAIS ANCL OVBE , « uv vv wwia 3 wim a wd a ww 0H 4 oN 0.95 10.2 18.5 BAB YBRIS «0 vv ivi 2 Wie + 08 WE WEE We 0.88 0.95 8.6 18.9 Females with deliveries . , . vu ss ssn smaes svn ona me ve 0.95 0.96 18.7 243 Under15years . . ............c uur nnennnnsnnns 0.96 0.98 19.1 40.4 Mental disorders . . .......... 0.95 0.94 19.0 16.7 Mental disorders, 15-44 YBArS. . . «+ «vv vu cus vn cms snus 0.96 0.93 22.4 14.5 Acute myocardial arcHon. « we sve vw wens wa 0m vas mm ee 0.72 0.75 3.5 4.0 CAlBIAtt, . +s swans sma Ns AE NHS ER ARE RIB HEHE 20 0.89 0.77 8.4 3.7 Atherosclerotic heart disease. . . . ................... 0.79 0.83 4.7 5.1 Inguinal hernia, male. . . ......................... 0.58 0.57 2.4 23 Cataract, BE YBaIS ANA OVO «ov vv viviv wv viv wi min sn win 0.88 0.76 7.3 3.6 Malignant neoplasmiol UNG « «aviv sms sm smn mw ows ww sin 0.49 0.61 1.9 25 Malignant neoplasm of breast, female . . . .............. 0.52 0.49 21 2.0 Malignant neoplasm of lung, male . .................. 0.45 0.41 1.8 1.7 Fracture of neck of femur, 65 years and over. . . . ......... 0.67 0.60 29 25 Congenital anomaly, under 15years . . . . .............. 0.95 0.96 15.5 21.4 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. 8 The price of the assumption is a variance that is at least double the true one. Using the overestimated variance will have an impact on trends of the year-to-year differences and on hypothesis testing done on these statistics. Exami- nation of the ratios reveals that for all statistics with an estimate larger than 800,000, the effect is huge, ranging from 8.6 to 22.4 for 1982-83 and from 16.7 to 40.4 for 1983-84. NHDS days of care The correlation coefficients for NHDS days-of-care statistics are displayed in table B. The total number of days of care among persons under 15 years has the largest correlation, 0.94, for 1982-83 and again, 0.96, for 1983-84. The number of days of care for malignant neoplasm of the lung among males has the lowest correlation for 1982-83, 0.34, and again for 1983-84, 0.29. All the correlations were positive. The ranges of the correlation coefficients are 0.94 to 0.34 for 1982-83 and 0.96 to 0.29 for 1983-84. The means are 0.74 and 0.73, respectively. The pattern by magnitude of the estimate is similar to that seen for number of discharges. For statistics with an estimate larger than 10 million, the correlations are above 0.84. Most of the other statistics have correlations ranging from 0.29 to 0.82. Exceptions are a correlation of 0.85 for cataract procedures for 1982-83 and a correlation of 0.93 for congenital anomaly among children under 15 years for 1982-83. Examination of tables A and B reveals that the correlations are always smaller for days of care than for number of discharges. One possible explanation is the introduction of the concept of diagnosis-related groups (DRGs) in hospitals. DRG’s were put into use during the years 1982 and 1983. The use of DRG’s would affect the number of days of care more than the number of dis- charges for these selected 19 variables. Because the correlations are all positive and most are greater than 0.8, making the assumption of independence results in relatively large overestimates of the variance of the difference. The ratio for days-of-care statistics reaches a high of 26.3 for persons under 15 years for 1983-84 and a low of 1.4 for malignant neoplasm of the lung among males for 1983-84. The value of 26.3 means that when the assumption of independence is made, the variance used is 26 times larger than a more accurate estimate of variance. The averages across the statistics are 6.4 for 1982-83 and 8.4 for 1983-84. In general, as the correlation coefficient approaches one, the effect of the independence assump- tion increases. NHDS surgical procedures The impression obtained from studying the correla- tion coefficients in table C is that these surgical statistics are as highly correlated from year to year as the number of discharge statistics are. For 1982-83, the correlation coef- ficients vary from a high of 0.93 for females ages 15-44 with bilateral occlusion of fallopian tube to a low of 0.43 for females ages 65 and over with a hysterectomy. The correlation coefficients for 1983-84 range from 0.96 for all procedures among females to 0.20 for females ages 65 and over with a hysterectomy. The average values are 0.80 for both time periods. In general, the statistics with larger estimates have larger correlations. The statistics with the three smallest estimates have the smallest correlations, ranging from 0.20 for 1983-84 to 0.55 for 1982-83. Table C gives the effect of the independence assump- tion (E) for the surgical procedures. The ratios range from 1.8 to 14.2 in 1982-83 and from 1.3 to 23.4 in 1983-84. The statistic for hysterectomy among females 65 years and over has the smallest ratio for both sets of values. The average ratio is 6.7 for 1982-83 and 10.1 for 1983-84. The variance assuming the correlation is zero is two times larger than the true variance when the correlation is about 0.43. Table B. Correlation coefficients and effect of independence assumption on variances for number of days of care for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 FOMAIG. ... ; ws vid sins tmumh amis sms da obs a s@s fsa 0.85 0.93 6.6 13.0 MER, «vs malts ws sms ms amid amis AME BEI EY 56 0.88 0.91 8.3 11.0 BE YES ANI OVBE. . « «viv ok Ant ma nisms Sone amen 0.87 0.91 7.7 719 154A YEAS dW. of cmsinie name mur m se Ae be 0.81 0.92 8.2 12.0 BEBAYBBIS . «cv vos imvms va sms omens sain ms swsais oi 0.84 0.89 6.3 8.8 Metal CHBOTOBIS . ... ous vuwisws evigme vusims 62d ux 53 0.92 0.93 12.1 14.4 UNABr 1S YOaIB . «vo vv sms sais ms 65 Ems sh sma sms oh vo 0.94 0.96 12.4 26.3 Females with deliveries . . ...............c.00.0.. 0.90 0.95 9.8 18.7 Mental disorders, 15-44 years. . . . .................. 0.92 0.93 11.3 14.5 Acute myocardial Infarction. . . ... wiv vws wu maemo ve ma a 0.53 0.63 22 2.7 Atherosclerotic heart disease. . . . ... .. «cu vi cws vei on 0.69 0.79 341 4.1 Fracture of neck of femur, 65 years and over. . . . ......... 0.53 0.41 2.1 1.7 Malignant neoplasmoflung ....................... 0.44 0.46 1.8 1.8 Malignant neoplasm of breast, female . . . .............. 0.47 0.32 1.9 1.5 Malignant neoplasm of lung, male . . ................. 0.34 0.29 1.5 1.4 inguinal hernia, Male. .. , vou smsme sms sul wummams vo 0.40 0.32 1.6 1.5 Cataract. «ov wn vwsms cwmsms ams aw ams Hs HEA EE FWD BE 0.85 0.74 6.6 31 Cataract, 65yearsand over . ...............0ouvunnn 0.82 0.70 55 28 Congenital anomaly, under 15years . . . ............... 0.93 0.89 13.1 9.1 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table C. Correlation coefficients and effect of independence assumption on variances for number of surgical procedures for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 All procedures: TOE] cms Ariat 5 Sie 9p rosviy wm: Fin mens Suen 3 w ans 0.89 0.95 8.4 20.6 BEHAAIS: Ue: 1: Fre: in hoes: $5 wir BGR ng sis spas op 1 ie 0.90 0.96 9.2 23.4 VE AAYBAIS. 4 ee va wx iv nim 0% 5 (Wi 0 0 HE 46 EB BE ERY 0.91 0.95 11.6 18.0 MIG, unr sneranpeminm iss se Bi RUE EHS 2H EWE 0.86 0.92 74 12.7 B5 YBaIS ANE OVBE « «si v wi» via: wi 06 % 5h 50% 0 4 0 lel 0 8 0 0.86 0.93 55 14.5 BE BA YBATE.:, 1x10. 2 45: ino © Tr #45) 00. vs 9 2000 wr 00 ris vem So oe wl 0.87 0.93 73 13.6 Under 1B years. . . .ovvvvisnonmsnma wn bdmmmm sms va 0.93 0.94 12.3 15.9 Chroumoision, Under 15Years . «. «sus rwswvems mma ws va 0.85 0.91 6.5 10.5 Cesarean section: BIFBGBE «+ 2 2/5 p03 700 200 0 5 own 78 0 2h meni 590 0 1 0 0 0.91 0.94 8.7 15.2 N58 YOAIGr. 2 § te 2 fo ht 0 9.50 4 50 a wit 19 oe i 0.91 0.94 8.6 15.2 HYSIBIOOIOMNY: ov 5 + 5 thin vi 5 iin mew 3.00 os do wo i ®ve we 0.88 0.92 8.5 13.1 Bilateral occlusion of fallopian tube, 15-44 years . . ........ 0.93 0.93 14.2 13.4 Extraction of lens, 65 yearsandover. . . ............... 0.86 0.79 6.8 4.5 Cardiac catheterization . . .: :vcss sn sms sas ms an a@s £5 0.90 0.92 8.9 10.7 PUOSIBIBCIONMY + «i «on vie min simi wm mies omonnin simamn on 0.68 0.71 3.2 3.4 Tonsillectomy, under 15years. . .........ccuovr news 0.74 0.75 3.8 3.9 Prostatectomy, 65 years and over. . . ................. 0.64 0.67 2.8 3.0 Hysterectomy, 45-84 YOars., . .. « «vv svn cv ws ewmiane vi 0.63 0.65 3.0 27 Direct heart revascularization. . . . .......... coven 0.84 0.85 6.4 6.6 Myringotomy, under 15 years . . ...... «ccs sic ews sn 0.82 0.83 4.6 55 Arthroplasty and replacement of hip: BE YEAS BNAOVEF . ois ro csms mums swore wimama om 0.53 0.47 2.0 1.8 FOMBIB, soo cvs dndims Rulvwy Qs Wa po gwen sav ma xn 0.55 0.43 241 1.7 Hysterectomy, B35 years and Over. . . . +. c. «vs ot ems vn ve 0.43 0.20 1.8 1.3 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. NHIS acute and chronic conditions Table D shows the correlations and the effect of the independence assumptions on the variances for 1982-83 and 1983-84. (The statistics, standard errors, and standard errors of differences are presented in table 4.) The corre- lation coefficients for 1982 and 1983 range from a value of -0.12 for acute conditions among females ages 6-16 years to a value of 0.20 for chronic conditions among black persons ages 35-44 years. The correlations are mostly in the range of the value of zero; of the 20 variables, 6 are negative. The mean is 0.02, which is not different from zero. The correlations for the three statistics with an estimate greater than 100,000 are 0.08 for total number of acute conditions in the population, 0.02 for total number of chronic conditions in the population, and -0.09 for the number of chronic conditions among females. Unlike the Table D. Correlation coefficients and effect of independence assumption on variances for number of acute and chronic conditions, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 Acute conditions Total. vv vp sms mp rms sms@s ao AE IBEW BAH PEWS 0.08 0.12 11 1.1 WHE, 28-84 YORIS., . «i s.o vi sins sais mama smanu EMass -0.11 0.07 0.9 1:1 Female, B~1BYBAIS . ui: «viv nic smsmas nosdio bd mbama sa -0.12 0.09 0.9 1.1 Male, B-1BYBAIS « uv rie vin vimn mim ms fom iin w som wm in wwe ® ine 0.14 -0.15 1.0 0.9 While, 85-84 YBaIS.. : «+ covms resign smems smewn swnan 0.01 -0.07 1.0 0.9 Married, 48-BAYBAIS. .. . + voc: cnr mn ims ss pew Faia ua 0.03 -0.08 1.0 0.9 Female, S85-BAYBAIS. . . . cco sos ns swan sms as smi §0 0.00 0.05 1.0 1.0 TEYeIS ANA OVBY . , cv siisima wiv wi suse ims omens vi 0.08 0.13 14 13 Separated, 17-24 y€arS. . . . . . . oc viii -0.03 0.00 10 1.0 Black, 65-74years. . . . ....... 0.10 -0.01 14 1.0 Chronic conditions Totals os wn v ins ton aia Mn Auth hos S56 hd me 8imy ol 0.02 0.35 1.0 1.5 FBIEID: «5 st voi 2 0 157 2 oot fe 0 0 55 100 5 5 6 0 0 m0 0 0 0 -0.09 0.39 0.9 1.6 25-34 YBAIS . . . . ovr trices ae a -0.10 0.20 0.9 1.2 MaAGH, 25-8 YBAIS. . 4 + wis wis sims mvs 5a wm EE gE 0.08 0.21 1.1 1.3 White, 45-BA YEAS. . « «cox ms seis mas Take A HE AHL ET 20 0.02 -0.14 1.0 0.9 Male, B-=1BYBAIS . « «iv i + wiv nod 33 0 ok 0h 9 hs WE 0 0: 8000 0.12 -0.01 1.1 1.0 Female, 45-54years ............ «uv mnmnnnnnn 0.04 -0.08 1.0 0.9 Income less than $5,000, 17-24 years. . . .............. -0.03 0.01 1.0 1.0 BIACK, 88-8 YBAIB.. . wv + wits sic ma wwe BEE yw aE 0.20 0.11 1.2 1.1 Income less than $5,000, 65-74years. . . .............. 0.06 -0.03 i fs 1.0 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84, 10 statistics for NHDS, the statistics with the largest esti- mates do not have the largest correlations, nor do the statistics with the smallest estimates have the smallest correlations. The values for 1983 and 1984 also are in the vicinity of zero, with the smallest value being —0.15 for acute condi- tions among males 6-16 years; the largest correlation is 0.39 for chronic conditions among females. For 1983-84 the average correlation is 0.06. The correlations for the two years do not show extreme variation with two excep- tions. For example, the correlation for acute conditions for the total population is 0.08 for 1982-83 and is 0.12 for 1983-84. Because the correlation coefficients are all in the neighborhood of the value of zero, the effect of the independence assumption is not great. The ratio measur- ing the independence effect for 1982-83 varies from a low of 0.9 for acute conditions among females ages 6-16 to a high of 1.2 for chronic conditions among black persons ages 35-44. For 1983-84 the ratio goes from a high of 1.6 for chronic conditions among females to a low of 0.9 for acute conditions among males ages 6-16 and three other statistics. The average ratio is 1.0 for 1982-83 and 1.1 for 1983-84. The conclusion one would make from these ratios is that when one compares estimates between years by assuming that the statistics are independent, the vari- ance estimate is not an overestimation. NHIS physician visits and short-stay hospital episodes The statistics, standard errors, and standard errors of differences for the three years are displayed in table 5. Examination of the correlations for 1982-83 displayed in table E shows that they vary from a low of -0.16 for short-stay hospital episodes among divorced individuals to a maximum value of 0.39 for short-stay hospital episodes among white persons. In this set of variables, no very clear pattern is seen between the size of the estimate and the size of the correlation. Two of the 20 variables have a negative correlation. The average value of the correlations for these 20 variables for 1982-83 is 0.15. Of the 20 correlations for 1983-84, 7 are negative; the negative values range from -0.01 for physician visits among black persons and physician visits among divorced persons to —0.33 for short-stay hospital episodes among black persons ages 45-54. The range for the positive values is from a low of 0.06 for physician visits among males to a high of 0.36 for short-stay hospital episodes among white persons. The mean is 0.10 for 1983-84; therefore, the averages of the correlations for number of physician visits and short-stay hospital episodes are bigger than the means for the number of acute and chronic conditions. Although the correlations for these statistics are slightly larger, they are still relatively low correlation coefficients. The last two columns in table E give the effect of independence assumption for the year-to-year differences. Only 2 of the 20 values for 1982-83 and 7 of the 20 values for 1983-84 are less than one. As an example of the effect of assuming independence, consider the correlations for the statistic short-stay hospital episodes among white persons —0.39 for 1982-83 and 0.36 for 1983-84. Although the correlations are not large, the effect on the independence assumption approach is a ratio of slightly more than 12 for each time period. Thus, by making the assumption that the variables are independent, the vari- ance is 1%2 times larger than the true variance. The average ratios are 1.2 for 1982-83 and 1.2 for 1983-84. Table E. Correlation coefficients and effect of independence assumption on variances for number of physician visits and short-stay hospital episodes, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 Physician visits TOM, ven srs ms A AN ES LHF EEE AP HERE ED YY 0.30 0.28 1.4 1.4 MIB: ics cv a TEER BEE EE S50 hE Be RR 0.10 0.06 1.1 1.1 BIBI. 7 5 2555: 5 00 mime i dm ees ms wom cae 0 2 20 ne #11 wi TR 8 0 0.14 -0.01 1.2 1.0 White, 17-24 years. . . . . . ... 0.13 0.34 1.1 15 White, 55-64 years. . . . . . 0.18 0.29 1.2 1.4 MABE B YBBR... oon vi wom sn ai 0 5 wi Wr £3 EE BEG 0.12 0.09 1.3 1.1 Married, 38-dA'YORIS. .. ov vs vw amy PH IHS 1h 8 EE EY WAS 0.11 -0.08 1.1 0.9 DIVOICBT +o amu sw Gd AMUN D PB HEE UE AE 3 GS E00 AE 0.02 -0.01 1.0 1.0 Female, 65-74 YBAIS. : ¢ ws vs bmi vo tus anis ints my va 0.05 -0.09 1.1 0.9 Female, 75yearsandover. ..............covueuusn 0.37 0.11 1.6 1.1 Short-stay hospital episodes Total: vuom na nm s Bas mR woe TET BF FEN SESE E EH 0.27 0.29 1.4 1.4 SVIVEE 0 oth once otf st 0 soon 8 ast ch 1 km, of Sess, ot 0.39 0.36 1.6 1.6 Female . . .. 0.33 0.10 1.5 1.1 IASI o.oo 0 ms ci 00 on on 0, a 5 FRY CR 0 0.04 0.15 1.0 1.2 TSYOABANCIOVEE . » .ovvv ov woom ve www ® ws wis & 88 0% 5% 0.21 0.24 1.3 1.3 Marned, 85-64 YORIS: + «un + 4 «wis v5 $45 3 FEE © EE TR 5 0.17 -0.12 1.2 0.9 B18 YOUNIS . os cv tr HME CF URE US ARE A RTRE LETS FEE -0.01 0.09 1.0 1.1 OIVOICBH vam + 5.50 00 20% 05 2.800 58 Shs 4 # FTE & wlan bos -0.16 -0.07 0.9 0.9 Male, 17-24 years . . .......... iii 0.08 0.25 : I 1.3 Black, 45-54 years. . . . ..... 0.10 -0.33 1.9 0.8 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 11 NHIS restricted-activity days, bed days, and work-loss days The correlations for restricted-activity, bed, and work- loss days are similar to those seen for the other two sets of variables from NHIS. (See table 6 for the statistics, stand- ard errors, and standard errors of differences.) Table F reveals that the lowest correlation coefficient for 1982-83 is =0.14 for restricted-activity days among black persons 45-54 years and the lowest correlation coefficient for 1983-84 is -0.20 for bed days among persons 6-16 years. The highest correlation for 1982-83 is 0.39 for restricted- activity days among white persons. The maximum correla- tion seen for 1983-84 is 0.29 for restricted-activity days among the total population. The averages are 0.09 and 0.07, which fall between the averages for the other two sets of NHIS variables. The effect of the independence assumption as measured by the ratio varies from 0.9 to 1.6 for 1982-83. For 1983-84 the minimum ratio is 0.8, whereas the maxi- mum is 1.4. The means are 1.1 and 1.1. Table F. Correlation coefficients and effect of independence assumption on variances for number of restricted-activity days, bed days, and work-loss days, by selected characteristics: United States, 1982-84 Correlation coefficient Independence effect (E) Characteristic 1982-83 1983-84 1982-83 1983-84 Restricted-activity days THROMB cos. vc on vem res sc vans so a wton i im ft wi is BE 0% 4 Se 0.16 0.29 1.2 1.4 WIHIB oo cv 0m wre a oem i030 #0 0 0% 6 oe 0.39 0.28 1.6 1.4 MIE 4 ois we £0 0 0 EB pe a Ek 8 Dn ee 0.01 0.20 1.0 1.3 B54 YOUIS : vc om tub sa 2 vata tat mE ami ms Lye 0.04 -0.07 1.0 0.9 BIBBK oii 1 0 te £0500 Bel 2 Ai 2m 0 05 Es 0 RR ER -0.06 0.11 0.9 1:7 Male, 75 years and OVI. .. .. « ove vu whos odin wia iow 475 3 42 -0.07 0.17 1.0 1.2 Black; 4B-BAYSAIS. . uv vw 5 65 vm 5 9 5 2 00 65 4 105 4 00 3 0 Bail -0.14 0.04 0.9 1.0 Bed days MOBY + ov + 2 v0 wt movin ae oon var fe 1 pos oe fe BR Be 0 AEB 8 0 be FOE 5 | EL 0.13 0.04 14 1.0 WBE: oo. cos ai mone son wen 8 i eat 0 in Pte te ir 2 i 0 2 P § OB 0.34 0.05 1.5 1.0 FemalB . ;u us su snsrwrmashs ed ams sos asmasws oo 0.20 0.06 12 1.1 MalB: uw: rmims smismermsmnams sasws Fase £0 pied wa 0.00 0.08 1.0 1.1 Male, Married». ; so ims cms wr EI MB ERE sm mE wine a 0.11 -0.06 7.1 0.9 B-1BYORIS. + iv sttins 9s 5% EME Gnd B5 $n aos Bm rms bw 0.17 -0.20 1.2 0.8 FEmale, atk ... «vive ts as $EI In iMr sHEWS SMBS £1 -0.08 0.15 0.9 1.2 Work-loss days TOA). ws suiems sms wy cms s SHE MEST REE EEG 0.05 -0.01 1.0 1.0 WIE, INBITIB0 «5 ovis ve wits wwe wien 5 win lo wes mid 6 bin 0 0.32 -0.05 1.5 1.0 POMBE : suis vs sh NN RSNA RE RNY MIRE OR AAAS 0.10 0.02 11 1.0 25-B4YBAIB ; vs 5ie vis Be BEEN EEE RT RE EE ie EW -0.04 0.02 1.0 1.0 BYBEIE ov verse me sri mn 0 ne er oo wg 8 ERE R AR ETERS 4B -0.11 0.07 1.0 4.3 SOPBIBIEH 11.» vv i + wvves: sate ain wiah 5 sow ow ono vs a E80 eB 8 0.06 0.13 1.9 1.2 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 12 Impact of inference One of the consequences when there are correlations among the sample units is an impact upon the inferences analysts draw from the findings. Kish (25) discussed the magnitude of the effect that clustering has upon the construction of confidence intervals. Using his strategy and this body of empirical evidence, one can see the effects on confidence intervals of assuming the correlation is zero. Krewski (26) and Krewski and Rao (12) showed that under certain conditions the quantity 6-0 \VAA! is asymptotically distributed as a standard normal random variable N(0,1) when L (the number of strata) goes to oo. This theory was derived for linear and nonlinear statistics and is valid for any stratified multistage design in which the PSU’s are selected with replacement and in which independent subsamples are taken for those PSU’s se- lected more than once. The PSU’s in NHDS and NHIS are not selected with replacement, and the effect of this difference on the result is not known. To observe the magnitude of the effects of independence upon the confidence intervals, correlations l= from each of the six sets of variables for NHDS and NHIS that vary across the observed range were used. The vari- ables, correlations, and effects are displayed in table G. The column labeled E shows the ratio of the variance assuming the correlation is zero to the true variance, whereas the next column gives the ratio of the standard errors. The last column shows the probabilities of exceed- ing the true population value for o = 0.05. When a = 0.05, the coefficient 1.96 is multiplied by the standard error. One would expect that 5 percent of the time an error will be made. However, this percent is either higher or lower depending upon whether the true variance is larger or smaller than the variance utilized. For example, if \/E is 1.3, the number of standard errors being used is 1.96 x 1.3 = 2.55. Thus, 0.0120 of the time an error will be made. Table G indicates that when the correlations are large, the confidence level is zero. This means that using the variance estimate calculated under the assumption of zero correlation leads to no incorrect statements. Correla- tions that are negative result in confidence levels ranging from 0.0524 to 0.0688. Confidence intervals constructed based on variance estimates that ignore correlations can be distorted. These distortions would also be present in hypothesis testing procedures. Table G. Effects of independence assumption on confidence intervals using statistics from the National Hospital Discharge Survey and the National Health Interview Survey, by selected characteristics: 1982-84 Probability of incorrect statements Characteristic Correlation E VE (a = 0.05) Congenital anomaly, Under 18 Years. . ... « «uso smu sw toms wan ou 0.9570 21.4 4.6 0.0000 VE BA YBAIS ov vin ems UR EE WE DE AHS Ak BR Ae SE Ae 8 0.8813 12.0 3.5 0.0000 OAATBC: oie hiv 4 3 RRL E I 0 0 a CARE BE Rh © Sse 8 (8 Do es 0.7347 3.1 1.8 0.0006 Arthroplasty and replacement of hip, 65 yearsand over . . ........... 0.5272 2.0 1.4 0.0052 Chronic conditions, female . . . . .......... i 0.3861 1.6 1.3 0.0120 PhYSICIan VIEHs, UNCBP 8 YBa... « ww «ao wiv ad = @ a 4 wis #0 a vac ow #00 0.1162 1.3 1.0 0.0376 Work-1oss:days; 1emale , sues nv ss sna sv ane sug wh 36 peas © eas 0.0197 1.0 1.0 0.0478 Work-1088:0ay8, T0880 uw vw wis im 2% wi 0 me 8 BT ne wR EE a -0.0131 1.0 1.0 0.0524 Short-stay hospital episodes, married, 55-64 years . . . ............. -0.1233 0.9 0.9 0.0658 Short-stay hospital episodes, divorced . . . . .................... -0.1628 0.9 0.9 0.0688 13 Components of variance The correlations for NHIS are not as large as those for NHDS. In both designs the PSU’s are the same from year to year; it is the second stage of sampling that varies from one year to the next. In NHDS the same hospitals remain in the survey and, for the specified year, discharges are sampled. The PSU’s in NHIS are groups of counties and metropolitan areas; the same PSU’s are used from year to year, but different households within the same segments are selected. The large correlations for NHDS probably reflect the fact that the composition of hospitals is not likely to change when measured by discharge and surgical procedures performed from one year to the next. These features are structured by the facilities and staff of the hospital. In contrast, measures of morbidity, such as acute conditions and limitation of activity, are more variable. If the difference in the values of correlations depends upon the type of primary sampling units, most of the variability in NHDS should be among hospitals, whereas for NHIS the largest component should be the variability among the households within the PSU’s. Casady (27) and Bean and Schnack (28) demonstrate how to apply the BRR method to obtain estimates of the components of variance. To estimate the within component of variance, de- noted as Vy, each of the PSU’s is considered to be a pseudostratum. Here, the sampled elements are randomly placed in one of two equal-sized groups. Constructing a half-sample thus consists of choosing one of the two groups of elements from each of the PSU’s. The data from a half-sample are subject to the same estimation proce- dure as the data from the total sample, creating another estimate of 6. By means of a second orthogonal pattern, B estimates of 6 are produced. Then an estimate of the within component of variance is: B Vi 6) => © —0)%B i=1 The BRR estimator of Vy, was applied to both the NHDS and NHIS data. Each of the sampled PSU’s in the sample was considered to be a pseudostratum. Then within each of these PSU's, either the hospitals in NHDS or the households in NHIS were randomly allocated into one of two groups. 14 Instead of displaying the within and between compo- nents of variance for each of the NHDS and NHIS statistics, the mean, maximum, and minimum of the percent contribution of the within component of variance are presented by year in tables H-N. Tables 7-12 show the percent contribution for each variable by year. For several statistics in NHIS the within component of variance esti- mate is larger than the estimate for the total variance. NCHS assumes that the between component of variance is zero in this case. As observed in table H, the within component of variance for the number of diagnostic discharges contrib- utes, on the average, 13-15 percent of the total variance. Therefore, most of the variability is between rather than within hospitals. For the days-of-care statistics, the mean of the within component of variance increases to 21-23 percent of the total variance (table J). The maximum across the three years is 89 percent. Similar results are found in table K for the within component of variance percent contribution to the total variability for the number of surgical procedures. Here the means are 17 percent for 1982, 18.5 percent for 1983, and 12 percent for 1984. This body of empirical evidence suggests that, regardless of the statistic, the between component of variance is consider- ably larger than the within component of variance. The opposite is true for the 60 NHIS statistics. For the acute and chronic conditions, the within component of variance counts for at least half of the variance (table L). The maximum is 97.9 percent for the year 1984. For the other 40 statistics, the average contribution is higher than 50 percent; the mean percent contribution ranges from 71.6 percent to 77.8 percent (tables M and N). These values provide empirical evidence that for NHIS most of the variability is among the households in the second stage of sampling. Table H. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of patients discharged from short-stay hospitals: United States, 1982-84 Summary statistic 1982 1983 1984 Percent MBBN .: wvitins sw v ms sms mn b 15.3 12.6 13.1 MBXIMUM. iv. co voir min ms mn » 64.3 43.1 43.9 Minimum. .......... 0.8 0.5 0.6 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84, Table J. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of days of care for patients discharged from short-stay hospitals: United States, 1982-84 Table M. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of physician visits and short-stay hospital episodes: United States, 1982-84 Summary statistic 1982 1983 1984 Summary statistic 1982 1983 1984 Percent Percent MBean sss ens nn en ma vs 22.9 22.6 21.5 MBEAN . re ie in stm wm am 75.5 76.7 77.0 MEXIIGIL. «505. 4% 20% 2 onion fed oc 2s 89.3 68.8 54.3 Maximum. . .............. 98.5 97.7 93.2 Minimum... ..... LL. 27 2.3 3.0 Minimum... ...... 35.9 271 56.8 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table K. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of surgical procedures for patients discharged from short-stay hospitals: United States, 1982-84 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. Table N. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of restricted-activity days, bed days, and work-loss days: United States, 1982-84 Summary statistic 1982 1983 1984 Summary statistic 1982 1983 1984 Percent Percent Mean .................. 17.0 18.5 12.4 Mean, ....sbtsvenw enna 71.8 771 77.8 LC 69.9 70.8 B57 Mestimuin,: «coe ow se» swe 4 97.4 99.9 97.8 Minimum. cove nvma vas enss 0.2 0.1 0.1 MINIM . . sues evans vans 25.1 41.5 49.7 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table L. Mean, maximum, and minimum of percent contribution of within component of variance to total variance for number of acute and chronic conditions: United States, 1982-84 Summary statistic 1982 1983 1984 Percent Mean . oc vine ans ns avdahs un 57.5 57.9 61.7 MMPI. 5 4510 ala & msinle os 96.9 91.6 97.9 Minimum. ............... 1.9 1.6 2.3 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 15 Discussion Both NHDS and NHIS are conducted on a continuing basis for the purpose of producing various estimates. Data from these surveys can be employed to estimate parame- ters for all the years combined, parameters for the current year, and parameters measuring change from one year to the next. As discussed by Cochran (29), each of these estimators requires a different design to achieve maximum precision. Sampling theory indicates that for estimating change the same sampling units should be retained for both surveys, whereas different sampling units are better for estimates using data from all time periods. An efficient scheme for constructing a current estimate is to retain a fraction \ of the sampling units from the previous survey while selecting 1 — A fraction of new units. The optimum percentage to retain depends upon the correlation. Mak- ing specific assumptions about the design, Cochran (29) shows that for sampling on two to four occasions, as the correlation increases, the optimum to retain decreases. For example, when sampling at two time periods, if the correlation is 0.7, the optimum to retain is 42 percent, but if the correlation increases to 0.8, the percentage should be only 38 percent. For this design, the best type of estimator is a double-sampling regression estimate. The findings for number of discharges, days of care, and surgical procedures clearly indicate that variance estimates based on the assumption that statistics from year to year are independent are in error. The overestimation of variance becomes larger as the correlation nears one. These empirical results demonstrate that NHDS can be used to detect very small differences from year to year. However, with large correlations the design of 100-percent overlap is not the most efficient for maximizing the preci- sion of yearly estimates. As Cochran (29) demonstrated for simple random sampling and a regression estimator, if the correlation is 0.9 and the number of occasions is four, the percentage gain in efficiency is 59 percent when the overlap is 50 percent; the gain in efficiency drops to 40 percent when the overlap is 75 percent. Research is required to determine if this holds for NHDS. Because the hospitals are the same for each survey, the NHDS design is not very efficient at producing data for combining across the years. However, for reasons of economy, selecting a different sample each year may not be feasible. A possible compro- mise that should be considered by NCHS is a design whereby part of the sample is changed from year to year. Another consideration is using a regression-type estimator 16 that would take into account the estimates produced from the previous year. The formula for the variances of differences given in the appendixes to NHIS reports is based on the assump- tion that the sample for a given year, say 1982, is independ- ent of other sample years even though the primary sampling units are the same across the years. By assuming independence, data are compared using the formula: VY, -Y) =V() + V(Y) The investigation of this assumption for 60 NHIS statistics indicates that the correlations range from a negative value of -0.20 to a high positive value of 0.39. The evidence here suggests that NHIS is great for adding yearly data together, but the design is not as efficient for examining differences from one year to the next. The results do not provide any clear findings on how an annual point estimate would compare with a yearly regression estimate. A caveat to these conclusions is that the empirical results are for totals only. Correlations for means and proportions were not calculated. Parsons (30) estimated correlations for selected statistics from the 1985 and 1986 NHIS. For eight estimates of totals, the correlations ranged from 0.03 to 0.36, which are similar to the values observed in this study. The range of correlations for 26 means and proportions was from a low of -0.03 to a high of 0.38, which is similar to values observed for totals. No results are available for means and proportions for NHDS. A further evaluation of means and proportions for both surveys is warranted. In summary, making the assumption of independence in NHDS may result in extreme overestimates of variance, at least for totals. The results clearly indicate that NHDS is ideally suited for measuring change over time. This feature of the survey should be utilized more than it currently is. In NHIS when the assumption of independ- ence does not hold, the findings show that the variance estimate is not an excessive overestimation. The variance estimated was never more than 1.6 times as large as it should be. One possible explanation is the difference in primary sampling units. Empirical evidence suggests that for NHDS the within component of variance is less than 26 percent of the total variance, whereas for NHIS the within component of variance accounts for at least half the variability. References 10. 11 12. 13. 14. 15. Simmons WR, Schnack GA. Development of the design of the NCHS Hospital Discharge Survey. National Center for Health Statistics. Vital Health Stat 2(39). 1977. Shimizu IM. The new statistical design of the National Hospital Discharge Survey. In: Proceedings of the Survey Research Methods Section of the American Statistical As- sociation. Anaheim, California: American Statistical Associ- ation. 702-6. 1990. National Center for Health Statistics. The National Health Interview Survey design, 1973-84, and procedures, 1975-83. National Center for Health Statistics. Vital Health Stat 1(18). 1985. Massey JT, Moore TF, Parsons VL, Tadros W. Design and estimation for the National Health Interview Survey, 1985-94. National Center for Health Statistics. Vital Health Stat 2(110). 1989. McCarthy PJ. Replication: an approach to the analysis of data from complex surveys. National Center for Health Statistics. Vital Health Stat 2(14). 1966. McCarthy PJ. Pseudoreplication: further evaluation and application of the balanced half-sample technique. National Center for Health Statistics. Vital Health Stat 2(31). 1969. Kish L, Frankel MR. Balanced repeated replication for analytical statistics. In: Goldfield ED, ed. Proceedings of the Social Statistics Section of the American Statistical Associ- ation. Pittsburgh: American Statistical Association. 2-10. 1968. Kish L, Frankel MR. Balanced repeated replication for standard errors. J Am Stat Assoc 65:1071-94. 1970. Frankel MR. Inference from survey samples: an empirical investigation. Institute for Social Research. Ann Arbor, Michigan: University of Michigan. 1971. Koch GG, Lemeshow S. An application of multivariate analysis to complex sample survey data. J Am Stat Assoc 67:780-2. 1972. Bean JA. Distribution and properties of variance estimators for complex multistage probability samples: an empirical distribution. National Center for Health Statistics. Vital Health Stat 2(65). 1975. Krewski D, Rao JNK. Inference from stratified samples: properties of linearization, jackknife, and balanced repeated replication methods. Ann Stat 9:1010-9. 1981. Gurney M, Jewett RS. Constructing orthogonal replications for variance estimation. J] Am Stat Assoc 70:819-21. 1975. Krewski D. On the stability of some replication variance estimators in the linear case. J Stat Plan Infer 2:45-51. 1978. Koch GG, Freeman DH Jr, Freeman JL. Strategies in the multivariate analysis of data from complex surveys. Int Stat Rev 43:59-78. 1975. 17. 18. 19. 20. 21. 22. 23, 24. 28, 26. 27. 28. Freeman DH Jr. The regression analysis of data from complex surveys: an empirical investigation of covariance matrix estimation. Institute of Statistics Mimemo Series no. 1020. Chapel Hill, North Carolina: University of North Carolina. 1975. Freeman DH Jr, Freeman JL, Brock DB, Koch GG. Strat- egies in multivariate analysis of data from complex surveys II: an application to the United States National Health Interview Survey. Int Stat Rev 44:317-30. 1976. Freeman DH Jr, Freeman JL, Koch GG, and Brock DB. An analysis of physician visit data from a complex sample survey. Am J Public Health 66:979-83. 1976. Freeman DH Jr, Brock DB. The role of covariance matrix estimation in the analysis of complex sample survey data. In: Namboodiri NK, ed. Survey sampling and measurement. New York, San Francisco, and London: Academic Press, Inc. 121-40. 1978. SAS Institute, Inc. SAS’s user guide: basics. 1985 ed. Cary, North Carolina: SAS Institute, Inc. 1985. Stanek EJ. The properties of balanced half-sample variance estimates in complex surveys when strata are paired to form pseudo-strata. Biostatistic-Epidemiology Program Series no 77-8. Amherst, Massachusetts: University of Massachusetts. 1977. Simmons WR, Baird JT. Pseudoreplication in the NCHS Health Examination Survey. In: Goldfield ED, ed. Proceed- ings of the Social Statistics Section of the American Statis- tical Association. Pittsburgh: American Statistical Association. 19-30. 1968. Lemeshow S. The use of unique statistical weights for estimating variances with the balanced half-sample tech- nique. In: Goldfield ED, ed. Proceedings of the Social Statistics Section of the American Statistical Association, part II. Boston: American Statistical Association. 507-512. 1976. Graves EJ. Utilization of short-stay hospitals, United States: 1984 annual summary. National Center for Health Statistics. Vital Health Stat 13(84). 1986. Kish L. Confidence intervals for clustered samples. Am Sociol Rev 22:154-65. 1957. Krewski D. Jackknifing u-statistics in finite populations. Commun Statisti-Theory Meth A(7):1-12. 1978. Casady RJ. The estimation of variance components using balanced repeated replication. In: Goldfield ED, ed. Pro- ceedings of the Social Statistics Section of the American Statistical Association. Atlanta: American Statistical Associ- ation. 352-56. 1975. Bean JA, Schnack GA. An application of balanced repeated replication to the estimation of variance components. In: Goldfield ED, ed. Proceedings of the Social Statistics 17 29. Section of the American Statistical Association, part II. Chicago: American Statistical Association. 938-42. 1978. Cochran WG. Double sampling. In: Sampling techniques. 3d ed. New York, Santa Barbara, London, Sydney, and Toronto: John Wiley & Sons, Inc. 327-58. 1977. 30. Parsons VL. Estimation of year-to-year covariances for a national health survey. In: Proceedings of the Survey Re- search Methods Section of the American Statistical Associ- ation. New Orleans: American Statistical Association. 210-5. 1988. List of detailed tables . Number of patients discharged from short-stay hospi- tals, standard errors, and standard errors of differ- ence, by selected characteristics: United States, 1982-84 . is vvinns salve sms SERS 55 ER GEERT SE . Number of days of care for patients discharged from short-stay hospitals, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 . ov vrrrn nnn rns vr amy . Number of surgical procedures for patients discharged from short-stay hospitals, standard errors, and stand- ard errors of difference, by selected characteristics: United States, 1082-84. . uc suv is soni ss smi mans . Number of acute and chronic conditions, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 ,........... . Number of physician visits and short-stay hospital episodes, standard errors, and standard errors of dif- ference, by selected characteristics: United States, 1982-84. . «vc ww wis wavs ws wow ws ves rons wee ene 2 . Number of restricted-activity days, bed days, and work-loss days, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 . Percent contribution of within component of variance to total variance for number of patients discharged 20 20 21 21 22 22 10. 11. 1 from short-stay hospitals, by selected characteristics: United States, 1982-84 ............coivivn... . Percent contribution of within component of variance to total variance for number of days of care for patients discharged from short-stay hospitals, by se- lected characteristics: United States, 1982-84... .... . Percent contribution of within component of variance to total variance for number of surgical procedures for patients discharged from short-stay hospitals, by se- lected characteristics: United States, 1982-84... .... Percent contribution of within component of variance to total variance for number of acute and chronic conditions, by selected characteristics: United States, FOBD-8H + iis isis» 2 rie iris WE pn BEE ESE i Percent contribution of within component of variance to total variance for number of physician visits and short-stay hospital episodes, by selected characteris- tics: United States, 1982-84. .................... Percent contribution of within component of variance to total variance for number of restricted-activity days, bed days, and work-loss days, by selected characteris- tics: United States, 1982-84. ...covvnnnurvvinsans 23 24 24 25 19 Table 1. Number of patients discharged from short-stay hospitals, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number of Number of Number of patients Standard patients Standard patients Standard Characteristic discharged error discharged error discharged error 1982-83 1983-84 Number in thousands Female. . ......... 21,554 635 22,027 687 21,380 695 284 184 VE-AANOBIS,. ; wv: vv ie #00 00 haemo ar ig 14,515 520 14,537 547 13,973 530 216 161 Malg . ovnema arama ens mg ware wh mame 14,428 436 14,805 446 14,296 439 186 140 BEYoars and Over. « «vv ins sa vant np ens an 9,982 316 10,752 362 10,801 375 150 12] ABBA YOBIS, + « 2vis smn swum bly Mase 8,094 245 8,100 249 7,808 255 119 82 Females with deliveries. . . . ............. 3,695 216 3,795 229 3,734 224 73 65 LINGer 18 YORI. www mw wiv wir ww a 4% gor wom Wa 3,392 214 3,443 249 3,094 233 75 54 Mental disorders: « » « sus vs ovr rs vers mrs 1,614 103 1,627 108 1,630 105 34 37 Mental disorders, 15-44 years . . . ......... 892 71 912 77 934 74 22 28 Acute myocardial infarction . . ............ 631 29 642 30 670 29 22 21 Cataract . . . «ooh ic ees 520 38 562 44 467 31 20 28 Atherosclerotic heart disease . . . .......... 465 25 438 28 325 21 17 16 Inguinal hernia, malB . . « » vss rvs ew nss » 417 15 403 17 378 18 15 16 Cataract, 65years and over. , . « « «vo «vs vio 399 30 451 36 384 26 17 23 Malignant neoplasm of lung. . . ........... 301 14 320 | 325 21 16 17 Malignant neoplasm of breast, female . ...... 210 11 230 12 225 13 11 13 Malignant neoplasm of lung, male. . . ....... 188 11 203 1" 205 14 12 14 Fracture of neck of femur, 65 years and over . . . 177 10 184 11 198 11 9 10 Congenital anomaly, under 15 years . ....... 161 28 178 34 164 31 1 10 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table 2. Number of days of care for patients discharged from short-stay hospitals, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number of Standard Number of Standard Number of Standard Characteristic days of care error days of care error days of care error 1982-83 1983-84 Number in thousands FBMIBIB.. 51m vrs sven oc svi vavsm cow save ws 50a #3 4000 nt 4 146,117 3,856 145,768 4,083 134,533 3,772 2,189 1,541 MRS sms moms vw 5 sem pe ew ws 0 108,167 3,042 109,100 3,046 100,161 2,981 1,493 1,285 BE Years ana OVEr: « + sum « wine © 4 se 5 4 ow 100,807 3,148 104,284 3,346 96,016 3,219 1,654 1,394 VE-AAYOBIS, . o.s5 £6 SHE ¥ We WE LR E YE we 74,416 2,468 73,328 2,674 68,586 2,381 1,268 1,035 ABBA YBAIS. « + ov vrin ov vw A Fd 63,517 1,860 61,514 1.217 56,172 1,724 1,011 822 Mental disorders. . ................... 19,541 1,610 20,259 1,763 19,420 1,676 686 641 LINAS 1B YEAS: . wv wivis iv 3 www vow wv vy 15,544 1,124 15,740 1,375 13,919 1,349 505 376 Females with deliveries. . . . ............. 13,172 695 13,341 756 12,810 720 329 242 Mental disorders, 15-44 years . . . ......... 10,534 1,131 11,102 1,271 11,037 1,239 506 466 Acute myocardial infarction . . . ........... 7,074 357 6,980 343 6,655 310 338 282 Atherosclerotic heart disease . . . .......... 4,130 278 3,681 245 2,386 176 209 150 Fracture of neck of femur, 65 years and over . . 3,318 193 3,267 219 3,131 202 202 230 Malignant neoplasm of lung. . . . . ......... 3,215 173 3,375 174 3,089 169 187 177 Malignant neoplasm of breast, female . ...... 2,082 140 2,167 118 1,881 149 135 158 Malignant neoplasm of lung, male. . . . ...... 2,062 121 2,094 116 1,916 129 137 147 Inguinal hernia, male . . . ............... 1,872 88 1,620 67 1,416 77 87 85 CAUBIAOY ; 3 wu sma Rms ms sams vuia ws a 1,519 104 1,431 107 1,105 72 58 73 Cataract, 65 yearsand over ............. 1,184 86 1,167 91 897 59 53 65 Congenital anomaly, under 15 years . . ...... 891 187 985 206 932 213 77 98 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. 20 Table 3. Number of surgical procedures for patients discharged from short-stay hospitals, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number of Number of Number of surgical Standard surgical Standard surgical Standard Characteristic procedures error procedures error procedures error 1982-83 1983-84 Number in thousands All procedures: TOR: feir tsti 170 wovs) Rehet, ER rh RY 33,519 1,186 35,223 1,340 35,811 1,375 619 423 FEMAIE: vst 3 =m iw 0 F und oF aur 2 fm 19,889 793 20,737 881 20,897 867 391 256 VB-AA YOAIS 11 six 5.3 5 05 ae ow GY 14,281 647 14,373 667 14,094 643 273 219 MRIS... © 55 50m 5 of Bw 5 5 0 He = 3 0 I ek WE 13,631 465 14,486 517 14,914 559 261 214 BBYoars and over . ... uv cnn mms Bn 8,132 356 9,295 487 10,108 531 253 189 YB-BAYBATS «+ + sin na worm oh 3% wot 0 411 000 7,747 309 8,077 338 8,243 355 170 133 UNGBr 1B Years . « vow vs wus o 5 wi box Sk 3,360 189 3,478 214 3,366 217 81 76 Circumcision, under 15 years: . . .. «cv. cv un 1,223 68 1,226 77 1,251 80 41 34 Cesarean section: AILARBE... 1 viv iw it 28 58 MIRE 0 3 Bi 4 nie 686 48 766 60 785 63 26 22 BMUYBAE «vo rvsie: or riers gah 1 pe 5) © iitrrg 683 48 764 60 783 63 26 22 HYSIBrettomy «x ve mms vow 0s soem dw 615 48 635 46 638 47 23 18 Bilateral occlusion of fallopian tube, 15-44 YRaIS. . . .. inhi 558 54 532 51 466 45 20 19 Extraction of lens, 65 years and over . . . ..... 425 34 473 36 398 31 19 23 Cardiac catheterization. «« « 4: ws «vam s ws 420 36 478 43 516 49 19 20 Prostatectomy . o vcnw srw ms uptime orem 328 18 332 19 347 22 15 15 Tonsillectomy, under 15 years . . .......... 249 18 255 17 209 18 13 12 Prostatectomy, 65 years and over . . . ....... 241 14 256 15 264 7 12 13 Hysterectomy, 45-64 years . . . ........... 171 13 169 10 180 13 9 10 Direct heart revascularization . . . . ......... 154 18 173 18 171 19 10 10 Myringotomy, under 15 years. . . . ......... 132 12 152 17 114 14 10 9 Arthroplasty and replacement of hip: BBYyearS anti Over . . ....cvcnn toss 93 6 112 8 135 11 7 10 ROMBIB vi ve 5 te oe vw off 70 0 og 81 91 6 103 8 124 10 7 10 Hysterectomy, 65 years and over . ......... 55 4 49 5 58 6 5 7 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table 4. Number of acute and chronic conditions, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number of Standard Number of Standard Number of Standard Characteristic conditions error conditions error conditions error 1982-83 1983-84 Acute conditions Number in thousands BOW. «tere wnssdin nescms Amable sme 25305 379,598 7,665 400,707 6,635 410,812 6,387 9,730 8,634 White, 25-34 years . . . ............ 58,686 2,435 60,095 2,076 63,126 2.275 3,369 2,969 Female, B-1BYEAIS. . wu vovms vss susie 45,124 1,826 45,470 1,841 45,810 1,770 2,744 2,441 Male, B~1BYBBIS ov sivas i dims 3 Hains 5 Wie 41,766 1,855 45,766 1,903 47,082 1.7V7 2,603 2,744 While, B8-BAYBAIS . ; vs ww vis cus ws noisms 20,578 1,116 20,338 1,022 21,522 1,181 1,509 1,687 Married, 45-54 years. . . . ....... 19,526 1,165 18,625 1,192 17,403 990 1,645 1,610 Female, 55-64years . . ................ 14,648 959 14,008 1,024 14,012 919 1,404 1,344 75Years and Over. . . ccs sm amin sis aw aw 7,969 751 9,581 808 10,245 844 1,058 1,092 Separated, 17-24years . ............... 908 240 898 220 797 262 330 342 Black, B5~74.YOAIS 5 su sin viv wis tv swam wi wins 678 151 1,482 376 1,564 323 391 499 Chronic conditions TOBE ie ve 2.30 5 0 0 4 Sab 60 0 #0 08 BO A 227,114 2,970 229,322 3,271 231,606 3,179 4,382 3,686 FOMBIB. . vs om vo 3% 5a 0 smh om wil hs 5h 53 117,579 1,703 118,625 1,764 119,771 1,719 2,559 1,930 25-34 Y8AI8. . s.r vn OE EE AEE 38,499 834 39,149 815 40,003 934 1,227 1,112 Married, 25-34YEaIS.. . . ss v0 55 vw wi ie na 8 27,520 778 27,272 750 28,199 818 1,038 986 While, 48-BA VRAIS , . uv + v5 sid wd wo nals wey 19,400 605 19,470 573 19,374 580 823 869 Male, B-1BY8aI8 .....ovvmvmsvmenms sins 19,225 731 19,251 669 19,153 618 928 913 Female,45-54years . . ................ 11,540 419 11,478 315 11,502 398 515 526 Income less than $5,000, 17-24 years . . . . . .. 3,892 319 3,936 358 3,386 301 487 465 Black, 35-44 VRAIS . us wv v vp am bone 3% 0 ma 2,889 233 2,999 289 3,168 266 334 370 Income less than $5,000, 65-74 years . . . .. .. 4.568 134 1,407 145 1,124 143 192 206 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 21 Table 5. Number of physician visits and short-stay hospital episodes, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number of Number of Number of visits or Standard visits or Standard visits or Standard Characteristic episodes error episodes error episodes error 1982-83 1983-84 Physician visits Number in thousands TOMMY ie i mr i 00509 0) 85 2 90) 4 om 3 56040 799,980 8,325 808,320 10,064 841,760 10,109 10,992 12,108 MBI. «cc iz smn han rman a ow vn rm a 322,307 4,630 331,788 4,944 345,555 6,255 6,416 7,722 BIBCK. 5 158 #5: Sih £5 TH 2105 105% 10 7 93,158 3,576 92,198 3,682 103,190 6,182 4,765 7,225 WHILE, 17-28 YBBIB . ... vv sus ns sms wm sma a 84,617 1,990 82,592 2,427 79,350 1,911 2,933 2,531 WHS, SB-BAYORIS . «vo sis nis sma so sins 4 83,482 2,848 80,999 2,866 83,550 2,297 3,654 3,108 UNABrB YEAS ; uu wus viv vars i ne us aii» 77,967 1,617 83,171 1,825 87,455 2,123 2,293 2,673 Married, 35-44 years. . . ............... 69,487 2,113 71,771 1,866 75,432 2,907 2,656 3,571 DIVOTEBE... . «os is 7m swine wan 58m 5 wo 45,325 1,716 49,610 2,255 52,605 3,120 2,799 3,867 FEMAale, 85-74 YBAIS .. vv + iss s5is sw sda v 48,856 2,115 46,099 1,828 49,763 1.777 2,722 2,658 Female, 75 years andover . ............. 33,828 1.212 36,185 1,507 41,131 2,428 1,547 2,710 Short-stay hospital episodes TOURE co rie v3 ioe ns 1 61 5 3 a WUE 908 06 wp 3% 30,041 455 30,241 402 29,294 435 520 501 WARES. © crs vo ms hs 1m 40% 6 8) £ 008 S005 99° 25,854 421 26,161 390 25,534 404 448 450 FOMBIG. + ova noo 2 0% & 3 Finds Hominid & 6 17,806 307 17,949 291 17,484 310 347 403 MEBIB zd nas sass Bs @md osm es mabe was 12,235 250 21,292 233 11,809 247 335 313 75years and over ............ 0000s 3,128 121 3,330 136 3,522 126 162 162 Married, 55-64 years. . . ............... 3,013 117 3.312 141 2,826 110 166 189 BAB YORE . . . sis vp wis 46 a Wk ew 1,697 87 1,674 71 1.511 70 113 95 DIVOICBO «vv + 5 wv 6 5 mn on wiiians & % 4% 9% $33 1,682 95 1,725 94 1,678 81 144 129 MBI, 17-24 YBAIS) ».iivvine + 4pm isc es Fame a 1,048 76 993 61 882 66 94 78 Black, 45-84years . .................. 393 47 335 36 325 41 56 63 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. Table 6. Number of restricted-activity days, bed days, and work-loss days, standard errors, and standard errors of difference, by selected characteristics: United States, 1982-84 Standard error 1982 1983 1984 of difference Number Standard Number Standard Number Standard Characteristic of days error of days error of days error 1982-83 1983-84 Restricted-activity days Number in thousands TBUBN sv me 3 0 0 ors ime i 0 3 5 0 a6 5 SE 0 3,253,254 59,036 3,318,069 47,805 3,427,193 50,960 69,712 58,830 WRB. vv ov 5906 5 0% 9 8 W538 Hid atin oe 6 est 2,746,275 48,420 2,813,111 46,009 2,904,705 51,284 52,172 58,632 VIBES ot 3) 5 0 51 5 ERA 0 Frc 0s ds om 1,372,863 34,439 1,364,731 27,972 1,454,136 31,144 44,203 37,368 BE-BAYOBIS. . cs tvs ma mee 514,393 20,568 501,147 19,684 511,634 20,659 27,929 29,529 BIEOK. 710 bo & wd fo Brstob 5% Eo Al B00 ym const 438,323 29,174 451,903 22,874 460,593 19,943 38,133 28,573 Male, 7Syearsandover . . .............. 92,191 8,587 119,477 9,041 125,864 10,213 12,509 12,472 Black, 45-84 Y8aI8 . . . . . ocr vans uns Es 66,922 6,939 51,848 5,468 54,137 6,186 9,432 8,104 Bed days TOUR = sts a 85 i 530 emi 510 ie 1,444,556 34,768 1,529,698 31,078 1,508,203 29,954 43,529 42,235 WHE. «sve ss vn an aS mE Fem ws 1,182,455 27,857 1,264,351 28,185 1,251,628 27,523 32,083 38,494 FBMBIG. + vows miming ama we EWE ESS 852,706 22,910 897,783 25,100 884,031 21,982 30,397 32,341 MBE cinema visa sam a ad ae A 591,850 20,636 631,915 15,848 624,172 16,446 25,967 21,849 Male, MBIBO. +. vs sic vn ums mu smn mmsms 326,080 16,128 339,074 13,514 349,820 12,702 19,896 19,127 B=ABYBAIS . . iv vv vin iy es a ws 145,403 6,858 157,084 6,016 155,876 6,615 8,303 9,779 Female,black ...................... 144,512 11,742 143,765 11,225 135,590 8,141 16,909 12,869 Work-loss days Total rcs germs smi maama im avis vmvme 452,615 12,151 419,249 11,784 513,896 13,757 16,538 18,231 White, married . . .................... 269,623 10,207 245,046 9,405 295,453 11,193 11,486 14,963 FEMAIE.. . .v. sms uh ew sll nore aos IE 350% y 227,451 8,684 212,161 8,949 254,601 10,434 11,825 13,611 ET rT 125,901 6,821 116,192 5,484 142,572 6,752 8,914 8,611 BACK civ ums viv 05 tv ws de Br 3 BE AE 53,450 4,834 49,996 4,978 70,026 5,304 6,980 7,012 BOPBIBIOE « caso vists mis ns oni sw ow 5 on 12,983 1,659 16,876 2,605 16,756 2,944 2,999 3,658 SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 22 Table 7. Percent contribution of within component of variance to total variance for number of patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Percent POMAIE ..0vm ivi bamnnmss nd Fae anmes BE new fees eres 0.8 0.5 0.6 MIBIE .... +. + cov am mi oe rom i SC th 4 eh 0 TR 8 BE PRR RE BYR 0.9 0.9 1.0 VB BBYERIB von ow oc 50 gy 3 Ge 3 Srienin < SALG § Bt DARE ERE 1.3 1.0 1.7 BS years ANAIOVBY. «vu 5x vin amin sm FE eM + 0 B® 1 8 wos & BEE 48s 1.2 15 12 ABBA YTRIS von s {55 20 FEE SFT UE Ga B85 8 85 ¢% 40 bre SH 25 29 3.0 Females With dBlIVBHBS. . . ov uo + samara saw ns wasn % wm 3 wesw 24 1.8 23 URGE IBVBAIS 0 vonins aol bedi oe 0 hie mht B00 5 I oe 0 0 3 1.5 1.4 1.6 NASIR THSOIBBIR,. .. oven rie imo iit bi, Bi 0 0 Bhs 0 0 le 4 lH 3.7 3.6 3.8 Mental disorders, 15-44 YEAS . . « « » vws winwms #iniem® wh 0k «®lm 200 4 4.0 3.9 4.1 Acute myocardial INfarction . . «« «crs as sara rare wr wre ammo 14.9 171 20.6 SREBIIICE SL + 5 5 0 0 io 60 0 8 0 tt 3 00 Se 10) 0 7 cm 0 Br 9.3 57 12.3 Atherosclerotic Near CiSease. . .v . : cc cuss th vn s Uva wmv ose 14.2 17.2 14.3 IQUE REINA, BIE. «i ip oo ov wie io 0 5 ion oof ors not lL 98 Ee 64.3 29.4 36.9 Carat, BEYBAIS ANG OVBE... . vo win civ win voy fit ins sib aif 8 dw 4% 9.6 5.7 14.0 Malignant neoplasm Of WONG. « «vou ws 5 & 5 5 81% mie 48% ob lie 81% wrk imide 0 41.4 30.2 19.9 Malignant neoplasm of breast, female . ....................... 35.4 43.1 27.1 Malignant neoplasm of UNG, Male. . « v. «vs vs sur srs sms vue Euan 38.4 37.8 34.2 Fracture of neck of femur, 65 years andover. . . . ................ 45.3 32.7 43.9 Congenital anomaly, under 15years . . . ...................... 35 3.0 6.4 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table 8. Percent contribution of within component of variance to total variance for number of days of care for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Percent FOMAIR ouvir sss sass smiis fMews smanpamonans vu 27 3.5 3.6 MBIB , i svc in vic mn a Bains EE amin CURA SEA ERA RE EE 55 5.5 5.6 YB-AAYOBIS o.voin ch moons Fis da uals SHAME REEWE wlth oie By 3.4 4.2 3.8 BE VEAISTAI OVBE. «oc. viv 4 oom nimi wims win wos i bl Aoi 88 9% 50 RI Sm 2 3.3 3.6 5.1 ABBAYERR sami sn sma msm Imm i ws Lh sR smTR I Ams Er rE Ha 6.5 9.4 72 Memtal SOTHBIS. « « ov + vs vv sme cn sms memo puss dns svn sxe 6.8 4.9 3.0 INOS 1BYBAIS «ov inn avi iw 0 i Fun 31 6 (8 woos 91s Hf oe 5 5 i wi B08 1% 4.1 23 5.0 Females With QOlVENBS: . uo +o sss ua tas sR as AE EEE so daa £580 3.1 37 4.2 Mental disorders, 15-44 years. . . . . . ... cc. viii 6.4 5.9 35 Acute myocardial INArCHON . . . « «+ «vo ws so rms ware wns urns ny 252 27.6 30.4 Atherosclerotic heart disease. . . . ...............covviunnun 18.6 30.6 22.0 Caitaract, CB years ANGIOVEBY. .. uw iv v5: #10 0/4 5 913 5 5 3 861 0 1s #0 hig 3 006 4 92 4 53.3 51.2 47.1 Malignant neoplasm Of HUNG. ox « xu wi sie 50 wis 30 w 30 30s 3 08 ok 5 20m 8 49.5 58.2 53.8 Malignant neoplasm of breast, female . . ...................... 55.9 66.1 42.8 Malighart neoplasm of JUNG, TABIB. . « «vc + vw ssw + wesw £5 8m soe vs 64.3 68.8 54.3 TAGLIE) BOTINR, FBI. + vin rv 0 woo mw om md be A mm 00% T8 K Wh 0 89.3 43.2 52.0 CRBTBOE. riven mons wise, pied wily 0 fe 000 A 0 ate 8 ody 10 0%: Be Fit co 14.3 12.3 24.7 Calaract, BB YSIS ANA OVE. . cu + 5 5 56 vu sive sv Wain om Sain sw vows 15.7 14.2 27.1 Congenital anomaly, under 15years . . . ................c.0uu. 7.4 13.6 13.4 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. 23 Table 9. Percent contribution of within component of variance to total variance for number of surgical procedures for patients discharged from short-stay hospitals, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Percent All procedures: TOBY + vie viv win simis iw sims 68 SMT Mss AW EMa ESET EOE 2H 0 0.1 0.1 0.1 Female... corms ens mars nionis IRIs SHENG IRIAT ABE 4T £0 0.4 0.4 0.4 V5BAYBAIS . i ov con sh ss vu ins vmems ameme ameme rome mn bY 0.7 0.6 0.8 MIB ih Salt Ba LRT BEG BE FE rn wes een 1.0 0.9 1.1 BEyears and Over . .. ..... concn enn rms vas see 1.3 0.8 0.7 AB BAYES sv. som kin so Si 5 HGR HEE RE RE WEBER 23 1.2 1.6 INCE 1BYRBAIS «i « wiv te vik iid eds Mia Abid Wa ARE REE EERE EAE 2.0 2.2 2.3 Circumcision, under 15years. . . ........... cure nnrnennns 7.0 53 7.7 Cesarean section: BUBGBEL oo 0 0065 55 & al 0 Brn lh Bre BU 9532790 92 5 tv, 0 1 ic a 3 wie 6.5 9.4 5.2 V5-AAYOBB vx v vs kn th EWI r I ERISA EAL AEE ESF SHG + 6.5 9.4 5.1 FIYSIBIBCIOMY + wv ov 0s iv + nti 50 0 3% W803 EH 500 & HBS BE 5 7.8 1141 9.2 Bilateral occlusion of fallopian tube, 15-44 years . ................ 6.2 7.5 6.9 Extraction of lens, 65 yearsand over. . . ......... i. 7.6 5.6 8.2 Cardiac catheterization . . ............ 5.7 7.4 5.0 PIOSIBIBOIONY «r+ + 4 45 iv 6 we: 8 win 10 on 1 0 [5 nt 32 mn 1 19.0 23.6 215 Tonsillectomy, Under 1B YBAIS « . « vs x vs cwsw i sum was ow sae «ws 19.2 35.9 19.2 Prostatectonyy, B5Y8ars ant OVeF. . . «s+ + + & + a 5 iw bb % 45 49 8a» 40 33.2 32.5 24.0 Hysterectomy, 45-—BA YBAIS , . «xv + «vs us bw sms as vd Sua 3 wa 5 EW 27.3 37.8 36.8 Direct heart revascularization. . . . ............... 0... 17.0 16.7 13.1 Myringotomy, under 15 years. . . . 28.1 18.5 11.2 Arthroplasty and replacement of hip: OSYORMS AN OVEY + + vos vss esata BIRO SHE PH IDI £0 H6 4 £5 2% a 64.9 60.4 22.8 FOMBIB «vom run imemusmss Sidi in sRi dw iNeiPiRs RIES 58.5 67.2 27.3 Hysterectomy, 65 years and over . . . ........ oe 69.9 70.8 55.7 SOURCE: National Center for Health Statistics. National Hospital Discharge Survey, 1982-84. Table 10. Percent contribution of within component of variance to total variance for number of acute and chronic conditions, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Acute conditions Percent TOMB 7 oie 000 30 rs 5 ia ss LE 0 rl tn ls) AR 8 4.3 3.8 4.4 White, 25-34 years. . . . ........ ivi rin rnin 39.3 72.1 60.2 FOMAE, B-1BYBAIB. . . vos sw vm oun oie wiv wwor eww we wy 68.5 78.9 53.5 NEG, B=1B YBAIS «vow viv vo Sink Dimas les 5 65a BEY 5 E50 55.7 64.9 97.9 WHE, BB-BA YSIS... » vos» bass 508 Faw EEE EH ELT EES 48 SWE 96.9 91.6 85.1 Mario, 45-BA YOM. « «oun b ive ess sa ves vs LEE LEED EE LEE 92.8 56.9 " Pompey, BE5-BAYOMB. «ov 2.5 6 5 5 ow ov a 5 dos do od ts Som 8 " 65.2 94.3 TE YOAIB BNO OVE. © 5 23 + 6 4 0 va woe 0, 005 wwii vd ooh mo ion itomn ow) 10s cw 79.4 91.4 62.0 Separated, 17-24 YEArS . . . . «oot 4) Mh 70.1 BIBOK, BETA NBEYS: , vs cio 3 5 410 0 003 13 5 407% S08 1500 50%: 35 6% 9 8% 8 71.0 80.0 94.2 Chronic conditions TOMB 2 5s BIE RE DES 5 57 0 5 0 3 0 50 0 he 790 i) mt 0 aE 1.9 1.6 23 FOIVAIB cv fui ni VIR Bie 5B BD: Ri BI ot 5 (5D Fk % TH 9 to BE 0 27.1 24.1 31.5 PEBAYORE ... co onc vv sms sn srsy sR sRS EHIRR EERE TY y 76.5 61.5 49.9 Married, 28-84 YBRIS. . . vi. wu sins i548 FREE SRI TH EM £05 60.0 42.0 48.4 White, 4B-BAYBAIS . . + v2 vvis vn smn simp sims wm amn $08 656 bmn 74.3 725 92.2 Malo, BIB YBAIS . vic «voc hit finns sain wn mn inn vn mn none wm 40.0 45.6 59.8 Female, 45-54 VAIS. . . . . ov vi M M M income. ess than $5,000, 17-28 YEAS . + ui vss wis wo sms ww yw & ws inn 73.4 37.3 60.0 Back, SB=A4 YOAIS , 4. . vor wa suir Hrs WE AEH BRET FRR BE EA 27 66.1 65.6 Income less than $5,000, 65-74 Years . . . . . .... vv v einen. M 85.9 88.2 1 The within component of variance estimate was larger than the total variance estimate. SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. 24 Table 11. Percent contribution of within component of variance to total variance for number of physician visits and short-stay hospital episodes, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Physician visits Percent TOMA 5.0% Hus ims a vin AA gE AE EW He BE RRA Bae AW 53.3 35.7 72.7 RENE... + coc cos i 50 won ov ew em ie we WE EA HE HR ee 59.6 71.8 M BlAGK: ws ms vs na sms main mnwn Rave BRAS ERE Fm Ewe Amie BY 35.9 27.1 67.9 WHE, A724 YBBIS « «os vv rvs win cmeun ges Brame Eo ws swim oi 87.1 82.7 80.3 While, BB-BAYBBIS : 1: ms 151% 8 EEF ER Ams RH SMA SW AWE CHE wS ne 71.5 78.3 83.0 LINCRrG YORE ois ms 2 mais A035 5 E005 6h 5 Machi mE lb wos WHR Eo 76.3 63.7 56.8 Married, 35-44 Y@ars. . . . . . ote 68.9 91.4 64.7 INVOIBEILL. .. +. «os monn oi cor wisn monn oo 0 ow 0 cat 300 9 6009 500 90 5 0 0 or a 0000 00 sm 0 00 0 Bi 98.5 88.6 93.2 Female, 65-74 YEAIS . . . «oo ite ee M 78.0 " Female, 78 years and Over . . «us cus mw sms is sm ele ® oy (iw Swe yo 98.1 89.3 74.9 Short-stay hospital episodes TROURI c. .cx item 0 0 mo con nen 0 0 he oF 0 9 9 F685 0 Be 581 20 5 B01 62.6 87.2 75.8 WVARBE. .. + vn wim ven nice es 0 1 8 in on 3 9 om i fT mc A abe 0 SA 64.4 83.6 68.4 FEmMall vino cnmiv dae 185408 SR IMO IRTH EAMES 0k 056 a ws ve 73.3 87.7 73.9 WIBIE rif 5 im 0 0 0 REL 201 50 nd OE 0 U6 I, 0 5 98.4 M 92.3 TE YORIS AIC OVE © 5 5k 5 vik 4 0 5 9.000 50 0 Bh 0 9 82.5 89.9 M Married, 55-64 YEArS. . . . ov vi 88.7 M M BIB YBRIS «co vom cv vv 31 i 9 ew oe ee RE 76.4 O 83.2 DIVOICEBH, i» + 4 viv 5 5 sms 3 Bowe @ 16 F000 5 W300 505% 4% $900 0% 20h) 0 83.3 97.7 90.7 Male, 17-08 YORIS ou wu 06 m5 5 0 GRE 6 0 Hal 5 8 bis wos 90 54 5 nos 92.7 " 76.3 Black, A8-BAYBAIS 1 ss 2c sx mans ned a meh Ed AE WE Kava va sbey 62.9 74.4 77.8 The within component of variance estimate was larger than the total variance estimate. SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. Table 12. Percent contribution of within component of variance to total variance for number of restricted-activity days, bed days, and work-loss days, by selected characteristics: United States, 1982-84 Characteristic 1982 1983 1984 Restricted-activity days Percent TOE «vcs pe ams ms wns ma Bs Es SHE HE TH 5s CHET G a H5 E8038 45.0 82.1 62.1 WIE, vn ams 53 SEHR ERR RF CHE DE SHR HAE GSA TMs Has Bde 59.4 77.7 49.7 IAI! 5 5 05 5 2 Bo 7 ied 8 on 902 re vt a v8 © Fh oer af Rn Ao 54) i BE 56.0 89.2 80.2 BE-BAYBAIS . . . occ vhen ra ae ee a ny a 83.5 62.0 75.8 BIOK. . .cvvvruswrsmummsmismews shows amen asms «hides 25.2 41.5 88.8 Male, 75 years anti OVBI. ««; «vv ws sass uu s@3 SR 8s isms & eins 84.2 M 97.8 Back; A558 YOUNIS ; 45. cos 05 CMs 9 cus Eoin mk Bhp me nM EEE Fue 84.8 M 84.8 Bed days TOW viv covmn naps twa my conga sis 68 IBIE HE FMP ny An vi diy 58.9 77.4 12.7 WIIG: cove 1h sme FRIES Emi mn sMEWE IRIN EGE RR SWE 2h amy 76.6 79.1 65.8 FBINAIG 0 iin cmon nmsms nmames smemn smsma mame ame sm 2ms 89.2 64.1 77.2 ABLE 0) eee se dos tate omg ets ed ed rl 2B 22 She 3 he £08 RE 64.9 ™ (" MEE, HOHE » ovens sr a@s swans anisms ER EEE SRI ME Ba Ei0 3 50s 69.0 M M" B1BYOBIS . iyi «ns MERE LEAD EEE Bs Pees Aim NEE Fwd 82.3 90.9 79.4 FOrmale, IVI ou vais cv sms bw sms vim ss ham i sR hs Se se 59.3 76.9 Mm Work-loss days BEEN co.» 203 0 0m 2 an ot es on 2 ew 4 PAB BE BB Mm M M White, marfied . . . . . oo. M 77.6 89.9 FONG wis co vw 5 m0 5 0 00000 500 Bo 05 506, ht 0 6 a 97.4 Mm 87.5 PE-BAYOAIB viva us vB ES EES ER Baas nh EE EEE EE na EE 84.4 " M BIRO: 55: 510.5 7 15 55 5 2 080 5 15 fal i wr 10 Se 00 5 it 96.1 91.3 " BOPBYBIBG: «. ovr nos 3 mis 0 vs 2m se a mia 0 mens 0 0 Sh i Fer " 84.4 77.1 The within component of variance estimate was larger than the total variance estimate. SOURCE: National Center for Health Statistics. National Health Interview Survey, 1982-84. ¥¢ U.S. GOVERNMENT PRINTING OFFICE: 1992— 3 12 - 0 82/ 6 0 0 03 25 . = a hs ) . - CE NATIONAL CENTER FOR | INARI Ie Vital and Health Statistics A Method to Redefine Stays on the 1985 National Nursing Home Survey March 1992 CENTERS FOR DISEASE CONTROL Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested Citation Jonas BS, Madans JH, Rothwell ST, Bush MA, Feldman JJ. A method to redefine stays on the 1985 National Nursing Home Survey. Vital Health Stat 2(115). 1992. Library of Congress Cataloging-in-Publication Data A Method to redefine stays on the 1985 National Nursing Home Survey. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 115) (DHHS publication ; no. (PHS) 92-1389) “March 1992." Includes bibliographical references. ISBN 0-8406-0455-6 1. National Nursing Home Survey (U.S.) — Evaluation. 2. Nursing homes — United States — Length of stay — Statistics. |. National Nursing Home Survey (U.S.) II. National Center for Health Statistics (U.S.) Ill. Series. IV. Series: DHHS publication ; no. (PHS) 92-1389. V. Series. DHHS publication ; no. (PHS) 92-1389. [DNLM: 1. Length of Stay — United States — statistics. 2. Long Term Care— United States — statistics. 3. Nursing Homes — utilization — United States. 4. Research Design. W2 A N148vb no. 115] RA409.U45 no. 115 [RA997] 362.1'0723 s—dc20 [362.1'6'0973] DNLM/DLC for Library of Congress 91-36707 CIP Vital and Health Statistics A Method to Redefine Stays on the 1985 National Nursing Home Survey Series 2: Data Evaluation and Methods Research No. 115 This report describes a method for standardizing definitions of episodes of nursing home care in the 1985 National Nursing Home Survey. The method shows how the information on nursing home admissions and discharges collected on the Current and Discharged Resident Questionnaires can be used to redefine the endpoints of nursing home stays. The report also explains how errors caused by missing and inconsistent nursing home admission and discharge data were resolved. Er rE viyB, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control National Center for Health Statistics Hyattsville, Maryland March 1992 DHHS Publication No. (PHS) 92-1389 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Division of Analysis Jennifer H. Madans, Ph.D., Acting Director Jennifer H. Madans, Ph.D., Deputy Director Diane Makuc, Dr.P.H., Chief, Analytical Coordination Branch Sandra T. Rothwell, Chief, Longitudinal Analysis Branch Jennifer H. Madans, Ph.D., Acting Chief, Demographic Analysis Staff Contents U.S. DEPOSITORY AAA JU « Ga 15992 Tr OAUC HON. sso ims smear msiems NIMs FRIAS (RIB B inion RINE IRAE HERI RE FRING I ANCE ERs an hess Em io Bagkatonnt cus wienins mass apm miEs ann 0S RNR Ry tm Ee Re PRE EBERT EM PR Ia Procedures for redefintion of admission and discharge dates of the sampled stay ......................... 00. Procedures for editing date InCONGISENGICE uu rm mr inspmimsmscmnamamsscmmemsionciismpimsmminsions ss nssiammss RE SOTTIVE GIES 2 or 0700 2 5000 17 sme se is 56 3 1 0 0 0 R08 0 0 4 0 0 0 ff Changing InCONSISIENt GateS «cx oc vv ma ams inv ca tmm mem s Boman asap sm ems soils Homes MAE 30 00 L000 0h rns i010 ued TMputing MISSING GALLE vows rrns brn mrn ues hus dsa mans ueksnmbmbn sensi aRrnsimytnr@P sia R sm swnins mis Results Of date rede MITION. «vv ucviimar sess snsansnsnsesensssssntnsesssssnssssssmsssssssssavsvsnssnsosss List of text tables moowp m Number and percent of intervals spent entirely in hospital by interval length .............................. Number and percent distribution of cases by type of date COTECHONS. +. vv cont imtmsersarsmemsensmnsmomnss Number and percent distribution of eligible cases by redefinition status of admission dates ................. Number and percent distribution of eligible cases by redefinition status of discharge dates .................. Number and percent distribution of Discharged Resident Questionnaire (DRQ) cases by length of stay (LOS) for sampled and redefined admission Bnd diSChArge AALS. « «uv vv ivan vr mimn ims mmems ined og wimeswsiwsomsmses Number and percent distribution of cases eligible for redefinition by redefinition status of admission date and SOIL 8tatUS Of JaLe BETAS warms vai mam eam ais EE 4s WoT bh AF RII 5 40 17 1 A 210 1 A 110 Worse $05 9 4 16 50 10 045 1 A RMS MATH 0 G. Number and percent distribution of discharged resident questionnaire cases eligible for redefinition by H, redefinition status of discharge date and edit Status OF Gate AYIAYE «ux rina vsmsimss nin vesansss sis omensss im Number and percent distribution of discharged resident questionnaire cases eligible for redefinition by redefinition status of length of stay (LOS) and edit status of date arrays...............cooiviiiiiinnn, List of figures 1 2. 9. Illustration of variation in the definition of nursing home stays ............cooiiiiiiiiii iii iii... Questions added to the 1985 National Nursing Home Survey to obtain nursing home and hospital utilization TUESLOVEY 05 1500 0 3 sh 0039 30 3 0... 0.0 0000 4.00 0 8 0 0 0 0 0 40 0 0 06 00 160 0 0 Number and percent of cases in the combined CRQ/DRQ sample with and without information on additional TERING TONE SUIVS, 5 von win 10 1 0 nin von riers 8) 500 0% fn lal 0'9 8 WR $0 7,004 L806 20 RT 4 90% rs 00 of 0 00 0 fle 4 0 nde 00 Schematic representation of the sampled admission/discharge date redefinition procedure. .................. Number and percent of cases in the combined CRQ/DRQ sample by whether dates were redefined.......... Percent of cases in the combined CRQ/DRQ sample by difference between sampled and redefined admission dates (N=LASY) oor urmv russ nes tenn ins vanes 18 vip BLE FEE 0 © S050 17% wl $F SLES 1B 08 © Pain $8 ms ba Percent of cases in the DRQ sample by difference between sampled and redefined discharge dates (N=747).. Difference between sampled and redefined admission dates by whether cases were edited, CRQ and DRQ IN EL BAY, 0 cir eve Taro wom rds rg i 300 os 3007 4 0s 0 0 To 4 6) a, 00 Difference between sampled and redefined discharge dates by whether cases were edited, DRQ only (N=747). 10. Difference between sampled and redefined length of stay by whether cases were edited, DRQ only (N=747).. xx aN BH ND 11 Co C0 J Wn 10 11 12 12 Oo 11 12 12 Symbols Data not available . Category not applicable Quantity zero Quantity more than zero but less than 0.05 Quantity more than zero but less than 500 where numbers are rounded to thousands Figure does not meet standard of reliability or precision A Method to Redefine Stays on the 1985 National Nursing Home Survey by Bruce S. Jonas, Ph.D., Jennifer H. Madans, Ph.D., Sandra T. Rothwell, M.P.H., Mary Ann Bush, M.S., and Jacob J. Feldman, Ph.D., Division of Analysis, National Center for Health Statisitcs Introduction The purpose of this report is to describe a method for standardizing definitions of episodes of nursing home care in the 1985 National Nursing Home Survey (NNHS). This standardization is necessary because not all nursing homes use the same definition of a nursing home stay and, thus, reported lengths of stay are not always comparable. Length of stay in nursing homes is important for policy decisions such as allocation of resources and planning for long-term care insurance. The construction of length of stay estimates is complicated in surveys, such as the NNHS, which use facility based definitions of a nursing home stay as the sampling frame for current and dis- charged residents. It is necessary to consider a resident’s entire pattern of nursing home usage, including multiple nursing home stays and intervening hospital utilization, in calculating length of stay. Studies of length of stay have been published using the 1977 NNHS (1,2). However, in that survey, information was not collected on nursing home stays other than the sampled stay. More recent surveys do include information on additional nursing home stays and hospital utilization. Adjustments to length of stay estimates have been made by Spence and Wiener (3) using data from the 1985 NNHS, and Short et al. (4) using the 1987 National Medical Expenditure Survey. The methods used differ from the method presented in this report. Background The 1985 National Nursing Home Survey (NNHS) collected extensive information about nursing homes and their residents. The sample frame consisted of all nursing homes listed in the 1982 National Master Facilities Inven- tory (NMFTI), nursing homes identified in the 1982 Com- plement Survey of the NMFI, facilities that opened between 1982 and June 1984, and hospital-based nursing homes identified by the Health Care Financing Adminis- tration. The sample frame contained 20,479 nursing homes. The sample was selected using a stratified two- stage probability design. The first stage was the selection of the individual facilities. A sample of 1,220 nursing and related care homes was selected, of which 1,079 nursing homes (88.4 percent) participated in the survey. The second stage was the selection of current residents and discharges within homes. This stage was carried out by the interviewers at the time of their visits to the facilities. The sample frame for current residents consisted of all people on the register of the facility on the evening prior to the day of the survey. A sample of five or less current residents per facility was selected. The sample frame for discharges consisted of all discharges (whether the resi- dent was alive or dead) that occurred during the 365 days prior to the survey date. A sample of six or fewer dis- charges per facility was selected (5). Data were collected on 5,243 individuals (97 percent response rate) who were current residents in these insti- tutions at the time of contact (Current Resident Sample) and on 6,023 discharges (95 percent response rate) that occurred during the 365 days preceding the date of con- tact (Discharged Resident Sample). Detailed information was collected from nursing home facility records and appropriate staff regarding dependence in activities of daily living; functional impairments; diagnoses; the receipt of services; cognitive and emotional status; sources of payments; and hospital use and nursing home stays prior to, during, and subsequent to the sampled stay. The data collection instruments were the Current Resident Ques- tionnaire (CRQ) and the Discharged Resident Question- naire (DRQ). One of the objectives of the NNHS was the collection of information on patterns of nursing home use. However, the sample design of the NNHS is not compatible with this objective. First, information was collected on discharge events rather than on discharged residents. Since an individual could have been admitted and discharged 2 several times during the course of a 365-day period and since each discharge is listed separately, individuals can appear on the sampling list multiple times. It is also possible for a current resident to be included in the discharge sampling frame if he or she was discharged during the 365 days prior to the survey and then readmit- ted to the sample facility. Thus, an individual might be selected into the sample(s) more than once. The second problem with the sample design concerns how endpoints of a nursing home stay are defined. Stay endpoints are identified by the facility and there is varia- tion across facilities as to how stays are defined, particu- larly in how temporary transfers to hospitals are treated. Some nursing homes consider any transfer to the hospital to be a formal discharge and the resident’s return to be a formal readmission as illustrated by facility A in figure 1. In other facilities, as illustrated by facility B, no formal discharge is made and the nursing home stay includes the hospital stay. These administrative rules concerning of- ficial discharge practices determine the content of the discharged resident sampling list and the admission date associated with the listed stays in both the current resident and discharge samples. Thus, while the two examples in the figure represent equivalent episodes of nursing home use, the manner in which they would be listed and subse- quently sampled would lead to length of stay calculations that would not lead to this conclusion. The discharge sampling list would contain one listing for facility B but two for facility A. Selection of either of the two stays in facility A would yield data on only part of the nursing home stay, and the length of stay calculated from the recorded admission and discharge dates would underesti- mate the true length of stay. In order to calculate length of stay in a consistent fashion, it is necessary to develop a uniform set of rules for the definition of a nursing home stay in terms of the appropriate admission and discharge dates. H - Hospitalization HM - NH discharge A - NH admission Facility A: Ail] H heeeeeeeee———{l ES SOURCE: National Center for Health Statistics, Division of Analysis Figure 1. Illustration of variation in the definition of nursing home stays Additional questions were added to both the CRQ and DRQ to obtain information on nursing home utiliza- tion that would allow the analyst to control the effects of facility variation in recordkeeping. The purpose of this report is to describe how the information on nursing home admissions and discharges collected on the Current and Discharged Resident Questionnaires can be used to rede- fine the endpoints of nursing home stays. As has been found in many other surveys, which collect retrospective information on dates of occurrence, the data on additional nursing home use contained considerable missing and inconsistent entries. Thus, the data first had to undergo extensive editing. These editing procedures are described below. This report describes how these editing and redef- inition procedures affect the data themselves. The results presented are unweighted and are not intended to be used as national estimates. Procedures for redefinition of admission and discharge dates of the sampled stay The first step in the redefinition process was the creation of a person based file where each record con- tained the data for a single resident. Multiple records of the same individual were sorted chronologically and only the earliest record was retained in the file. No data are lost as a result of this procedure since each record should contain all information on a resident’s nursing home and hospital use. This resulted in files which contain 5,200 current resident cases and 5,981 discharged resident cases. The procedures for redefinition of admission and discharge dates were developed with the goal of providing a uniform approach to identifying the endpoints of each nursing home stay. A nursing home stay is defined as the time spent in the nursing home inclusive of time spent in an acute care hospital if the resident returned to the nursing home after the hospital stay without spending any time in the community or in another nursing home. The questions added to the CRQ and DRQ to help make this determination are listed in figure 2. Information was obtained on the dates of other stays the resident had in that facility. Space was available to code up to eight pairs of additional admission and discharge dates. For current residents these stay pairs occurred prior to the sampled admission date. However, for the discharged resident sample, the stay pairs can be any combination of stays prior to the sampled admission and subsequent to the sampled discharge. Information was also obtained on whether the discharge was to a short-stay or general hospital. When the stay just prior to the sampled stay ended with a discharge to a hospital, additional informa- tion was collected in the CRQ on the number of nights the resident spent in the hospital during that stay. For this one interval only, it is possible to determine whether the interval between the two nursing home stays was spent For up to 8 stays: ® On what other dates was resident admitted to and discharged from this facility? ® Was this discharge to a short-stay or general hospital? For first stay only: © Number of nights resident spent in hospital? Figure 2. Questions added to the 1985 National Nursing Home Survey to obtain nursing home and hospital utilization history 4 Total cases 11,181 No additional dates Had additional dates 2,871 a A A 0 20 40 60 80 100 Percent of cases SOURCE: National Center for Health Statistics, Division of Analysis Figure 3. Number and percent of cases in the combined CRQ/DRQ sample with and without information on additional nursing home stays exclusively in a hospital. For the other stays, information on the duration of the intervening hospital stay was not collected. The number of cases for which information on addi- tional stays in the nursing home was reported is given in figure 3. No additional stays are listed for 74 percent of the cases (8,310 residents). Thus, redefinition of admission and discharge dates to reflect the true parameters of the nursing home stay might be required for 2,871 cases or 26 percent of the combined CRQ and DRQ samples. Specifically, the admission date for the sampled stay was to be redefined if 1) the resident had previously been in that facility, 2) the resident had been discharged to a hospital at the end of the previous stay, and 3) the entire interval between discharge and the admission date of the sampled stay had been spent in the hospital. Similarly the discharge date for the sampled stay was to be redefined if 1) the resident had been discharged to a hospital at the end of the sampled stay, 2) the resident returned to the nursing facility, and 3) the entire interval between nursing home discharge and subsequent readmission had been spent in the hospital. Because information on place of discharge was collected for all reported stays in the nursing home, but information on the number of days spent in the hospital was only obtained for the most recent hospitalization immediately prior to the sampled stay, the following rule was developed to determine which intervals should be considered to have included only a hospital stay. Intervals of 21 days or less were treated as exclusive hospital stays while intervals of more than 21 days were treated as having included some time spent in the commu- nity. The goal was to preclude redefinition if it appeared the resident spent any time in the community. The 21-day cutoff rule was developed by analyzing the hospitalization that occurred prior to the sampled stay where number of nights spent in the hospital was ob- tained. Ninety-nine percent of the intervals that were 21 days or less were spent exclusively in short-stay or general hospitals. Conversely, only 38 percent of the intervals that were greater than 21 days were spent exclusively in short- stay or general hospitals (table A). Table A. Number and percent of intervals spent entirely in hospital by interval length Nature of interval Interval length Hospital stay Mixed Number PIayS OriBSs vs nan swims us in 472 1 Morethan21days ............ 44 72 Percent 27 days Oriess ...... Su tii volun 99.8 0.2 Morethan 21 days . . .v.v ov uv vn 37.9 62.1 If the following three conditions are met: 1) the resident had a previous stay in the home, 2) the stay ended with a discharge to a hospital, and 3) the interval between the discharge and subsequent readmission is 21 days or less, the new admission date is set to the admission date for the previous stay. The process continues with the investigation of each preceding stay until a true discharge is encountered. Similarly, if the following three conditions are met: 1) the resident had a subsequent stay in the home, 2) the sampled stay ended with a discharge to a hospital, and 3) the interval between the discharge and subsequent readmission is 21 days or less, the new dis- charge date is set to the discharge date for the subsequent stay. The process continues with the investigation of each subsequent stay until a true break is encountered. A schematic representation of the redefinition process is presented in figure 4. In the first scenario, the admis- sion date is not changed because the previous stay did not end with a discharge to a hospital. In the second and third scenarios, the admission date is set to the admission date of the previous stay because conditions for redefinition are met. In these examples, the admission date is not set to an even earlier admission date because conditions for redef- inition are not met in the second interval preceding the sampled stay. In the fourth scenario the admission date is set to an earlier date that includes two previous stays because conditions for redefinition are met in both inter- vals. The sampled discharge dates were redefined in an analogous manner in the forward direction. H - Hospitalization @ - Current resident HM - NH discharge A - NH admission a A - Redefined admission . . . . . . . . ' ne J S—— 1 . . * ' yy ' : A Ek —8 H k -—@ 2 r ta A—HN, Hjp——" H p— 3 A . : , : A “ll H A ll H A ® 4 v 21 . 21 ' Sampled stay + days + days SOURCE: National Center for Health Statistics, Division of Analysis Figure 4. Schematic representation of the sampled admission/ discharge date redefintion procedure Procedures for editing date inconsistencies The information on the dates of additional stays contained errors and omissions, requiring extensive edit- ing to resolve the inconsistencies. Of the 2,871 cases which contained information on other stays in the nursing home, 1,084 (37.8 percent) had date inconsistencies that re- quired editing. For purposes of this discussion the entire set of eight additional admission and discharge dates is referred to as the “date array” and any one admission and discharge date pair is referred to as a “stay pair.” The report of any date information was considered documentation of a nursing home stay. The sampled admission date, sampled discharge date (DRQ only), and interview date were taken to be correct and the date array was evaluated in relation to these dates. Information from the entire date array was used to reconstruct stay pair information. Stay pairs were deleted only as a last resort and admission or discharge dates were not moved across stay pairs. Dates were instead corrected or, if necessary, imputed whenever possible. When two or more solutions were available for reconstructing a consistent date array, we chose the one that changed the least number of dates in the existing date array. An effort was made to be conservative in the reconstruction of date arrays with respect to the 21-day cutoff that would determine eligibil- ity for date redefinition. Finally, the same mechanism was used to resolve all cases with the same type of inconsistency. Date arrays were reviewed, errors detected, and edit rules were applied to resolve errors. A consistent date array is defined as one with the earliest admission and discharge pair in the first position, with subsequent stay pairs following in chronological order, with missing date codes (i.e., “989898”) filling the remainder of the array, and with date intervals that are mutually exclusive, i.e., no two stay pairs have overlapping or embedded date infor- mation. In addition, for current residents, all dates listed must fall before the sampled admission date. For dis- charged residents, the dates can fall before the sampled admission date or between the sampled discharge date and the interview date. All records were reviewed using computer edit pro- grams and manual inspection to determine whether they met the consistency requirements. Corrections made to inconsistent date arrays can be classified into three groups: 1) re-sorting of dates, 2) changing dates, and 3) imputing missing dates. 6 Re-sorting dates The simplest errors were those where the stay pairs were out of chronological order. The entire order could be reversed or individual stay pairs could be out of chrono- logical order. Occasionally admission and discharge dates within a stay pair were transposed. These types of errors were corrected first and involved reordering the stays. Changing inconsistent dates After the date array was sorted into proper sequence, the records with remaining problems were reviewed to determine whether, in the judgment of the authors, they appeared to have transcription or keying errors. Errors of this type were corrected. Errors that did not appear to be the result of transcription or keying errors required fur- ther attention. For example, stay pairs embedded within other pairs were deleted and the longer stay was main- tained on the file. Overlapping stay pairs were combined into one longer stay covering the entire time period. Stay pairs occurring after the sampled admission date for current residents, which could not be corrected through one of the above rules, were deleted. Finally, discharge dates for the most recent stay for discharged residents, which indicated that the subject was still a resident in the facility at the time of the field interview, were checked against question 11d, “Is still a resident [of the sample facility]? and question lle, “Was dis- charged alive?” When the answers to these questions indicated that the resident had been discharged, the discharge date was corrected either by imputing a dis- charge date (see section on imputation below) or by inserting the date of death depending on the information provided in question 11d and question 1le. Imputing missing dates Date imputation rules were developed for cases in which one or more dates in the date array contained missing information. Two types of missing date informa- tion occurred. In the first type, the missing dates are bracketed on either side by valid dates. In this case the rule employed is to impute the missing date to the mid- point of the interval defined by the valid dates. For example, in the following date array, (010184 989898) (010185 020185) (021585 030785) (031185 031985) (989898 989898) (989898 989898) (989898 989898) (989898 989898), the first pair’s discharge day, month, and year were imputed to be the midpoint between the first and second admission dates (070184). In the second type, missing dates could be at one or both ends of the valid subset of the date array. In this case the rule employed was to delete the stay pair. For exam- ple, consider a record with the following date array: (989898 121584) (010185 020185) (021585 030785) (031185 031985) (032185 032585) (032785 033185) (041085 043085) (989898 989898), where the sample ad- mission date is 091085, the sample discharge date is 111585 and the interview date is 121785. Since this first missing admission is at the beginning of the series and before the sampled admission date, there is no possibility of using the information in the rest of the series to determine the range within which the date might fall. In this case no midpoint imputation is possible. The decision to delete such stays, thus not considering them in the redefinition, is the most conservative alternative since it results in the shortest length of stay. The figures and tables that follow in the report are based on this decision. An alternative method for dealing with these missing dates would be to set the missing date to the date preceding or following the valid discharge or admission date in the stay pair. In order to preserve the first stay pair in the example above, the missing admission date would be set to “121484,” the day preceding the existing dis- charge date. The choice of a single day represents a conservative limit on the length of the newly preserved stay, and this imputation method would also likely under- estimate the true length of stay. If this second approach were employed, the results of the redefinition process would be slightly different. Fifty-six additional cases would have their admission dates redefined, 40 current residents and 16 discharged residents, increasing the total number of redefined admission dates specified in table C, page 8, from 1,434 to 1,490. The percent of total cases undergoing redefinition of admission and/or discharge dates would increase slightly from 17.6 percent to 18.1 percent result- ing in only minor changes in figure 5, page 8. Missing discharge dates at the end of the series would be handled in an analogous manner. However, there were no cases of this type where the imputed date would have become the redefined discharge date. Thus, there would be no differ- ence in the results of the redefinition for discharge dates between the two methods of handling missing end dates. Of the 16 additional discharged resident cases where the alternative imputation method resulted in redefinition of the admission date, one had a redefined discharge date using another redefinition rule. Therefore, the total num- ber of discharged resident cases with any redefinition of dates increases from 1,319 to 1,334 in table E, page 10. The median length of stay for these cases would change minimally. It should be noted that cases often had more than one of the types of errors described above. These cases were handled using the most appropriate combination of rules. The complexity of these cases was such that writing computer algorithms to effect the rule logic was impracti- cal and most of the cases were scrutinized individually and corrected by hand. The types of inconsistencies found among the 1,084 cases are shown in table B. In 129 cases (11.9 percent) the only correction required was the reordering of the date array. In 773 cases (71.3 percent), changes in one or more dates were required in addition to any possible reordering. Finally, in 182 cases (16.8 percent), the cor- rections included imputation of one or more dates and may also have included reordering or date correction. Table B. Number and percent distribution of cases by type of date corrections Percent Correction Number distribution Total. vsme rime sms mene emmy 1,084 100.0 Re-SOMING ONY «iv cvs wv nimis wus 129 119 Datecorrection . ....:vs:2 104 773 71.3 Date imputation. . . ........... 182 16.8 Results of date redefinition Figure 5 shows the results of the date redefinition procedure. As noted, 74 percent of cases did not contain information on additional stays in the nursing home and, therefore, were not candidates for redefinition. Eighteen percent of the total cases, but over two thirds of those eligible for redefinition, had their sampled admission or discharge dates redefined. The remaining 8 percent had additional dates but the dates for the sampled stay were not redefined. The admission date was redefined in 648 (68 percent) of the CRQ cases eligible for redefinition. Of the 1,920 DRQ cases eligible for redefinition, only the admission date was redefined in 572 (30 percent) cases, only the discharge date was redefined in 533 (28 percent) cases, and both endpoints were redefined in 214 (11 per- cent) cases. In total, 1,434 cases had their admission date redefined and 747 had their discharge date redefined (tables C and D). No additional stays PL RON) Date redefinition SOURCE: National Center for Health Statistics, Division of Analysis Figure 5. Number and percent of cases in the combined CRQ/DRQ sample by whether dates were redefined Table C. Number and percent distribution of eligible cases by redefinition status of admission dates Percent Redefinition status Number distribution Total eligible +. + uo wp vm ow soma’ 2,871 100.0 Admission date redefined. . ...... 1,434 49.9 Admission date not redefined . . . .. 1,437 50.1 Table D. Number and percent distribution of eligible cases by redefinition status of discharge dates Percent Redefintion status Number distribution Total SlQIBIG . is suis wa smsms sme 1,920 100.0 Discharge date redefined . . ...... 747 38.9 Discharge date not redefined . . ... 1,173 61.1 Figure 6 shows the magnitude of the changes made to the admission dates for those current residents and dis- charged residents in which the sampled admission date was redefined. In 56 percent of the cases, the admission date was redefined as having occurred at least 1 year earlier than the recorded admission date. In nearly 20 percent of the cases the redefined date is more than 4 years earlier than the recorded date. Thus, redefinition tends to make a large difference in the date for a majority of the cases where the admission date was redefined. The impact of redefinition on the discharge date for those discharged residents undergoing a change in dis- charge date is shown in figure 7. In order to be selected into the discharge sample, the discharge had to have occurred within the 365 days prior to the field interview. As a result, discharge dates could only be brought forward by a maximum of 365 days. The difference between the redefined and recorded discharge date is more than 3 months in 56 percent of the cases. In about one third of the cases this difference is 6 months or more. Thus, even when the range of time difference is restricted to less than 1 year, redefinition makes marked differences in the sampled discharge date. The impact of redefining admission and discharge dates is illustrated by comparing the distribution of length of stay based on recorded admission and discharge dates 25 r 20 15 - 8 2 & 10 5 0 Less than 3-6 months 6-9 months 3 months SOURCE: National Center for Health Statistics, Division of Analysis 19.3 9-12 months 3-4 years 4 years or more 1-2 years 2-3 years Figure 6. Percent of cases in the combined CRQ/DRQ sample by difference between sampled and redefined admission dates (N=1,434) Less than 3 months 3-6 months 6-9 months 9-12 months SOURCE: National Center for Health Statistics, Division of Analysis Figure 7. Percent of cases in the DRQ sample by difference between sampled and redefined discharge dates (N=747) to length of stay based on the redefined dates (table E). Using the original 5,981 discharged residents, the median length of stay increases 58 percent from 135 days when based on the recorded length of stay to 213 days when based on redefined length of stay. Focusing on only those 1,319 discharged cases where dates were redefined, the median length of stay increases from 134 days when based on recorded endpoints to 562 days when based on rede- fined endpoints, an increase of 419 percent. In addition, about 55 percent of stays are less than 6 months long when the length of stay is based on recorded dates for both the total group and those cases undergoing redefini- tion. When redefined dates are used, 47 percent of the total sample of discharges have a length of stay less than 6 months and only 22 percent of the subset of redefined stays have stays under 6 months. Similarly, 34 percent of the 5,981 sample stays have a length of 1 year or more when using recorded dates, but 41 percent of discharge cases have stays 1 year or more in length when redefined dates are used. Of the 1,319 redefined cases, 30 percent of the discharges have recorded lengths of stay of 1 year or more whereas almost 62 percent of the redefined dis- charges have recalculated lengths of 1 year or more. Furthermore, 37 percent (494) of the 1,319 dis- charged residents with redefined endpoints, had their discharge date set to the interview date since the redefini- tion process determined that they were still in the nursing home at the time of the interview. Subsequent waves of data collection in the National Nursing Home Survey Followup will extend the discharge date estimates for these residents as well as for current residents and there- fore the statistics presented above must be considered conservative bounds on the differences between sampled and redefined length of stay. Table E. Number and percent distribution of Discharged Resident Questionnaire (DRQ) cases by length of stay (LOS) for sampled and redefined admission and discharge dates Sampled LOS Redefined LOS Percent Percent Length of stay Number distribution Number distribution All DRQ cases TOR wo 5 50 wat 3 0 3 oe 5 30703 0 ee EE 0 BEE BG 5,981 100.0 5,981 100.0 Less than I MOMNS osc snr rr rmasvsws snamenanmes 2,662 44.4 2,249 37.6 BeBOTIIG vo rv nine 18 hie tie Bos Bi vias Hi ES wars ima 621 10.4 590 92.9 C-OMOMNS . ..cnvvnvns vrs cna me Ea 406 6.8 412 6.9 O-=TZ MONA 4 cvs vms wn smome amas dWaME §HE HE 8 269 4.5 294 4.9 Y=2WBAIB ox ves mn vw HWE ERTS EMRE ERE 643 10.7 736 12.3 Rr-BYBBIB xu 2 vou niais ai pi TR wh Swe ene dn bb aE eb 392 6.6 441 7.3 BA VYBAIS . ws 5 IGANE «MEPS SHE PR Zhe WE nw ww 225 3.8 306 5.1 MOT INBN AYBAIS ios vse co smmemsmarame wih va emms 763 12.8 953 16.0 Median (08yY8) «ilu vss mvs mimeibis male mem hues 135 213 DRQ cases undergoing redefinition TOURL ov vn 0 grommongims oe ig ch 2 SEER BREE 8 v8 FLEES 4 igs ress ro 8 1,319 100.0 1,319 100.0 Lessthan 3 mONNS. . , . cuca r rss nrbns $a ens wrens 571 43.3 158 12.0 BB OMIYS 420 vy we ww ws vm te Sw AE RE RRS 171 12.9 140 10.6 B-DINOMNG vc 14 coh 51% wi ew vai m9 At 0 che 8 107 8.1 113 8.5 S-12H0GING » v0 ts £UARE Bll b Haw a 73 55 98 7.4 T=BYBAIB . v vvvv ran bas ERR SEES hn mw 154 11.7 247 18.7 DB YORI, i dv ows Hawg RLS 5 ATR i TE WS Ee 94 74 143 10.9 Brel YAS «win v +10 1 5 wins wi 0 0% Wie 9% BSE ER ER BE 45 3.5 126 9.6 Mora than'a YBars . . «dy «ivi w wo iv sew viv vias wmv os 80 ol 8 104 7.9 294 22.3 MEGIEN{CEYE) "1 15 cmv dd sim adn 48 208 4 0s sor in wi din oe 134 562 Effects of data editing on date redefinition Because the procedures used to edit the date arrays could affect redefinition and length of stay computations, the edit rules were designed to err in the direction of not redefining dates and therefore limiting length of stay. Table F shows the percent of edited and unedited cases undergoing redefinition of the sampled admission dates for the combined current resident and discharged resident sample. Thirty-one percent of the edited cases were Table F. Number and percent distribution of cases eligible for redefinition by redefinition status of admission date and edit status of date arrays Edit status of date array Redefinition status Not of admission date Cases Edited edited Number THA: sare 2 mems sms 2,871 1,084 1,787 Redefined . . ....... 1,434 341 1,093 Not redefined. . . . . .. 1,437 743 694 Percent distribution Total: + sme vs nims nus 100.0 100.0 100.0 Redefined . ........ 49.9 31.5 61.2 Not redefined . . . .. .. 50.1 68.5 38.8 redefined while 61 percent of the unedited cases were redefined. Thus, for the cases where inconsistent informa- tion was found and corrected, the percent of those cases redefined is half of that for the cases that were correct to start. In addition, if redefinition occurred, the change in dates was more limited. As shown in figure 8, the admis- sion dates for edited cases were more likely to change by less than 3 months, 3-6 months, and 6-9 months than are the unedited cases. Similar results were found for discharge dates. Table G shows the percent of edited and unedited cases undergoing redefinition of sampled discharge dates. Eleven percent of the edited cases were redefined while 47 percent of the unedited cases were redefined. Again, if redefinition occurred, the change in discharge dates was more limited for edited cases. As indicated in figure 9, the discharge dates for edited cases were more likely to change by less than 3 months (78 percent versus 42 percent). Combining the results found for admission and dis- charge dates, table H shows that the length of stay was changed in 30 percent of the edited cases while the length of stay was affected in almost 80 percent of the unedited be! J SERRIHRRKY 15 o> CO bo 5 Poledele RERREE 2 ore , XC) XR Percent ) 2 4, 2X2 QQ S 2 ON XD ” XX % B55 3 10 Ke ON (X20 S505 CD 0 ®, 2 XX, 5 55 25 KK 505% OC ve 0; Q > 9 x0 % % ) 5 5 0 5 & 9, o, QQ () % 0 3 3-6 months o Ko Less than 3 months SOURCE: National Center for Health Statistics, Division of Analysis 9-12 months Bl coed BB Not edited RREXAXRL 2 LX XXXL C50 0.0. 9.0.0 0, XX QQ 5 % % % SXXRRR 0: oe 5% &% CX 2 250505 Q RR SO (> 2 {XX 30585 CQ 2 & 3 KK 5 SKK & 0. > Q 2 2 > 8% SKK RRS RRS &% % ) 30 3 (Q QQ 2X2 . CQ KK XXX 8 55 3S 2 2 2 SX % RRS & X52 CD) 0, 0 K) XX (XD CHAS CAR RR RR CX 250% 9 9 2 2 XD 2 2 > QQ KR) XX 8 KL CIS OC) JX % RRS RRS > 2 C0 2 SL > & % & COC & SRS 5% X &% & RA ERK LER 82588 255 Dotetele! SE OO 5 & 5 9 BS 2% 8 CO) PASS RE 1-2 years 2-3 years 3-4 years 4 years or more Figure 8. Difference between sampled and redefined admission dates by whether cases were edited, CRQ and DRQ (N=1,434) 11 Table G. Number and percent distribution of discharged resident questionnaire cases eligible for redefinition by redefinition status of discharge date and edit status of date arrays Table H. Number and percent distribution of discharged resident questionnaire cases eligible for redefinition by redefinition status of length of stay (LOS) and edit status of date arrays Edit status of date array Edit status of date array Redefinition status of Not Not discharge date Cases Edited edited Redefinition status of LOS Cases Edited edited Number Number VOB in msi wm em 3 1,920 423 1,497 Total co: 0s cnimn eu 1,920 423 1,497 Redefined . . ....... 747 45 702 Redefined . . ....... 1,319 129 1,190 Not redefined . . . . . .. 1,173 378 795 Not redefined. . ..... 601 294 307 Percent distribution Percent distribution TOM. oom 0 2508 100.0 100.0 100.0 Total. «sums mimenms 100.0 100.0 100.0 Redefined . . ....... 38.9 10.6 46.9 Redefined . . ....... 68.7 305 79.5 Not redefined . . ..... 61.1 89.4 53.1 Not redefined . . . .. .. 31.3 69.5 20.5 larger percents in the next three intervals: 3-6 months, 100 6-9 months, and 9-12 months. The remaining intervals of ee Edited BB Not edited 1 year or more tend to be about even between edited and 80 unedited cases. However, the percents for edited cases are z 60 based on relatively small numbers. The net impact of the 3 date edits is that both the proportion of edited cases € 40 Re) undergoing redefinition as well as the magnitude of the SS increase in length of stay for those edited cases that were 2 RSS 5 RS redefined is less than that for cases that were correct as o LIE 5 ti] me | recorded. Lgss than 3-8 months 6-9 months 9-12 months A specific example of a decision to err conservatively with respect to admission and discharge date redefinition SOURCE! Nebicnal Snpin Kaen Statetca, Delon of Avilysls is the midpoint rule for imputed dates. An examination of Figure 9. Difference between sampled and redefined discharge dates by whether cases were edited, DRQ only (N=747) cases which contained additional dates. Figure 10 shows the difference between sampled and redefined length of stay for edited versus unedited cases. Edited cases have a larger percent in the less than 3-month interval (35 per- cent versus 30 percent), while the unedited cases have date array patterns revealed that in over 93 percent of the cases where a discharge date was missing, the interval between the two successive admission dates was greater than 42 days. Thus, imputation using the midpoint date in these cases kept the separation interval greater than 21 days on either end. The imputed date served primarily to mark the place in the date array and there was no redefinition over the interval. 40 ] % x 8 ele 35 CX 2 2 XS & % 5 5 255 vs Q 2 Q 0. 2 0 9%, 2 2 ) SOX 2552 Q Xs 5% S 5S x0 *. , 5% & 5 5% *, 9, 5 & 8 5 5 CX Q Percent OC 3X5 XL 2% 2 2 0 28 Q 2 2 0 2% 2s CQ ®, 555% B58 R55 OO) ) SRR OX 55 2 28 5% RAD 250 2 > 0 KR 2 2 9 QO OX 55 5 2 RZ oO Q 9 x2 Ps 25 3X OX 2 0 re 22552 255 530% 8 XX) 2 Q XD) 0. 0% 2 CCX XXX > 5 CQ & 10 p= XK RR XK x0 o. 2 0S % 3 5 20 2% 35 & 3 0 5 &% % ERX o SO SOURCE: National Center for Health Statistics, Division of Analysis 9-12 months Bl cote B33] Not edited oot XXX XXX XXX OC) 3353 20 Cd 0038; 55 5% & 2% CO 47 22 > ® 5% XJ 5 5 % % 5% % 3 3 2 Ro 5 KQ 2, CQ 1-2 years 2-3 years Figure 10. Difference between sampled and redefined length of stay by whether cases were edited, DRQ only (N=747) 12 Summary A method has been described whereby episodes of nursing home utilization can be reconstructed using new information collected in the National Nursing Home Sur- vey (NNHS). Additional questions made it possible to combine nursing home stays that were broken only by visits to short-stay hospitals and to improve compatibility of stays reported by facilities that do and do not consider such hospital visits as reasons for formal discharge. Addi- tional stays in the same nursing home were found in one fourth of the cases in the NNHS. However, data on additional stays had high levels of missing and incomplete data. Forty percent of those cases required further editing. As noted, many cases were so complicated that the date fields needed to be evaluated and corrected manually. Editing was performed in such a manner that, when data on additional stays were incomplete, the editing tended to limit further redefinition of recorded admission and discharge dates. When data on additional stays were present, admission and discharge dates were redefined in 70 percent of the cases overall: 56 percent for the edited cases and 77 percent for the unedited cases. Redefinition tended to increase markedly the length of stay for the affected cases and this translates into a change in length of stay estimates for the entire sample. For the entire dis- charged resident file, the median length of stay based on the recorded endpoints was 135 days as compared to 213 days for redefined endpoints. In addition, the 1985 Na- tional Nursing Home Survey Followup will provide more information on this cohort. For many of the current residents as well as for discharged subjects who were still residents of the sample nursing home on the 1985 NNHS interview date, subsequent data collection waves will ex- tend the length of stay and further increase the estimate of median length of stay. 13 References 14 “Meiners MR, Trapnell GR. Long-term care insurance pre- mium estimates for prototype policies. Medical Care 22: 901-911. 1984. Liu K, Manton KG. The length of stay pattern of nursing home admissions. Medical Care 21: 1211-1222. 1983. Spence DA, Wiener JM. Nursing home length of stay patterns: Results from the 1985 National Nursing Home Survey. The Gerontologist 30: 16-20. 1990. Short PF, Cunningham P, Mueller C. Standardizing nursing- home admission dates for short-term hospital stays. Medical Care 29: 97-103. 1991. Hing E, Sekscenski E, Strahan G. The National Nursing Home Survey: 1985 summary for the United States. Na- tional Center for Health Statistics. Vital Health Stat 13(97). 1989. 3% U.S. GOVERNMENT PRINTING OFFICE: 1992 — 3 1 2— 0 82% 60002 New Electronic Data Product Releases National Center for Health Statistics CENTERS FOR DISEASE CONTROL National Hospital Discharge Survey on Diskettes — 1989 Diskettes containing data on hospitalization in the United States during 1989 have recently been produced by the National Center for Health Statistics (NCHS) and are available for sale from the National Technical Information Service. The diskettes, along with documentation, provide automated access to data on hospital utilization by age and sex of patient, geographic region of the United States, diagnosed conditions, and surgical and nonsurgical procedures performed. The data are from NCHS’ National Hospital Discharge Survey (NHDS), an annual survey designed to obtain information from medical records of sampled inpatients. The data diskettes include software for accessing the NHDS database fields, computing rates for selected popu- lations, and creating smaller files for use with other soft- ware packages. The database files were developed by downloading text files similar to the detailed tables in Vital and Health Statistics, Series 13, “Detailed Diagnoses and Procedures for Patients Discharged From Short-Stay Hos- pitals, United States.” The technical characteristics and the software and hard- ware requirements for the diskettes are: Technical Characteristics: ® First-listed Diagnoses e Procedures e Days of Care All-listed Diagnoses (available upon request) Hardware and Software Requirements: A. Hardware ® (40K RAM eo [BM compatible computer B. Software e PC/MS DOS release 2.0 or higher eo Multi-year data access system software (included) eo dBase III files (Diskettes for data year 1989 can be purchased by complet- ing the order form on the back. To order NHDS diskettes for data years 1985-88, contact the National Technical Information Service at the address below.) Mail Order To: NTIS National Technical Information Service U.S. DEPARTMENT OF COMMERCE Springfield, VA 22161 (703) 487-4650-TELEX 89-9405 Purchaser: Telephone Name, Organization Address City, State, Zip, Attention: Method of Payment [C] Charge my NTIS deposit account no. J Purchase order no. [C] Check enclosed for $ J Charge my: American Express Visa MasterCard Card expiration date Account no. Signature (Required to validate order) | Send me an application for an NTIS Deposit account (NTIS-PR-33) Ship to: (Enter if different from address at left) Name Organization Address City, State, Zip Diskettes 1989 Data Files QUANTITY UNIT PRICE TOTAL PRICE 312" PB 91-506881 $80.00 312" (HD) PB 91-506873 $80.00 514" PB 91-506865 $80.00 5v4” (HD) PB 91-506857 $80.00 1989 All-listed Files QUANTITY UNIT PRICE TOTAL PRICE 312” (HD) PB 91-507111 $50.00 5va” PB 91-507087 $50.00 5%” (HD) PB 91-507079 $50.00 Documentation PB 91-204362 $15.00 TOTAL Series 2 No. 116 NR% oigle| Health Statistics From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics Comparability of the Birth Certificate and 1988 Maternal and Infant Health Survey March 1993 = “ - U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES 7 yy Public Health Service . v Centers for Disease Control and Prevention Ta National Center for Health Statistics CENTERS FOR DISEASE CONTROL AND PREVENTION Trade Name Disclaimer The use of trade names is for identification only and does not imply endorsement by the Public Health Service, U.S. Department of Health and Human Services. Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested Citation Schoendorf KC, Parker JD, Batkhan LZ, Kiely JL. Comparability of the birth certificate and 1988 Maternal and Infant Health Survey. National Center for Health Statistics. Vital Health Stat 2(116). 1993. Library of Congress Cataloging-in-Publication Data Comparability of the birth certificate and 1988 Maternal and Infant Health Survey/by Kenneth C. Schoendorf . . . [et al.] p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research; no. 116) (DHHS publication; no. (PHS) 93-1390) ISBN 0-8406-0473-4 1. Pregnancy — United States — Statistics — Evaluation. 2. Childbirth — United States — Statistics — Evaluation. 3. Birth Certificates — United States — Evaluation. 4. National Maternal and Infant Health Survey (U.S.) I. Schoendorf, Kenneth C. II. Series. ll. Series: DHHS publication; no. (PHS) 93-1390. [DNLM: 1. National Maternal and Infant Health Survey (U.S.) 2. Birth Certificates. 3. Health Surveys — United States. 4. Questionnaires. W2 A N148vb no. 116] RA409.U45 no. 116 [RG530.3.U] 362.1'0723 s—dc20 [304.6'3'0973021] DNLM/DLC 92-49066 for Library of Congress CIP Vital and Health Statistics Comparability of the Birth Certificate and 1988 Maternal ana PUBLIC HEALTH Lisa Infant Health Survey NOY 03 1993 INCE y - UNIVERSITY OF CAI IE Rep VI UALIF. BERKELEY Series 2: Data Evaluation and Methods Research No. 116 This report compares responses to the 1988 National Maternal and Infant Health Survey maternal questionnaire to similar items from the birth certificate, including demographic factors and items pertaining to the current and past pregnancies. TI Wiss, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics Hyattsville, Maryland March 1993 DHHS Publication No. (PHS) 93-1390 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jack R. Anderson, Acting Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Office of Analysis and Epidemiology Jacob J. Feldman, Ph.D., Associate Director Division of Analysis Jennifer H. Madans, Ph.D., Acting Director Jennifer H. Madans, Ph.D., Deputy Director Jennifer H. Madans, Ph.D., Acting Chief, Demographic Analysis Staff Diane M. Makuc, Dr.P.H., Chief, Analytical Coordination Branch Sandra T. Rothwell, Chief, Longitudinal Analysis Branch Contents DE EO CITC ENO rs rrr srw v0 or 2 po 0 es ae no ss i robs: Ap Be Ep Fo ay aE hp 1 oe Were Bh BIR rei Bem ae EWR Dat SOUTH AE ITIITEATIOIIS wow: «vr irins rome irigiee ise a woes tase sintop es pi or ws e750. x Po secs ftp ma Fehon i: i are 40» 0 ts el © 00h do 3 Vs et DECGIATIIC TIERRTITES ons» ore ssinrsas ose oe mics oer 1 0 ko mt i oH Tse ge hcg gh i fr Ep ge og fief A Race and Hispanic origin of mother and father. . «cc cows vuneiocnmmse snipe sow ns pews soms ss amas wae ss baie ess NAIVIEY OF TOOT 50s 05m mams moms wim mia ws oa bis m0 me 20009 RES 00500 0 1GR 205010 000 HER A0 E OAR 8H B08 0008 HR # ae ALE Of MONET BN RHEL. . os cain 5 0m Bons n® 308 5 RS 5 RR RE Fk 08 IE FSFE Wiig HE A 0 0 0 HW 4 Education. of mother GRE IBTREE . «sis ovens im steis nies iene wie siete se aise wenden te abate eo oe 6 41s W080 ERD site Sebraie Marital SEALS... vos urs isn mnie eis irs a enn sella es eT i ri EU ee Gt 0 ERNE RSE BIE td Pregnancy HIStory IVSHSUICE www ove uis wun ms wisi ms mime mie ios i homens rs vw S00 » pls 12 of meme sobs 5 lt 18 8 108 Wh0 kahve md at dre of Live DIrth OPAET «0s ve + mses mmm uno om se 2 001% 9 000 0 £5000 10.6 HRAS 0 SERIE HALE T8 16 H40 55 0 909108 Rie 9 010 (Rw 0 i 0 8 i 8 HE 8 Prior fetal dRath8. oso vimim mans was mam 518 0 ERARE ABE 030 HE RRA SE 0000 REE 5100 F FRE TED FEA AE HS HER HR 2 Pregnancy MICASUTES . vo seams samme GR 58 000 01 FREE RHE GRE 50020 585 RES 0 00 58 6 RE BS 3 40 018 RAK 82 0510 810 5 408 48H a0 0550 08 5 PLUBATILY, 05 nn rotons sein mew rmnts cd # S 18 h B 80 3 1 ol FR, S85 4 HR J G8 (5 AT isc 9 ¥ 00 0 es ik Be AE Timing 0F prenatal Care MTEC . «vs sacs seems soa wan 55s wes Hae EERSTE £080 RLS HE REE TSHIRTS Hen T0850 104 90 0 Number of prenatal COE VASE. ; sv ws munis wus a 5855 5 90 W5 £ 54318 5 58 TAKES 84018 8 48 56 31010 018 10 $10 90 405 919 96 48 92 80 0 GEStRHONGL BLL... «5.4 sissies vi 2m vie ie 00h 0s 0 050 0 8 900K 918 5 Fk B13 0 Ri B13 5 uf BE 10 Bh HER Eh 0 08 Hk 900 if LAR Bi 0 Bk 32 90 B04 0B 00 TDISCUSSION « o5 0505 0 500 70 0.0 41 7 46 00000 0 50 0 0s 8 OF 06 J 90.0 60 Ba 5% io 00 90 0 Sr 50 0% 9 0 oo 0 mh wil 20 8 iw 0 5 0 BE (wi 0 i RO CTONCER - vn ves 7 vw 0s 8 re 2 0 id 0% 0S ETE 0 8 oF OE WE TUE BE NUE 0 EE WE EE 0 BE Be A FE TE HR 4 BE SR 1 8 ee of LiSt Of ASAI] TABI. + «5 vw m1 3 70 30:0 mis op 30 0000 08 000 30 4 0 9 0 0 0 06 0 0 7 3 0 WF 0.6 00 9 0 18 0 0% 0 0% URE WO 5 0 010 10 oe 0 0 List of text tables A. Number and rate of respondents and nonrespondents for the maternal questionnaire, by category of respondent and race of mother: National Maternal and Infant Health Survey, 1988 ......................... B. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for race of parents, by race: Tialicd Sates, TIE . io isi: wim tim sins a1 ims gm us 0 000m oh 0 0 003 4050 0 8 40008 10 1 0.0 0 gf 0 4 0 8 0 06 04 C. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for Hispanic origin of parents, by Hispanic origin: United States, 1088 vc urvvsnsmms mia ninns nt mnt sean na enins sons emens D. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for nativity of mother, by race and nativity: United States, 1988 ...viiiavensinsnss omen sme sns bos es assmine staves E. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for age of parents, by race and age: United Sates, 1988 . virus impr mru umm mamsmes as @wes sess ns masss sss smth F. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for education of parents, by race and education: United States, 1988 ... titi G. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for marital status of parents, by race and marital status: Linited States, 1988 ....coinivrrvios sminsinrsmenisnsnsmsomonvintms H. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for live birth order, by race and live birth order; United States, 1988 «viv vinsmnimresiniquistsnsnnmtanensmyinsmus mans: J. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for prior fetal deaths, by race and number of fetal deaths: United Sales, 1988 ..civonusvrmsrimnmaismnons nto msn ssn sm asmsions K. L. M. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for plurality, by race and plurality; United States, 1988 . ...ooceenanas murs sams ss smns ens snsss snes ssesssnssssdsnn eases Percent of responses from the mother’s questionnaire in agreement with the birth certificate for trimester of first prenatal care, by race and trimester: United States, 19088 ........ coo ie Percent of responses from the mother’s questionnaire in agreement with the birth certificate for number of prenatal care visits, by race and number of visits: United States, 1988 cuuninssrssrsssmrinsmsan vamssass snes Percent of responses from the mother’s questionnaire in agreement with the birth certificate for length of pregnancy, by race and gestational age: United States, 1988... .. iii eee Symbols --- Data not available . Category not applicable - Quantity zero 0.0 Quantity more than zero but less than 0.5 Z Quantity more than zero but less than 500 where numbers are rounded to thousands * Figure does not meet standard of reliability or precision (estimate is based on fewer than 20 births in numerator or denominator) Comparability of the birth certificate and 1988 Maternal and Infant Health Survey by Kenneth C. Schoendorf, M.D., M.P.H., Jennifer D. Parker, Ph.D., Leonid Z. Batkhan, Ph.D., John L. Kiely, Ph.D., Division of Analysis Introduction Information from birth certificates is widely utilized in the United States. Among the major uses of birth certifi- cate data are annual statistical tabulations and research in areas concerning maternal and child health. Since birth certificate data are so commonly used, it is important to examine the validity of the birth certificate information. One method of assessing the quality of birth certificate data is to compare the information on the birth certificate with an independent source of information from the same birth. This was done in 1985, when survey data from the 1980 National Natality Survey (NNS) was compared with NOTE: We are grateful to Mary Glenn Fowler, M.D., M.P.H., and Robert L. Heuser for their reviews of this report. corresponding birth certificate data for a sample of births that occurred in the United States (1). Although that comparison provided much useful information, its gener- alizability may be limited because the maternal survey portion of the NNS included only married mothers. The 1988 National Maternal and Infant Health Survey (NMIHS) allows for the assessment of birth certificate data for a broader cross-section of the population than the NNS allowed. The purpose of this report is to compare responses from the National Maternal and Infant Health Survey with birth certificate information for items com- mon to both sources of data. Data source and limitations The 1988 NMIHS was conducted by the National Center for Health Statistics to examine factors concerning maternal health, pregnancy outcome, and infant health. The NMIHS collected information on independent sam- ples of live births, fetal deaths, and infant deaths that occurred in the United States in 1988 (2). The analyses in this report focus only on information from the live birth sample. Data for each infant in the NMIHS live birth sample were derived from four different sources: A questionnaire completed by the mother anywhere from 6-30 months after the birth of the child; the mother’s prenatal care provider(s); the hospital where the infant was born; and the infant’s birth certificate. This report compares infor- mation from the maternal questionnaire only with infor- mation from the birth certificate. Future reports will include data from the prenatal care provider and hospital portions of the NMIHS. The live birth sample contains an over-representation of low-birth-weight infants (infants with a birth weight of less than 2,500 grams) to allow for detailed analysis of factors associated with prematurity and growth retarda- tion. Additionally, because black women in the United States have a high risk of adverse pregnancy outcome, the live birth cohort contains an oversampling of black infants. The NMIHS was made nationally representative by the calculation of a sample weight for each record that accounts for the survey’s sampling scheme and for survey nonre- sponse. In this report, the sample weights were not uti- lized in the calculation of comparability rates to permit the reporting of the actual numbers upon which the comparisons were based. No substantial differences were found when comparability rates with the sample weights were compared with comparability rates calculated with- out the sample weights. Of the 13,417 mothers that were contacted for the survey, 9,953 responded, yielding an overall response rate of 74 percent. This report is limited to the 4,956 black mothers and 4,695 white mothers that responded to the survey, because there were too few mothers of other races to allow for meaningful comparisons among those groups. Table A shows the total number of sampled live births for white and black mothers, along with response rates and reasons for nonresponse. Because pregnancy characteris- tics and outcomes differ by race, all comparisons, with the exception of Hispanic origin, are reported separately for black and white persons. For those comparisons, maternal race on the birth certificate was used to determine race. For some survey respondents, information for individ- ual survey questions or birth certificate items was missing. Comparability rates between corresponding items from the maternal survey and the birth certificate are based only on records that contain valid responses for both items. The number of missing observations for each item is Table A. Number and rate of respondents and nonrespondents for the maternal questionnaire, by category of respondent and race of mother: National Maternal and Infant Health Survey, 1988 White! Black Category of respondent Number Rate? Number Rate? Total number of births sampled. . . ................. ,947 sila 7,055 faa RESPONUBALS . + « . vv vos mnmn sma sm smnms smn mm nme 4,695 78.9 4,956 70.2 Nonrespondents: Unabletolocatemother. . . ...............cc... 544 9.1 1,119 15.9 RelUSEAIBUIVEY . « ; + vs ws wn ims an smi dm mins 2s 250 4.2 254 3.6 Could not contact mMOher. . . .: ss mews ne sms 2% ams 124 21 320 4.5 Certificate excluded by State. . . ................. 200 3.4 195 2.8 Nonresident of the United States . . . .............. 43 0.7 31 0.4 Mother gave baby for adoption . . ................ 22 0.4 32 0.5 Motherdeceased . . ... c«cssws smsmas snsws smwams 5 0.1 21 0.3 Mother claims no pregnancy. « . «+ cvs va cw own smems 16 0.3 18 0.3 Other noninterview . . . . ... «viii 48 0.8 109 1.5 1As defined by the birth certificate. 2The number of records in each category divided by the the total number of births sampled multiplied by 100. provided in the detailed tables. Additionally, in 1988, data concerning Hispanic origin and parental education were not reported on birth certificates in all States. Compari- sons for those items are based only on the records of respondents residing in States which collected that infor- mation on the birth certificate (see the individual tables for lists of the States). Birth certificate data were used as the denominators for the comparison rates in this report. Although this facilitates the examination of potential misclassification of information on birth certificates, it is important to note that there is no systematic method for determining whether the response on the birth certificate or the maternal survey is the most accurate, should they differ. Demographic measures Race and Hispanic origin of mother and father Over 98 percent of mothers reported the same race on both the birth certificate and the mother’s questionnaire (tables B and 1). The level of agreement was similar for black and for white mothers. For the father’s race, the comparability between the birth certificate and the mater- nal questionnaire was also approximately 98 percent among both black and white fathers (tables B and 2). For the mother’s race, the rate of item nonresponse was similar for black and white mothers. Approximately 2 percent of records were missing data on the mother’s race from the birth certificate and/or the maternal survey. However, approximately 43 percent of black mothers’ records and 11 percent of white mothers’ records were missing data on the father’s race from at least one of the sources. Among the black mothers missing data on the father’s race, 91 percent were missing data only from the birth certifi- cate. Among those records, over 98 percent of the fathers were reported as black on the maternal questionnaire. Data from the District of Columbia and the 30 States that collected parental Hispanic origin on the birth certif- icate in 1988 show that the agreement on Hispanic origin as reported on the birth certificate and on the maternal questionnaire was over 97 percent for both mothers and fathers (tables C and 3,4). Nativity of mother Overall agreement for maternal nativity was greater than 99 percent among both white and black women (tables D and 5). Nearly all of the mothers identified on their infant’s birth certificate as being born in the United States also reported being native born on the maternal questionnaire. The rate of agreement was slightly lower among mothers identified as being foreign born, but the number of foreign born mothers was small. Age of mother and father While the mother’s age at the birth of her infant is reported directly on the birth certificate, her age at deliv- ery in the NMIHS was calculated using the mother’s and the infant’s date of birth. The overall agreement for mother’s age, using grouped intervals, was approximately 4 Table B. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for race of parents, by race: United States, 1988 Race Mother Father Percent Total suv envnsmeresn sme 98.4 98.2 WHS ; wv sm ves sme mm 8m 5 55 55 3 98.3 98.4 Blagk , sv sn ons sms mw same os 98.4 97.9 Table C. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for Hispanic origin of parents, by Hispanic origin: United States, 1988 Hispanic origin Mother Father Percent TO ; cosmo v ams sa vw Imy 63 97.9 97.6 HISPANIC . ows ms ews ws sweme sma 98.1 85.8 NON-HISPANIC . v4 ows va swiss sy 97.9 97.9 NOTE: Hispanic origin was collected on the birth certificate in 1988 in the 30 following States and the District of Columbia: Alabama, Arizona, Arkansas,California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, Indiana, Kansas, Kentucky, Maine, Massachusetts, Mississippi, Montana, Nebraska, Nevada, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Tennessee, Texas, Utah, Washington, and Wyoming. Table D. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for nativity of mother, by race and nativity: United States, 1988 Nativity of mother White Black Percent TOE «up :upmmrtprmErsss cmame 99.6 99.8 Nave BOI +. wu vss sss pms me 99.8 99.8 Foreign bom ss ur vns ns wx sus us 95.8 % 97 percent among both black and white mothers (tables E and 6). When considered by individual year, the agree- ment for mother’s age dropped to 92.0 percent for black mothers and 93.2 percent for white mothers. Item non- response for mother’s date of birth on the maternal questionnaire was low; 3.9 percent for black mothers and 2.5 percent for white mothers. There were no missing data for maternal age on the birth certificates because records with missing values for that variable are given an imputed age. Table E. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for age of parents, by race and age: United States, 1988 Mother Father Age White Black White Black Percent AI BDTE 1s co: 1 i010 te i ot hers i 8k 0 rl be ds 0 tb 97.6 96.9 71.6 68.2 Underi8years. . ........cinnnn.n 93.8 96.5 * * T8—1QYCAIS 5 5s ci ah Sint Blk ht 4h 50h ABE oh 8 eu ptt 96.2 94.7 ® * D028 YOBISi 5 ra 57% ss Ls ish 0 him 0 om nn 97.7 97.7 62.0 66.6 25-20 YBAIB. 5 55 ll 0s SAE RE ARE AE.E0EGL SRE 97.6 97.4 69.6 69.1 BO-BA YORI: 5 tn ses ws 05 5 0 AER GIE BE REE SE Ha 98.5 96.7 76.5 73.6 BE-BOYSBIS, cv dm bins aE AR EME £0 End des 97.4 96.1 77.5 69.2 AD Years ANA OVEY «vs vs ss sR mn sHE TRL FREE 3 100.0 94.9 98.8 94.1 NOTES: Agreement is for grouped data. Father's age reported on the birth certificate is his age at the time of the infant's birth, while father's age reported on the questionnaire is his age at the time of completion of the questionnaire. Table F. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for education of parents, by race and education: United States, 1988 Mother Father Completed years of education White Black White Black Percent TO inser se sme nh PEs 9 ORT E3 H@E.50 865.4 85.9 79.5 84.3 75.4 TI Ves Or fBWEeE wi: nis sus 55 Gms vis amis da A 83.5 80.0 82.7 71.7 V2YSRIS . as iv sms as Ea ne RARE FEE EME 85.7 77.8 83.8 77.9 13 AB YORI vs sms es sas ms PIMs AMEE AEE 82.6 79.8 75.0 68.5 1B YEAS Or MOVE. 5 us + sss sats smbms va sis 92.7 86.2 92.7 79.5 NOTES: Data for California, New York State (exclusive of New York City), Texas, and Washington are excluded because the information was not collected on the birth certificate in these States in 1988. Agreement is for grouped data. The NMIHS did not collect information regarding the father’s age at the date of the birth of the infant. The NMIHS reports the father’s age at the time the maternal survey was completed (6-30 months after delivery of the infant), while the birth certificate reports the father’s age at the date of the birth of the infant. Consequently, the rate of agreement between the birth certificate and the maternal questionnaire by father’s age group was rela- tively low, approximately 68 percent among black fathers and 72 percent among white fathers (tables E and 7). When considered by individual year, the agreement was lower; 8.1 percent and 6.0 percent for black fathers and white fathers, respectively. Subtracting one year from the father’s age as reported on the NMIHS to estimate the effect of the reporting differences increased the rates of agreement between the two data sources. For the grouped ages, the agreement rate increased to 84.0 percent among black fathers and 89.4 percent among white fathers. Indi- vidual year comparability improved as well; 46.6 percent among black fathers and 54.3 percent among white fa- thers. As was the situation with race, more data were missing for father’s age than for mother’s age. For black mothers 45.6 percent and for white mothers 10.9 percent were missing data on father’s age on either the birth certificate and/or the maternal questionnaire. Education of mother and father In 1988, parental education was reported on the birth certificates of all States except for California, New York State (exclusive of New York City), Texas, and Washing- ton. The overall agreement for mother’s education was approximately 80 percent among black mothers and 86 per- cent among white mothers (tables F and 8). For both races, the highest comparability rates were among those mothers reporting at least 16 years of education. Agree- ment for father’s education was somewhat lower than that for mother’s education for both black and white fathers (tables F and 9). As was true for the mothers, the highest levels of agreement were among those fathers reporting at least 16 years of education. Several explanations may account for discrepancies in education level between the birth certificate and the maternal questionnaire. First, some parents may have completed more school in the period between the birth of the infant and the time the questionnaire was adminis- tered. Second, the questionnaire considers a high school graduate as having completed 12 years of education, regardless of how long it took to obtain a high school diploma. The birth certificate simply reports the number of years of schooling. Finally, the questionnaire but not the birth certificate, distinguishes academic from voca- tional training. For example, of those women reporting 12 years of education on the questionnaire, but more than 12 years on the birth certificate, approximately 75 percent of white mothers and 62 percent of black mothers reported additional vocational training on the questionnaire. Overall, 3.1 percent of black mothers and 2.1 percent of white mothers were missing data on maternal educa- tion. Father’s education was missing for approximately 50 percent of the black mothers and 13 percent of the white mothers. As was the case for father’s race and age, the majority of the missing data on father’s education was missing from the birth certificate only. Marital status The overall agreement for marital status at the time of birth of the infant was approximately 94 percent among black mothers and 96 percent among white mothers (tables G and 10). Comparability between the birth certificate and the questionnaire was the highest among unmarried black mothers and married white mothers. Approximately 5 per- cent of black and white mothers were missing information on marital status at the time of birth. Table G. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for marital status of parents, by race and marital status: United States, 1988 Marital status White Black Percent Total ................. 95.9 93.9 Married. . .............. 97.4 91.6 Unmarried. . ............ 88.6 95.1 For the majority of the United States, marital status is reported directly on the birth certificate. However, for eight States (California, Connecticut, Maryland, Michi- gan, New Hampshire, North Carolina, Ohio, and Texas) parental marital status is inferred by comparing the sur- names of the mother, father, and, if necessary, the child. This method of determining marital status on the birth certificate may explain some of the discrepancy between the birth certificate and the maternal questionnaire, par- ticularly among populations of women most likely to retain their maiden name after marriage. Pregnancy history measures Live birth order On the birth certificate, live birth order is determined directly from an item specifying the total number of live births the mother had, including the current birth. The NMIHS maternal questionnaire asked for detailed descrip- tions of factors associated with all previous pregnancies, including live births, stillbirths, miscarriages, and induced abortions. Live birth order from the maternal question- naire was calculated as the sum of the prior live births plus the current birth. The overall agreement rate for live birth order was approximately 82 percent among black mothers and 89 per- cent among white mothers (tables H and 11). The agree- ment was slightly less when birth orders higher than four were examined individually, not combined as in table H. For black or white mothers, the comparability was the highest when the current birth was reported as the first birth. The agreement between the maternal questionnaire and the birth certificate was lower with each subsequent birth. Information necessary to compare live birth order was missing for 4.2 percent of black mothers and 3.0 per- cent of white mothers. The birth certificate was more likely to report a high live birth order than was the maternal questionnaire (table 11). Among black mothers, 29.6 percent of records reporting a live birth order of four or higher on the birth certificate reported fewer than four live births on the questionnaire. Conversely, of records with a live birth order of four or higher on the maternal questionnaire, 9.8 percent reported fewer than four live births on the birth certificate. The situation was similar among white mothers, where the respective percents were 23.2 percent and 5.6 percent. Prior fetal deaths The number of prior fetal deaths from the birth certificate was defined as the sum of reported pregnancy terminations (spontaneous and induced) occurring before and after 20 weeks of gestation. The number of prior fetal deaths reported on the NMIHS maternal questionnaire was calculated as the sum of all reported stillbirths, miscarriages, and induced abortions. Table H. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for live birth order, by race and live birth order: United States, 1988 Live birth order White Black Percent AU DINS + cu vs vn vig vs sma 88.8 82.4 YSLIVE BIR. ov vas mn suns wn 97.8 95.9 Zr IVE BIN cv ivs vn vn ne vas 84.0 77.0 BOVE BIN. . wu vv swrmn vas 80.6 73.0 4th or higher live birth. . . . ....... 78.7 70.4 Table J. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for prior fetal deaths, by race and number of fetal deaths: United States, 1988 Fetal deaths White Black Percent TO sms swims swnws spss 68 3 81.0 77:7 NONE . oc covwssnsns smpmnews 86.0 88.7 1priorfetal death ......o000 4. 69.9 50.8 2 prior fetal deaths. . . .......... 57.7 352 3 or more prior fetal deaths . . . .... 51.8 35.1 Approximately 78 percent of black mothers and 81 per- cent of white mothers had the same number of prior fetal deaths recorded on the maternal questionnaire and the birth certificate (tables J and 12). The agreement was the highest for women with no fetal deaths reported on the birth certificate and decreased with each additional fetal death. Approximately 4 percent of black mothers and 3 percent of white mothers were missing information needed to calculate the number of prior fetal deaths. The distribution of the number of prior fetal deaths was similar whether data from the birth certificate or the maternal questionnaire was examined (table 12). How- ever, on an individual record level, the agreement between the two sources was not high. Unlike the situation for live birth order, there was no consistent trend for either the birth certificate or the maternal questionnaire to report a higher number of prior fetal deaths. Pregnancy measures Plurality The agreement for plurality of the delivery was over 99 percent among both black mothers and white mothers (tables K and 13). Singleton births were more likely to be identified on both the maternal questionnaire and the birth certificate than were multiple births. Data on plural- ity was missing from the questionnaires of 5.5 percent of the black mothers and 4.4 percent of the white mothers. The birth certificates had no missing values for plurality. Timing of prenatal care initiation The timing of the first prenatal care visit as reported on the birth certificate is derived from an item specifying the month of pregnancy prenatal care began (for example, first, second, etc.). The maternal questionnaire asked the mother, “How many weeks pregnant were you when you went for your first prenatal visit?” The overall agreement for timing of the first prenatal care visit varied by maternal race; 67 percent for black mothers and 85 percent for white mothers (tables L and 14). Among mothers who are reported as having received first trimester prenatal care on the birth certificate, over 87 percent of the black mothers and 95 percent of the white mothers also reported first trimester care on the NMIHS questionnaire. However, fewer than 40 percent of the women reported as having initiated prenatal care later than the first trimester on the birth certificate also re- ported delayed initiation of care on the maternal questionnaire. Opportunities for errors in reporting the timing of prenatal care initiation exist for both the birth certificate and the maternal questionnaire. The method of collection of data for the birth certificate varies depending on the site of delivery. Some centers collect information directly from the mothers, while others abstract it from medical records. There are opportunities for error in both meth- ods, particularly if the mother switched prenatal care providers during pregnancy (3). Information regarding the timing of prenatal care collected on the NMIHS question- naire may be inaccurate because of the time lag between prenatal care and the completion of the maternal ques- tionnaire. Up to 3 years can separate those events. 8 Table K. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for plurality, by race and plurality: United States, 1988 Plurality White Black Percent TOR wsnmsvsvssmsma swans a 89.7 99.1 SINGIOtON: os sowisvy sms up saa ms £3 99.8 99.5 TWIN Or igher « wu vwisms swiss sa 96.4 89.0 Table L. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for trimester of first prenatal care, by race and trimester: United States, 1988 Trimester first prenatal care White Black Percent Total ........ 84.9 66.9 Fst Himester: . ws sa smi swims v4 95.4 87.2 Second trimester. ; .: cw: ssw ins 39.7 33.2 Third trimesterornocare ........ 32.6 36.2 Table M. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for number of prenatal care visits, by race and number of visits: United States, 1988 Number of prenatal care visits White Black Percent Tol somsns emams sasns ining 46.5 39.6 NOVSHS «vcr swsms sus vs mime 51.7 50.4 TBYIOUS. ¢ v3 sms ws suis es sn smy 35.7 36.3 7-10VIBUS : v5 sss swans smsms 32.9 34.6 T1=1BVIBUS + « sus vs sus ms smo wy 57.3 45.2 172 Or Ore VIsHS «vu so sms su vs 44.0 34.2 NOTE: Agreement is for grouped data. Number of prenatal care visits The overall agreement rate for the number of prena- tal care visits the mother received was lower than 50 per- cent among both black and white mothers, even after grouping of data (tables M and 15). The relatively poor comparability between the birth certificate and the ques- tionnaire for this item may be partially due to the same factors that influence the reporting of timing of the initial prenatal care visit. An additional factor that may affect the reporting of the number of prenatal care visits is the variation in determining exactly what constitutes a prena- tal care visit. The reporting of appointments for pregnancy tests, sonograms, and other prenatal tests as a prenatal care visit may vary depending upon where the test oc- curred and whether the mother or the facility supplied the information for the birth certificate. Gestational age Gestational age from the birth certificate is measured by subtracting the date of the mother’s last menstrual period (LMP) from the date of the infant’s birth. The duration of the pregnancy as reported on the maternal questionnaire came from a question that asked, “How many weeks did this pregnancy last?” Overall, the agreement between gestational age, as reported in 4- to 5-week intervals, between the question- naire and the birth certificate was 66 percent among black mothers and 79 percent among white mothers (tables N and 16). The agreement was substantially higher among births of 38-42 weeks gestation than among births of less than 38 weeks gestation. This finding may support evi- dence showing that the LMP measure of gestational age is at its most accurate when assessing full-term births (4). Among both the black and the white populations, the proportion of infants reported in each gestational age category below 33 weeks was slightly higher on the mater- nal questionnaire than on the birth certificate (table 16). However, 20.2 percent of black infants were reported as having a gestational age of 33-37 weeks on the birth Table N. Percent of responses from the mother’s questionnaire in agreement with the birth certificate for length of pregnancy, by race and gestational age: United States, 1988 Gestational age White Black Percent TO sunwwwouwy my sea ny spss oy 78.7 66.0 Lessthan24weeks. . . ......... * 41.9 24-28 WBEKS . vi vv vw vs wn maw 73.7 64.8 20-32 WEBKS ww vs iv on mwa iw 0 4 61.6 44.4 BE-37WeEKS «svt nv mas meme wn 57.9 29.1 BE-A2 WEBKS wiv ix v wwii a vs wow in 5 win 91.5 89.7 Greater than 42 weeks . . . ....... * * NOTE: Agreement is for grouped data. certificate, compared with 11.1 percent on the maternal questionnaire. Fifty-three percent of black infants were given a gestational age of 38-42 weeks on the birth certificate, compared with 65 percent on the maternal questionnaire. Fifty-two percent of black infants reported as being 33-37 weeks gestation on the birth certificate were reported as having a gestational age of 38-42 weeks on the maternal questionnaire. The discrepancy among white infants was less, but still high. Among white infants with a gestational age of 33-37 weeks on the birth certifi- cate, 27.6 percent had a gestational age of 38-42 weeks on the maternal questionnaire. These variations in the report- ing of gestational age may be important when examining prematurity as a pregnancy outcome. Unfortunately, de- ciding whether the birth certificate or the maternal ques- tionnaire provides the most accurate estimate of gestational age is not possible from this information. Discussion The degree of comparability between data collected on the birth certificate and information reported on the 1988 National Maternal and Infant Health Survey mater- nal questionnaire was dependent on the item examined. In general, agreement rates were slightly higher among white mothers than among black mothers. Parental race and Hispanic origin, maternal nativity, maternal age, and plu- rality all had agreement rates of over 95 percent for both black and white mothers. Responses to other items common to the birth certif- icate and the maternal questionnaire, such as number of fetal deaths, gestational age, and measures of prenatal care, were not highly comparable between the two data sources. While the overall distribution of most of the variables was similar for the birth certificate and the maternal questionnaire, the lack of comparability for those items may limit the utility of individual-level analyses using that information. Particularly, indices of prenatal care adequacy that are based on gestational age and number of prenatal care visits as reported on birth certif- icates may be of questionable validity, since the agreement in the reporting of either of those factors was relatively low. Although the sampling and weighting scheme of the NMIHS was designed to produce a nationally representa- tive sample of U.S. live births in 1988, these comparability results may not completely represent the population as a whole. Since the NMIHS included an oversampling of low-birth-weight births, the findings of this report may be different than if a random sample of births was chosen. Overall survey nonresponse and individual item nonre- sponse were not considered in this analysis and may also limit the generalizability of each comparison. However, as mentioned earlier in this report, utilizing the weights that account for the survey sampling scheme and survey non- response did not affect the comparability rates. There is also potential for bias among the respondents of the maternal questionnaire. Mothers who received the ques- tionnaire late or who delayed completing and returning the questionnaire may have had more difficulty recalling details of the index pregnancy, particularly time- dependent items such as timing of prenatal care, than did mothers who completed the questionnaire promptly. Ad- ditionally, this time lag could have made it possible for responses to the maternal questionnaire to be confused with a more recent pregnancy than the one referred to on the birth certificate. The comparability rates between the birth certificate and the 1988 NMIHS maternal questionnaire are gener- ally somewhat lower than those between the birth certifi- cate and the 1980 National Natality Survey (1). The exception was the agreement for parental race, which was higher in this report. The comparability of other variables reported here, such as parental Hispanic origin, marital status, and number of prenatal care visits, was not mea- sured in the NNS report. A possible explanation for the apparent decline in comparability is that the NMIHS was composed of a higher risk population than was the NNS. The information on either the birth certificate or the maternal questionnaire may be less likely to be accurately reported for the NMIHS population. From these analyses, one cannot conclude whether the birth certificate information is accurate for any given item. However, a high rate of agreement between the birth certificate and the maternal questionnaire may support the validity of a particular data item, while a low rate of agreement may highlight a source for potential problems. This report should be helpful in determining which vari- ables commonly used in the reporting of national natality data are likely to be accurately reported on the birth certificate. References Fingerhut LA, Kleinman JC. Comparability of reporting between the birth certificate and the 1980 National Natality Survey. National Center for Health Statistics. Vital and Health Statistics 2(99). 1985. Sanderson M, Placek PJ, Keppel KG. The 1988 National Maternal and Infant Health Survey: Design, content, and data availability. Birth 18:26-32. 1991. Liberatos P, Kiely JL. Selected issues in the evaluation of prenatal care. In: Kiely M, ed. Reproductive and perinatal epidemiology. Boca Raton, FL: CRC Press, Inc., 79-97. 1991. Kramer MS, McLean FH, Boyd ME, Usher RH. The validity of gestational age estimation by menstrual dating in term, preterm, and postterm gestations. JAMA 260:3306-8. 1988. 11 List of detailed tables 0 12 . Number of responses by race of mother on birth certificate and mother’s questionnaire: United States, LOBE sms ms sede mon ori on stacey pains issn os icon . Number of responses by race of father on birth certificate and mother’s questionnaire: United States, FOBT 56 00 0 000 5 1 rn 1 omic i of . Number of responses by Hispanic origin of mother on birth certificate and mother’s questionnaire: United States, 088. . co uv rvomens su rmms mess ch vm reams . Number of responses by Hispanic origin of father on birth certificate and mother’s questionnaire: United States, 1988. to vmiarsmrns smrma su dms ympmy va nm . Number of responses by race and nativity of mother on birth certificate and nativity on mother’s questionnaire: LInited SLates, 1088. vv wx wir oie nm comin om mes its om wi wim nome . Number of responses by race and age of mother on birth certificate and age of mother on mother’s ques- tionnaire: United States, 1988. ................... . Number of responses by race and age of father on birth certificate and age of mother on mother’s question- naire: United States, 1988... .............. .. .... . Number of responses by race and education of mother on birth certificate and education on mother’s question- naire: United States, 1988. ...................... 13 13 13 14 14 14 15 10. 11. 12. 13 14. 15. 16. . Number of responses by race and education of father on birth certificate and education on mother’s ques- tionnaire: United States, 1988. .................. Number of responses by race and marital status of parents on birth certificate and marital status on mother’s questionnaire: United States, 1988. ....... Number of responses by race and live birth order on birth certificate and live birth order on mother’s ques- tionnaire; United States, 1988. . vcvvivrsveomer veins Number of responses by race and prior fetal deaths on birth certificate and prior fetal deaths on mother’s questionnaire: United States, 1988 ............... Number of responses by race and plurality on birth certificate and plurality on mother’s questionnaire: United States, 1988 ........ eB Kit a PE i El Hanes ore Number of responses by race and month of first prenatal care on birth certificate and week of first prenatal care on mother’s questionnaire: United States, JOBE smi ami rms wn emp mdm awh Ry nh Number of responses by race and number of prenatal care visits on birth certificate and number of visits on mother’s questionnaire: United States, 1988. ....... Number of responses by race and gestational age on birth certificate and gestational age on mother’s ques- tionnaire: United States, 1988. .................. 16 16 17 17 18 18 19 19 Table 1. Number of responses by race of mother on birth certificate and mother’s questionnaire: United States, 1988 Race on mother’s questionnaire Asian or Eskimo/Aleut/ Pacific American Race on birth certificate Total White Black Islander Indian Missing TO cs v7 sms i om bmp sm 3% DFE Pe Da 98,953 4,588 4,810 250 96 209 WHE 40s isso 5 Bid he 550 5 or 3 0 1m 2 0 3 09 4,695 4,516 27 21 29 102 BIER eo om 505 5% 4 55 BE RB RIB RET 4,956 47 4,778 22 8 101 Asian or Pacific Islander . . .............. 216 8 1 203 - 4 BIAWEHBAT +o v0 0 2 0 3% 40 0 80 0 0 i 0 0 ole 3 10 1 - 9 - - COMBE. vii) ee 10 0 0 10 0570 i 3000 70 fo 300 0 40 - - 40 - - EL EE TT 20 1 - 19 a — BHIDING 5: ire sama nm om om mobs 0 ae 252 on 45 2 41 - 1 OMYEr +5 is 29 00 8 000 # 5 2030 i 8 80 0 6 Bo 3000 20 101 4 - 94 » 3 Eskimo, Aleut, and American Indian. . ....... 70 8 —- 1 59 2 OMB is vn swam sm soim swam via mn s@ in 1 1 - - oe — MISSING. cv cov vnemssmembmnsmemn russ 15 8 4 3 - - Table 2. Number of responses by race of father on birth certificate and mother’s questionnaire: United States, 1988 Race on mother’s questionnaire Asian or Eskimo/Aleut/ Pacific American Race on birth certificate Total White Black Islander Indian Missing Molal cmv smnims amome sis ms nme ms spn 9,953 4,411 4,886 209 100 347 WIRES. , iv saison smsms smamani smewiman 4,178 3,992 19 20 25 122 Blok oui saison ime wie ame Ea By sme gd 3,088 41 2,929 15 7 96 AsianorPacificlslander . . ..... avn: sass 174 8 - 157 1 8 Hawallah ... 0 co s@e tos imemo ny sa 9 2 - 7 - - Chinese. ..... ucoicnimsngs sms ani 38 - - 38 - - JEADBNBBB vs ws wos WE Ams WE es Big win 13 1 - 12 - FHDING 2 smi da sme mn imi ihr Bodies 27 1 - 22 - 4 OEE 5; ia sei aR Rms Me mE us mE 20 87 4 ws 78 1 4 Eskimo, Aleut, and American Indian. . . ...... 52 7 —- 2 42 1 OUel cma pasate a ¢ REP RABE 4S Smead 2k 1 1 - - - WISE. wis ub ime pis Eg MEME AE smEAE fa 2,460 362 1,938 15 25 120 Table 3. Number of responses by Hispanic origin of mother on birth certificate and mother’s questionnaire: United States, 1988 Hispanic origin on mother’s questionnaire Non- Hispanic origin on birth certificate Total Hispanic Hispanic Missing Tolal sons nrmmsms mens neni mEs BEEBE 6,916 780 5,937 199 HISPANIC : sms mms ina mmeme mh omsmms se dessa 676 652 13 11 NON-HISPANIC + vo «wo mns mrss rms @ps De® ise 5,993 122 5,689 182 TT RR pe 247 6 235 6 NOTE: Hispanic origin was collected on the birth certificate in 1988 in the 30 following States and the District of Columbia: Alabama, Arizona, Arkansas,California, Colorado, Connecticut, Florida, Georgia, Hawaii, lllinois, Indiana, Kansas, Kentucky, Maine, Massachusetts, Mississippi, Montana, Nebraska, Nevada, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Tennessee, Texas, Utah, Washington, and Wyoming. 13 Table 4. Number of responses by Hispanic origin of father on birth certificate and mother’s questionnaire: United States, 1988 Hispanic origin on mother’s questionnaire Non- Hispanic origin on birth certificate Total Hispanic Hispanic Missing TOBY 5 ius en vm bids amps pote sa wg 5001 ne 6,913 727 5,788 398 PHSDANG . v3 vn v v5 smash Rms ma ewnins vay ws 564 522 23 19 Non-Hispanic 4,555 92 4,215 248 NABBIA 2 tne «ims Mim moron ii pk BR ag a 85 0 8 1,794 113 1,550 131 NOTE: Hispanic origin was collected on the birth certificate in 1988 in the 30 following States and the District of Columbia: Alabama, Arizona, Arkansas, California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, Indiana, Kansas, Kentucky, Maine, Massachusetts, Mississippi, Montana, Nebraska, Nevada, New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Tennessee, Texas, Utah, Washington, and Wyoming. Table 5. Number of responses by race and nativity of mother on birth certificate and nativity on mother’s questionnaire: United States, 1988 Nativity on mother’s questionnaire Native Foreign Race and nativity on birth certificate Total born born Missing White Tolga) oss snes amin ines wmeeps ws 4,695 4,100 463 132 NBHVB BOM . cvs snus smoms sa sms emaoms 4,204 4,079 9 116 POTBIGH BOF «vcs vai mmomavms aio £0 5h me pa 244 10 228 6 MISSING. « ws atismn cma mp mi ms Sa niin oi ems 247 1 226 10 Black Tol) vive svimsnmams ev ams sn sms en sma 4,956 4,443 272 241 INBUVE IOI + vv sms ww oi mw m0 oo oor oe» 6 8 3 #0 4,626 4,402 11 213 FOIBIGN BOI... .. cvs 50 wins 0 wiih cin 8 et oo mg ws 13 - 13 - INISEHAG. « xis vv ev i 0m is wom mali bmi eh 317 41 248 28 Table 6. Number of responses by race and age of mother on birth certificate and age of mother on mother’s questionnaire: United States, 1988 Age on mother’s questionnaire Under 18-19 20-24 25-29 30-34 35-39 40 years Race and age on birth certificate Total 18 years years years years years years and over Missing White TOW coims mm rman ems mb pms wms 4,695 154 318 1,189 1,456 1,070 342 48 118 UAdEr IB Years. . vs vn smn dn dws 45 165 152 7 1 - - — 2 3 18-19 VRAIS. . . viv cus ma rin sins vin 324 - 305 10 2 - - — 7 POP YORIS. «vv vv vn sivis sn smn sme 1,220 2 4 1,167 17 5 wt - 25 P5-POYBAIB, vows vwsinn inva ws 1,508 - 2 6 1,432 23 5 - 40 BO-B4YBAIB. «oon rrr ir rwrrns 1,080 - - 5 4 1,035 7 29 B5-BOYBAIS. «vx rrr mama 352 o- — — 1 7 330 1 13 AOYBBIS OF OVI . vc vv vrs va vin ws 46 - - - - - - 45 1 Black TOR ois ms wns ms mmuins Hie mn wns 4,956 475 614 1,536 1,202 666 233 39 191 UNCBr 1B Years. cv v.msvwems wes 509 469 16 1 - - - - 23 1BADYBAIS.: svi s wv vv wn ma tam 639 4 592 19 9 1 - - 14 20-24 years. . . o.oo 1,589 1 4 1,501 20 10 - - 53 25-20vears. .... coin iaun 1,238 - 2 6 1,181 18 5 - 46 BO-B4YBBIS: wu tv ui bhi ws mae 692 - - 9 9 634 4 - 36 BEBO YBAIS. i «vin nn un ea 246 1 - - 3 3 222 2 18 AO YBAIS OF OVE « viv wn sin min ui ini 43 - - - - - 2 37 4 Table 7. Number of responses by race and age of father on birth certificate and age of mother on mother’s questionnaire: United States, 1988 Age on mother’s questionnaire Under 18-19 20-24 25-29 30-34 35-39 40 years Race and age on birth certificate Total 18 years years years years years years and over Missing White Tolal iv vivms mums omens shvms 4,695 5 76 662 1,352 1,386 721 390 103 Under 1B years. . «« vn cos wn smsms 24 1 21 1 - - 1 V8-TOYBAIS. . oc cov mn vmimn vn ews 93 - 22 70 - 1 - - - PO-2AYORIS. , ns cvema sms va twa 756 - - 462 278 2 2 1 11 28-20 YOBIS. . » » cuvwia mien nn Ea 1,399 - 1 5 954 406 3 2 28 B0-B4YBAIB. . vs «rms www 5 Swe 1,184 - - - 3 897 270 3 11 BE-=BDYBAIB. . s sins ms amas vi an 529 - - - 1 4 407 113 4 AO YEAIS OF OVE « «viv vw vin vo sims 264 - - - 1 2 - 252 9 MASSA » css naan shine nis ius 446 4 32 124 118 74 39 19 39 Black TOE) ois iamsmd omedy dhims 2s 2k 4,956 32 229 1,269 1,366 940 500 388 232 Under 1BYears. . .. iva vo smn vw sins 54 10 38 3 1 - - 2 18-19 YBBIS. + is avi imn Am RHE 45 » 8 127 - 29 90 - — - 8 20-28 YBOIB: ss vw dius SRA Dib ER &R a 672 - 2 440 211 3 4 1 11 DE POYBBIS... : ws vmhis th Enh 4a Ema 854 - 2 10 872 229 9 6 26 B0-34YRBIS. . is vw sins smb wenn pma 594 — 1 2 9 420 130 9 23 BE5-BOYBBIS. «vs cvsms seams us Fs 287 - - 1 3 8 193 74 8 A0 Years Or OVer . ou vss sos vs vn sus 203 1 - - 3 4 3 174 18 MISBING. ; spss crews swams po ves 2,165 21 157 723 567 276 161 124 136 NOTE: Father's age reported on the birth certificate is his age at the time of the infant's birth, while father's age reported on the questionnaire is his age at the time of completion of the questionnaire. Table 8. Number of responses by race and education of mother on birth certificate and education on mother’s questionnaire: United States, 1988 Education on mother's questionnaire 0-8 9-11 12 13-15 16 years Race and education on birth certificate Total years years years years or more Missing White TORI . ois suie ms Risa sults 002M 3,468 112 447 1,400 840 644 25 O-BYSRS uc sv ums cnsmn cmon muy 105 72 14 9 4 2 4 S=11'YOBIS. + sn ts vmeiia sw ima 40 an 487 25 375 69 11 1 6 T2VOBIS , oo vs ving entms susws umn 1,419 5 47 1,207 135 15 10 13-15Y8arS. . . : vc cv vme va me winnie 781 2 1 97 641 35 8 1B YBAIS Or MOIS. v « vv vviv vis ms vivre 627 3 - 6 37 581 = NUESIAG. cms nu ems hizms Spas on 28 49 5 10 12 12 10 - Black Toll pve vs sma nn smn va emsinn vu 4,213 118 1,107 1,765 846 313 64 D-BYBBIS oo vu vovnva vwswan vmnmn nw 122 63 43 10 1 1 4 O-T1 VRAIS. . .. cov vs serra a 1,211 36 900 221 22 5 27 TYAS... vo vnsns smrin amen vo 1,808 16 137 1,392 223 22 18 13-18Y@aIS. .. . cc vv nvr ia ra en 708 2 13 99 556 27 11 18 YeAIS Of MOM. « + vv vv t wv rims imn ein 300 = - 8 33 257 2 MISBING. + vv en sons va sar on smzms vs 64 1 14 35 11 1 2 NOTE: Data from California, New York State (exclusive of New York City), Texas, and Washington are excluded because the information was not gathered on the birth certificates from those States in 1988. 15 Table 9. Number of responses by race and education of father on birth certificate and education on mother’s questionnaire: United States, 1988 Education on mother’s questionnaire 0-8 9-11 12 13-15 16 years Race and education on birth certificate Total years years years years or more Missing White TOMA « voi on 5% 0 vod sims: 0st ie i owe 3,468 120 408 1,384 627 820 109 B=BYBBIS «ov i bm #0% 0 0m 0 oi mh 1 dh 82 54 15 4 1 = 8 GAT YBBIS. cou mampnmoss dims ssa 343 23 243 62 3 - 12 PEAS. ir 17, + 5 9 0 5k 00 0 0 od 1,269 13 56 1,048 113 20 19 Ly ET ET ETT 598 po 2 96 442 49 9 1B YEAS Or MOC. « uv ni ini 0 0 36 5 wi 784 - - 17 40 724 3 DAUSSING wir io 0 ool 0 om 1 5 i 0 DIES #8 392 30 92 157 28 7 58 Black TOA 5 oc nie dime vi 08 0 00 0 Bo Fe So 4,213 100 686 2.17 632 311 367 O-BYBBIS «vn v0 4 vin 5.0 fio wow 30 5 10 ws 44 25 5 9 2 - 3 LEER SE rp 365 14 229 89 5 3 25 T2YOBIB « «vie vin ck 0% wi 01 410 ow a) ms nh 1,204 11 73 901 139 32 48 VB=ABYOAIS. wv vv viv vim vie 5mm bi mine 360 - 4 80 239 26 11 TB YOBIS OPINOIG.. « vv + 4 ais vou visi sma 223 1 4 8 31 171 8 MISSING, © 5 10 6 6 000 50s: I) 0) oe 9) £00 Bb 2,017 49 371 1,030 216 79 272 NOTE: Data from California, New York State (exclusive of New York City), Texas, and Washington are excluded because the information was not gathered on the birth certificates from those States in 1988. Table 10. Number of responses by race and marital status of parents on birth certificate and marital status on mother’s questionnaire: United States, 1988 Marital status on mother's questionnaire Race and marital status on birth certificate Total Married Unmarried Missing White FOE osama ns sus samy RL EER 4,695 3,663 772 260 Marmed. . c ovons su rms sis soma wd 3,851 3,576 96 179 KINMmaried . «ov sms cmamia smite spiimn ims 844 87 676 81 MISSING. ws ws tmp smams nme ms smemp sw s - - - - Black THE ovo swing ume ms sas ms ames 453 4,956 1,670 3,034 252 MEIC. x vw cws swim swiss cmsdh bus 1,767 1,522 139 106 Unmarried. vo. us ss iws amid: imams aes 3,188 148 2,894 146 MISSING: is .o 0 sma salsms smsms vmsme sos 1 — 1 i 16 Table 11. Number of responses by race and live birth order on birth certificate and live birth order on mother’s questionnaire: United States, 1988 Live birth order on mother's questionnaire 1st 2nd 3rd 4th or higher Race and live birth order on birth certificate Total live birth live birth live birth live birth Missing White THA ; ns crismils sRFmpiwriws sme sms 4,695 2,210 1,355 675 321 134 Istive bith. . «suv insur imemnvmemunms 2,030 1,913 29 10 4 74 2nd NebIth .vseriminr sus smimn ams 1.516 204 1,242 31 2 37 Brad live bith. ... cu cos cmems smrwr mx 737 61 67 582 12 15 Ath orhigherlivabirth. .. co vv vou vwviwn wus 408 26 15 51 303 8 MISSING: ons sms me same Rmems nwhimn 2m 9 6 2 1 - - Black TO vcvn vmtmn cnms smsing adam p71 4,956 2,193 1,248 775 556 184 1SLIVB DIAN. «vo ms cmsnis sews smems sms 1,865 1,717 55 11 8 74 nd VeDINh . oii cams sme mn pas sme 1,424 259 1,064 53 6 42 Brave birth. . «cc cuss cms smpme sus 907 102 91 633 41 40 4th or higher live birth. . . . . ............. 732 107 30 72 497 26 MIBBING «4 ix sini M® Bis me ww bb Hain wn none 28 8 8 6 4 2 Table 12. Number of responses by race and prior fetal deaths on birth certificate and prior fetal deaths on mother’s questionnaire: United States, 1988 Prior fetal deaths on mother's questionnaire 1 prior 2 prior 3 or more prior Race and prior fetal deaths on birth certificate Total None fetal death fetal deaths fetal deaths Missing White TOBY 55.2 4 5m 5% 50 Bad 00 0 Gi 5s Bo i wis 4,695 3,180 953 301 127 134 NONE oie 54 10 le 5 508 3 & Hai [5 BEE 3 E FEE 3,555 2,957 364 79 38 117 1priorfetal death . .. us inven ssny gs ues 771 155 529 61 12 14 2priorfetaldeaths. . .....cos wuss sans 244 40 43 139 19 3 Bormorepriorfetal deaths . ....... cous vv 112 17 17 20 58 - MISSING. os vw wt 0 nw 0 vn me Ba 8 8 WEE ERE 13 11 w 2 - - Black TORE ove 5 em 005 2s fn ash ELAS: itl od ates 4,956 3,632 756 267 117 184 NODE. ii 5) 3008 15.0 10 5 Bie, 00 IAT) 0 00 0 hn 3,666 3,138 274 99 28 127 1 priorfetal death . . «ue + cme wm wom Le Bm 823 318 399 51 18 37 Spriorietal'deaths, . ces ss nbam ss 0 um ES 280 106 51 94 16 13 Bormorepriorfetal deaths . . ...c. vv as 156 47 30 21 53 5 MISSING. « wasp ress srs Wb as vmE ww 5s 31 23 2 2 2 2 17 Table 13. Number of responses by race and plurality on birth certificate and plurality on mother's questionnaire: United States, 1988 Plurality on mother’s questionnaire Race and plurality on birth certificate Total Singleton Twin or higher Missing White Yota] covevnnne nanan basis ings BEd wme 4,695 4,307 144 244 SInglRloN;: +» vo 5 sve s 8 Hew BEES BEE EYE 4,402 4,302 10 90 TWinorhighery sus ssnvs rs nates 588548 uss sas 293 5 134 154 Black TOLL oo nnn on intion soosmstmninn ) weston 0 cts mies orion 4,956 4,560 178 218 SIOIGION. : s sass rridisTds nanan s amas ain 4,711 4,541 24 146 TWINOL NIGNBYE wn v5 6.58 5 RSF % 2.500 3b bad 4 245 19 154 72 Table 14. Number of responses by race and month of first prenatal care on birth certificate and week of first prenatal care on mother’s questionnaire: United States, 1988 Week of first prenatal care on mother’s questionnaire Race and month of first prenatal care on 1-13 14-26 27-47 No prenatal birth certificate Total weeks weeks weeks care Missing White THA 2000 +0200 0d hE 5d wh ® ad BaPd 4,695 4,012 454 38 75 116 T-SMONMNS : cs sar cr RET IR AUR TE RAE TEE 3,733 3,485 145 3 21 79 AB MOINS «v1 ces vr SUEZ ERE SF BRET RS 649 357 247 11 8 26 7-9momhs . cc sns ss snr in smn as REESE 141 62 43 21 8 7 Noprenatal Care. zs ss swans vs snn ss wn som» 64 22 5 2 33 2 MISBIAG. 00 v0 wc monn wm wine ir ope a HE 3989 108 86 14 1 5 2 Black TOR crisis TRC ARPA IRE AE SFOS 4,956 3,584 867 105 231 169 1-3 MONINS vos rrnp snare ss aaE ATURE 3 2,953 2,495 302 20 45 91 B=B'MOMNG ¢ ccs meve 2PBEWE ERLE + HELE 7 1,313 776 421 33 37 46 Z-9montNS uss srr srre ne VINES BRENT 282 130 86 40 12 14 Noprenatal Care; sso vussssnsse yes ans 250 74 33 6 125 12 MIBBIAG.. : wv +: vw 5 vas ww ES 0 FES 5 TEE 158 109 25 6 12 6 18 Table 15. Number of responses by race and number of prenatal care visits on birth certificate and number of visits on mother’s questionnaire: United States, 1988 Number of prenatal care visits on mother’s questionnaire Race and number of prenatal care visits on No 1-6 7-10 11-16 17 or birth certificate Total visits visits visits visits more visits White THA ois ins sm sion Bh PER #30 0s El § 4 ol i 4 4,695 75 475 1,211 2,167 767 NOVISHS : is soisms chisims dumb in amamys 64 33 13 10 6 2 V=BVISHS: cs is 50s wis 4miving An Ama HE HBL AE 507 14 181 178 101 33 T=TOVISHS + cosine swsms pa smatn ims ah; 177 6 108 387 520 156 Y9-TBVISHS » ov vis sn rain ns amos img 28s 2,027 7 59 415 1,161 385 Y7 Or TOT ISHS . vo =n 20is £8 fms ni gis wos 232 - 10 26 94 102 NISSING. . wis soi sms vn sms nh sms ng ams un 3 688 45 104 195 285 89 Black TOUR) viens an sais mais we BR HHL Ee 4,956 232 840 1,426 1,746 712 NOISES: , ou im switch amd ms smn smEES 2 250 126 61 29 22 12 BIH, osm sis Bp pans Pm NE FR ue 1,948 49 405 326 253 82 T=IOVIBUS ..» wv 2 ix Ho sms 0h BE Ew 1,376 20 176 476 516 188 PI ABVISHS ; os ss mesma mvs masin 285 Ey 1,549 16 110 417 700 306 170 MOG VIBHE . « «vv ims sv cme wn smsmus 193 1 8 30 88 66 MISSING: 2 5% vem 5 30 © 010 0 05 5 Gym i oh for 6 ot 473 20 80 148 167 58 Table 16. Number of responses by race and gestational age on birth certificate and gestational age on mother’s questionnaire: United States, 1988 Gestational age on mother’s questionnaire Race and gestational age on Fewer than 24-28 29-32 33-37 38-42 43 weeks birth certificate Total 24 weeks weeks weeks weeks weeks or more Missing White MORE ov 5 4 wi avivii of di Strasse 4,695 93 316 318 629 3,152 89 98 Fewer than 24 weeks . . . . ..... 56 41 11 2 1 1 - - 24-28WEBKS ... wk mare @ lin 245 23 174 30 7 2 0 9 20-32°WEBKS «i mans 2 mas 8 301 5 60 181 30 17 1 7 B3-37'WEBKS «vv vm 5iv bw i wok 691 8 25 58 392 191 3 14 38-42weeks . ...... 2,953 9 10 21 158 2,649 49 57 43 weeksormore . . ......... 255 - 2 3 15 199 33 3 MISSING: wis odin wo 0 50% 1% 8 194 7 34 23 26 93 3 8 Black Total 2. 505 smi sms ® ans s 4,956 133 419 357 549 3,222 58 218 Fewerthan24 weeks . . . ...... 135 54 41 9 9 16 - 6 24-28WOEKS . i «ois bo iain bonis 302 27 188 42 14 19 - 12 20-32weeks . . ..........., 337 15 76 138 36 46 - 26 B3-B7 WBBKS . vk win yn 999 10 45 98 276 519 2 49 BB-A2 WEBKS . +» vn bk wr 4 wey 2,636 11 17 32 166 2,276 35 99 43weeksormore ........... 261 - 2 6 9 217 17 10 MISSING. + sia sigan wm twa vo sm ws 286 16 50 32 39 129 4 16 “U.S. Government Printing Office: 1893 — 342-327/80002 Series 2 No. 117 409 112 y. 117 JBL Vital and Health Statistics From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics National Survey of Family Growth, Cycle IV, Evaluation of Linked Design July 1993 Ca U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES yy =o) [eR Rl ela lel) r Centers for Disease Control and Prevention CENTERS FOR DISEASE CONTROL %, fis, National Center for Health Statistics Ent AND PREVENTION Trade Name Disclaimer The use of trade names is for identification only and does not imply endorsement by the Public Health Service, U.S. Department of Health and Human Services. Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested Citation Waksberg J, Sperry S, Judkins D, Smith V. National Survey of Family Growth, Cycle IV, evaluation of linked design. National Center for Health Statistics. Vital Health Stat 2(117). 1993. Library of Congress Cataloging-in-Publication Data National Survey of Family Growth, cycle IV : evaluation of linked design. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 117) (DHHS publication; no. (PHS) 93-1391) Prepared by J. Waksberg and others. July 1993. Includes bibliographical references. ISBN 0-8406-0482-3 1. National Survey of Family Growth (U.S) 2. National Health Interview Survey (U.S.) 3. Family size— United States — Statistical methods. 4. Fertility. Human — United States — Statistical methods. 5. Family life surveys — United States. 6. Health surveys — United States. 7. Health Surveys — United States. 8. Sampling Studies. |. Waksberg, Joseph. Il. National Center for Health Statistics (U.S.) lll. Series. IV. Series: DHHS publication ; no. (PHS) 93-1391. [DNLM: 1. National Survey of Family Growth (U.S.) 2. National Health Interview Suirvey (U.S.) 3. Family Characteristics. 4. Research Design. 5. Population Growth — United States. W2 A N148vb no. 117 1993] RA409.U45 no. 117 [HQ762] 362.1'0723 s—dc20 [304.6'2'0723) DNLM/DLC for Library of Congress 93-24687 CIP For sale by the U.S. Government Printing Office Superintendent of Documents Mail Stop: SSOP Washington, DC 20402-9328 Vital and Health Statistics National Survey of Family Growth, Cycle IV, Evaluation of Linked Design Series 2: Data Evaluation and Methods Research No. 117 Research was undertaken to quantify the effects of costs of alternative methods for selecting sample women for the National Survey of Family Growth (NSFG) from the National Health Interview Survey (NHIS). This report presents estimates of the effects of alternative design options, obtained by statistical modeling techniques, for linking the NSFG with the NHIS; the cost data and the statistical precision of estimates were based on data from the NSFG, Cycle IV. The estimated survey costs and projected response rates for alternative linked design options and for the unlinked design are compared for fixed precision. The findings confirm that substantial gains in the NSFG design efficiency were obtained by linking the NSFG sample design to that of the NHIS. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics Hyattsville, Maryland July 1993 DHHS Publication No. (PHS) 93-1391 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jack R. Anderson, Acting Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for ~ Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Office of Research and Methodology Monroe G. Sirken, Ph.D., Associate Director Kenneth W. Harris, Special Assistant for Program Coordination and Statistical Standards Lester R. Curtin, Ph.D., Chief, Statistical Methods Staff James T. Massey, Ph.D., Chief, Survey Design Staff Andrew A. White, Ph.D., Chief, Statistical Technology Staff Preface This report is the third in the NCHS Vital and Health Statistics Series that evaluates the error and cost effects of linking the sample design of the National Survey of Family Growth (NSFG) to the design of the National Health Interview Survey (NHIS). NSFG Cycles I-III had been designed as stand-alone area household sample surveys. The sampling paradigm of the linked NSFG survey design is that the file of names and addresses of the NHIS sample households, including the information collected about them in NHIS, serves as the NSFG sampling frame. The first report, Integration of Sample Design for the National Survey of Family Growth, Cycle IV, With the National Health Interview Survey (Series 2, No. 96), estimated the NSFG design effects resulting from linking the NSFG to the NHIS design instead of designing the NSFG independently. The design effects based on statis- tical modeling investigations were quite encouraging and indicated that linkage was likely to produce substantial gains in the NSFG sample design efficiency. The second report, Linking the National Survey of Family Growth With the National Interview Survey (Se- ries 2, No. 103), presented results that were based on survey experiments in which alternative design options for linking the NSFG to the NHIS were tested. The experi- ments demonstrated conclusively the feasibility as well as the efficiency of linking the NHIS and NSFG sample designs. Although the findings did not lead to a definitive determination of the optimal design for integrating the NHIS and NSFG designs, the findings did provide the basis for making informed choices for integrating the NSFG Cycle IV design with the NHIS. This, the third report, evaluates the linked design that was executed in conducting NSFG Cycle IV. For fixed precision, it compares the actual NSFG Cycle IV costs with those expected for an unlinked design and for two somewhat more efficient linked designs that might have been executed under more ideal conditions. After reading an initial draft of this report, I asked Westat, Inc, to add a section that compared the linkage cost reductions actually realized in NSFG Cycle IV with those conjectured on the basis of the earlier experimenta- tion and to reconcile any differences. The reader is re- ferred to the section ‘Reconciliation with earlier projections” for these comparisons. Suffice it to say that the agreement is really quite close —the linked design was expected to and actually did reduce NSFG Cycle IV costs by about 25 percent. Plans are under way to link the NSFG Cycle V to the 1993 NHIS in much the same way that NSFG Cycle IV was linked to the NHIS. It is not often that researchers have the opportunity to develop and test innovative survey designs to proceed systematically in the manner that was done here —from modeling, to field experiments, and finally to the main survey. Hence, it is noteworthy that this in-depth research effort has been more than justified by the improvements that were achieved in design efficiency and by the analytic enhancement possibilities provided by the merged NSFG and NHIS microdata sets. Congratulations are owed to the Westat staff who prepared these three reports and the NCHS staff who worked with them. The NCHS staff included Andrew White, who provided technical oversight in the prepara- tion of the first report; Deborah Trunzo (nee Bercini), who provided oversight during the entire second phase of the research project; and Steve Botman, who provided oversight during the final phase of the project including the preparation of this report. As noted in the preface to an earlier report, this project would not have been possi- ble without the support and cooperation of Robert Fuchsberg and Owen Thornberry, the former and current directors, respectively, of the Division of Health Interview Statistics, and William Pratt, former chief of the Family Growth Survey Branch, Division of Vital Statistics. Monroe G. Sirken, Ph.D. Associate Director Research and Methodology Contents DE CE ny rel en rr ie Reo i POIs £0 We or ie ia Be ir 3 Heli ott Fe Stew Fleet 57 SN 25s Tn LIS FoR Ser Te 1 SATIRE 10 F nei 0G sez iii IO dUCION vom nem sms rm CP ERs ME HE HE EE RE EE i EAE EE ME AE EE RE ROE ED FE MRC HEART WEA BEE SELB CRY, vs nme eis ma HM SD KEE EEE ENR ABIES AA EEK BE REE OS ANE E DA a BS Chapter 1. Design Of CYClE TV .vvvn vom mtr snm ms seins semen noma mul ensusmss sams snsissesnesssnsss snes SS CHNIIYALY ee te gran risen se eatin el rgd ne pn teva ee ls es ot OE 7 1 wa 2 EE BE FT Fe Bs Design of the National Health Interview Survey ......... iii iii iii Subsampling of NITIS COMPIELEA TVLETVICWS «cm 5 2 55% soo 0.05 #5518 553 06 55 £504 508 5:30 0 5 8058 909 35 50000 4 50 ob 319 8198 6 4 04 309 3: ud Field adjUSLMBIIS «ov iv sis osm mins waives mas oan sie sie a 4/8 0 3 i005 00308 004 00358 900 808 501 50305 300 1s 00a #4 400 39 0 5.9 504 #9 Subsampling for nonresponse fOllOWUP ..... oot Chapter 2. Components Of VAIIANCE. . . o.oo titties eee aren eaaaanaeeannnns TO A] VOATTANICE 5555. 550915 5 6 508 50k 359.8 55.50 50 0 5 00 556 5 005 634 9 31550 0 900 99 0 5006 4 910 9 90 901 4 5 300 1 10 i i 9 NVI -SCTICIIL VATIAIICE «sir viv ws mmo 0000 6 00008 0 08 00 408005800 4 00 0 S000 0 8 SLE 0 A 43 100 4 010 0 a 0 RD EBON ot, ly lore els elenelalatalsrs tle ret eRe) sheers seems sli ers sl wes Eels se We hr A hE SE RE SH B13 EAL REL GE SINOEILITITIE 55.5.5 5:55 53 5150 305 iri 0 se conin 431 13 0 0 0 3084 8 30 1 5 ot 3 3 9 Tf 0 1 ne i lv 8 COVERS 5 vv 000 0 000 900 9000 0 30 50 500 9 50 66 00 SST 0 0 HOHE B00 40% #660600 00 000. 9 0 500 00 OF 0350 0 9 wm 59 wi Chapter 3. Model for components Of VATIBIOE « « «vo vv nvvio satis waar 8 500058803 5 308 59.5 505 99:0 98 5 04 8035 0009000000 9:4 0 TCOrRHCal [OLIN wis nnn ven in a 2 in in Fe in 1 me i en re 1 20 A 0 30 SA To I Boe AS WB WE RITE WB YIP BE 2 5 COCO NINO NPA WLWWLWWW ND = Chapter 4. Design SPECHICALION. : nis wien sis mai ais #0058 508.5805 5 85881850810 2185 5.8 518 16 58 45 os 5088185 008 008.5009 5 08 4140 900.80 0 11 First-Stage COMPONENTS «vv viviwrvws mss maine sama ss sis va ane a8 sa 3a meus sass sass saan ss ass snes ines 11 SCCONA-SIAZE COMMPOMEIIES rei + aii ciniia comers itv e070 00 3301.00 50 0 300 0 4 0 0 00 00 AO AO 90 a 12 Effects Of variation In WEIBIES: «occ o s vi0 55.5 0000050005 95.00 98 050000 0 300 9 950 3100 920 30000 wim 0 wT 90mm 0 300 pi 930 0 mi 13 Projected TESPONSE TALES «ov vv wiv sv sis vive vie moe 30 510 405 30 0.0 013 6305008 0303 315530 35 50 68 90.0 0 9.9 900 0 2 90 0 wo 00 ww 13 Chapter 5. Cost estimates and reconciliation with €arlier DrOJECHONS «sc xs sn sis samsnsss 6555 500 +0 aim bs ste sn wien 15 BO OT OIICEE «0 vv m vw ws wis x oy £30 #0 S00 B40 Si 00 00 0 90 E00 FE Ba) BIW 6B 4 5 BI EF BE DIE 9. 4800 6 BEE ER 6 A 17 Text tables A. Rules for selecting households from the National Health Interview Survey sample for Cycle IV of the National Survey of Family Growth, by year and quarter interviewed in the National Health Interview Survey and race and number of eligible women living in the household. ......... cc. iii iin, 5 B. b parameters for women, by race and components of variance..................... 42.60 nc oun ni 1 rd 0 iin 7 C. TFirst-stace intraclass correlation FOr DEIN IC vu vu vo vem vom oom om 5m 58 5005 205% 5 50555 BIA ® $15 9.50 5x eos 9 D. Second-stage intraclass correlation for Design C......oouiit titi iii ee 10 E. Components of within-segment design effects for Cycle IV, by race ........covviiiiiiiiniiiiniiinnnnnnns 10 F. Comparison of modeled and observed design effects for Cycle IV, by parameter and race.................. 10 pgp vOZE Vi First-stage contribution to design effect for Design A. ........ovivhis iui mines nsisssnsasnnissanns ans 12 Percent distribution and oversampling rate of eligible women by stratum, according torace ................ 12 Second-stage contribution to design effect for DESIGN A «vin sssimnsmsmnsssmnsnsossssmnswssssssionsyns ss 13 Percent distribution of population and retention rate by number of eligible women within household, according EO LACE 4 5 vw 5x 200 0 £10 00% 0050 08 008009 00d 06 35 00 0 506 08 RR LO 9 0 0 BEET HE BE ET 0 eB 0 13 Components of design effects for Design A, by parameter and r8C€ «cv vvsvrvrrvvnvivssvrvsnvinsvvorrres 14 Comparison of National Survey of Family Growth design effects, by race and various designs. .............. 14 Sample sizes for Design A, by race and type of household or respondent .............ccooiiiiiiinnn... 14 Summary design specifications for Cycle IV of the National Family Growth Survey........................ 15 Direct costs for design options in 1992 AOA. +. «cov vvuivvvruinn sna wonnin snsis macs siewoe ss ens esses sss 15 Symbols --- Data not available . . Category not applicable - Quantity zero 0.0 Quantity more than zero but less than 0.05 Z Quantity more than zero but less than 500 where numbers are rounded to thousands * Figure does not meet standard of reliability or precision The National Survey of Family Growth, Cycle IV, Evaluation of Linked Design by Joseph Waksberg, Sandy Sperry, David Judkins, M.A., and Valerija Smith, M.A., Westat, Inc. Introduction The National Survey of Family Growth (NSFG), Cycle IV, conducted in 1988, provides information on childbear- ing, contraception, and related aspects of maternal and child health for women who were of childbearing age at that time. The NSFG sampled women were selected from women who had previously been sampled for the National Health Interview Survey (NHIS). Complete NHIS inter- views included women who responded directly to the interviewer (self-respondents) and those who responded indirectly through proxies. In this report, both groups are referred to simply as “respondents” or as “the interviewed sample.” The NHIS sample women had been interviewed between the fourth quarter of 1985 and the first quarter of 1987. The purpose of the linkage was to reduce NSFG costs while keeping sampling error constant. A report by Waksberg and Northrup (1) had projected that the linked design would be more efficient than the approach of drawing fresh area-based samples, which was the ap- proach used for the first three cycles of the NSFG. Indeed, the achievement of cost or variance effi- ciencies in surveys such as the NSFG was a major goal in the redesign of the NHIS for 1985-94. The 1985 design did not use the decennial Census lists in sample selection that all the other current surveys of the Bureau of the Census use (e.g., the Current Population Survey, the National Crime Survey, the Survey of Income and Pro- gram Participation, and the American Housing Survey). When these lists are used, the sample addresses cannot be released for 70 years to anyone who is not a sworn employee of the Bureau of the Census. By avoiding decennial list-based sampling, the NHIS sample addresses may be made available to non-Census employees under strict privacy safeguards. A drawback for NHIS of using mixed area- and permit-based samples is that the area- based component is much more expensive than a compa- rable list-based component. Another drawback of the mixed area- and permit-based samples is that it is more difficult to avoid double coverage of some housing units with this dual approach than it is to avoid double coverage between decennial list-based and permit-based samples. The obvious question is whether the efficiencies for follow-on surveys to the NHIS pay for the higher cost of the NHIS itself. The purpose of this study was to answer part of this question and to quantify the cost savings of linkage for the NSFG. Although considerable research on this same topic (1,2) preceded the decision to use NHIS as the sampling frame for Cycle IV, such prior research can never quite operate under the actual conditions that prevailed during the data-collection operations. Studies similar to this study will need to be carried out on other follow-on surveys to the NHIS to answer the overall question of whether the area sample of old construction for NHIS better suits the overall program of the National Center for Health Statis- tics (NCHS) than would a list sample of old construction. Strategy The basic strategy was to project the cost for an unlinked NSFG design that would provide the same sam- pling precision and compare it with the cost for the linked design. The first conceptual problem encountered was determining which linked design would be considered. These considerations were brought on by the very special circumstances that surrounded the Cycle IV linked design. The survey was originally intended to be conducted in mid-1987 with women interviewed for NHIS in 1986. However, there were major revisions in the questionnaire late in the survey planning process, so the interviewing was conducted around March 1988 instead of July 1987. These delays raised costs and hurt response rates by requiring more extensive tracing. The other special circumstance was that, as a result of budgetary difficulties, the sample size for the 1986 NHIS was only one-half the intended size. NSFG requirements for black sample sizes could not be met from this reduced NHIS sample. As a result, black women who had been interviewed for NHIS in late 1985 and in early 1987 were added to the sampling frame. This dilemma was resolved by estimating the cost of three linked designs: ® Design B, which assumes that NSFG IV data collection was conducted around July 1987 and not March 1988; ® Design C, which was the actual 1988 design, was conducted around March 1988; and ® Design D, which assumes that the 1986 NHIS was at full strength and that NSFG IV data collection was conducted around July 1987 and not March 1988. These statements do not specify unique designs. Ad- ditional assumptions and estimates of components of variance are required. The same is true for Design A, the unlinked design. Before explaining how the components of variance were estimated, it is helpful to review the designs for NHIS and Cycle IV of NSFG. Chapter 1 Design of Cycle IV Summary The sample for Cycle IV of the NSFG is a subsample of women whose households had participated in the NHIS, a continuous survey of the civilian noninstitutionalized population of the United States. When the full NHIS sample can be used, interviews are obtained at 47,600 housing units each year in a fixed set of 198 metropolitan areas and clusters of nonmetropolitan counties. Data are collected on each household member about disabilities, health conditions, doctor visits, hospitalizations, and other health-related topics. A new set of households is inter- viewed each year. NCHS provided computer files to Westat of house- holds that participated in the NHIS together with address information, rosters, and some basic demographic data on household members. Households were included that had been interviewed for NHIS any time between the fourth quarter of 1985 and the first quarter of 1987, inclusively. From these, Westat selected the NSFG sample. House- holds were drawn from 156 of the 198 primary sampling units (PSU’s) in the NHIS design. In comparison, Cycle III was confined to 79 PSU’s. Spreading the sample across more PSU’s resulted in smaller sampling errors. No more than one woman was selected for the NSFG sample per household. Interviewers attempted to locate these women, following them to new addresses if neces- sary. After locating a sampled woman, the interviewer conducted a brief screener to confirm that she was indeed eligible (between the ages of 15 and 44, inclusively). Design of the National Health Interview Survey The form of the NHIS sample redesign of 1985 (3) made it possible for NCHS to transmit data on NHIS sample households to private contractors for use in con- ducting follow-on surveys, which are then said to be linked to the NHIS. The confidentiality of the transmitted data is protected under section 308(d) of the Public Health Ser- vice Act. The NHIS sample for the years 1985 through 1994 is restricted to 198 PSU’s. These sample PSU’s were se- lected from a much larger set of PSU’s that covers the United States. The sample selection method was based on a stratified probability design. This means that the PSU’s were grouped prior to selection to ensure that the selected PSU’s would be broadly representative of the total U.S. population in terms of several demographic and economic characteristics. Some of these PSU’s are so populous that they were included in the sample with certainty. These are called self-representing (SR) PSU’s. There are 52 SR PSU’s in the full NHIS design. The remaining 146 PSU’s had a chance of not being selected. These PSU’s would thus represent both themselves and other PSU’s that were not selected. Hence, they are called non-self-representing (NSR) PSUs. To allow flexibility to conduct the NHIS with any of several different sample sizes, the PSU’s are divided into four panels, each of which can be used to represent the Nation, if need be. The largest SR PSU’s are in all four panels. Medium-sized SR PSU’s are in two panels. There are 62 PSU’s in a single panel sample, 112 PSU’s in a two-panel sample, 156 PSU’s in a three-panel sample, and 198 PSU’s in the full design. Within each sample PSU, a sample of blocks (or small groups of blocks) was selected. In PSU’s in which black persons constituted 5-50 percent of the population, blocks in enumeration districts with a higher percent of black persons were selected with a higher probability than other blocks. Within each block, a cluster of eight housing units was selected. These housing units were spread as evenly throughout the block as possible. To gain better control over the size of the sample, housing units constructed since the 1980 census were selected through a sample of building permits rather than through area sampling. These units were selected in clus- ters of four instead of eight. To provide continuous coverage of the population throughout the year, the sample of households was spread over the 52 weeks of the year, with each week’s sample a random subsample of the total sample. Each year, a totally new sample of households is selected. However, they tend to be neighbors of the households interviewed the previ- ous year. Subsampling of NHIS completed interviews The procedure for selecting the NSFG sample from the NHIS sample was complex. In this section, the factors motivating the design are described in tandem with the design features themselves. For readers more interested in any effects of the design than in motivating factors, suffice it to note (a) that in some PSU’s, only black women were selected, (b) that neighborhood clusters of black women tended to be. larger than clusters of women of other races, and (c) that households containing more than one eligible woman of a race other than black were selected at a higher rate than households containing just one such woman. This last point considerably reduces the variability in the sampling weights of women of races other than black. The weights of black women tend to vary much more strongly than the weights of other women. The NSFG sample was drawn from women whose households had participated in the NHIS in the fourth quarter of 1985, any time during 1986, or in the first quarter of 1987. Because of a lack of adequate funding, the 1985 NHIS sample was restricted to three panels (156 PSU's), and the 1986 NHIS sample to just two panels (112 PSU’s). Funding was augmented for 1987; thus the 1987 NHIS sample is found in all 198 PSU’s of the full NHIS design. Unfortunately, even combining all six of the avail- able quarters together did not provide as many black women as were selected for Cycle III of the NSFG. The decision was thus made to select as many of these women as possible, subject to operational constraints and the restraint of selecting just one woman per household. The only black women who were not selected were those who resided in the 42 PSU’s that were used by NCHS only in 1987. It was judged that the travel costs per completed interview would have been too high for the women in these PSU’s. Combining all six of the available quarters together provided many more women of races other than black than were required for the NSFG. In deciding how to subsample, the general preference was to take the most recently interviewed because they would be the least likely to have moved since the NHIS interview. (Such a proce- dure does not introduce bias because each week’s sample is a random subsample of the total sample.) It appeared, however, that the household information from the first quarter of 1987 might not be available for sampling in time; therefore, an initial decision was made to restrict the sample to 1986. Subsequently, timing ceased to be as tight and more funding was made available, so the sample of women other than black was expanded to include some from the first quarter of 1987. The first step was to select households. The second step was the selection of persons from households. Rules for selection of households are summarized in table A and listed below: ® All NHIS sample households in the 156 PSU’s used in the fourth quarter of 1985 containing one or more eligible black women were selected. e All NHIS sample households from 1986 containing one or more eligible black women were selected. ® All NHIS sample households from the first quarter of 1987 containing one or more eligible black women were selected if they lived in one of the same 156 PSU’s used in the fourth quarter of 1985. ® All NHIS sample households from 1986 containing more than one eligible woman of a race other than black were selected. e The NHIS sample is split into 52 subsamples corre- sponding to weeks of the year. Households from 30 of these 52 subsamples from 1986 containing exactly one eligible woman who was not black were selected. Drawing fewer of these households than of households containing more than one eligible woman who was not black makes up for the fact that only one of the women in each multieligible household could be interviewed. If the sample had not been selected in this manner, women who are not black from the following sorts of households would have been underrepresented: mother and daughter both between 15 and 44 years of age, sisters both between 15 and 44 years of age, and unrelated women both between 15 and 44 years of age. Note that black women in multieligible households are underrepresented because there were not enough eli- gible black women in the total NHIS sample to allow the subsampling of black women in single-eligible households. For this and other reasons, it is thus crucial for microdata users to use the provided sam- pling weights. e A few NHIS sample households (1 week out of 13) were selected in two of the available panels from the first quarter of 1987 if they met a specific criterion. The criterion was that the household contain exactly one eligible woman who was not black and no eligible black women were selected. ® Households assigned to 12 out of the 13 weekly NHIS subsamples in two of the available panels from the first quarter of 1987 were selected if they met a different specific criterion. That criterion was that the house- hold contain exactly two eligible women who were not black and no eligible black women. eo All NHIS sample households in two of the available panels from the first quarter of 1987 containing three or more eligible women other than black women but no eligible black women were selected. Within a given household, all eligible women had the same probability of selection. The probability of selection was one divided by the number of eligible women. Eligi- bility was defined in terms of exact age on March 15, 1988: A woman had to be 15-44 years of age on that date. There was one minor exception to this rule. Within multiracial households selected from the first quarter of 1987, only black women had a chance of selection. Each of the black women in such a household had the same probability of selection. Field adjustments There were rare instances in which the sampled woman was under age 15, over age 44, or not a woman at all. (NHIS age and sex information were imputed if missing, causing some errors. Even where the data had not been imputed, other errors were found.) In these cases, the interviewer selected among other eligible women then residing in the household. If there were no eligible women, the case was dropped. Subsampling for nonresponse followup After all efforts to complete an interview were ex- hausted by local interviewers, a 50-percent subsample of nonresponse cases was selected for intensive followup. This subsampling, designed to reduce interview costs, was accomplished in two ways. In PSUS for the six largest metropolitan areas, where there were large numbers of nonresponse cases, the nonresponse cases were sequenced by an identification number, and a systematic sample of one-half of them was drawn. The remaining PSU’s were sequenced in descending order by the number of nonre- sponse cases they contained. A 50-percent sample of these PSU’s was selected systematically. Among the selected cases, those that appeared to be convertible were assigned to a corps of traveling interviewers and assistant supervi- sors who had previously demonstrated superior ability in refusal conversion. Prior to the followup, the response rate was 77.9 per- cent. Of the 8,450 final respondents, 220 were obtained as a result of the nonresponse followup. Counting each of these 220 interviews twice, because each woman repre- sents herself and one other woman, boosts the response rate from an unadjusted 80.0 percent (8,450/10,562) to an effective response rate of 82.1 percent ((8,450 + 220)/ 10,562). Table A. Rules for selecting households from the National Health Interview Survey sample for Cycle IV of the National Survey of Family Growth, by year and quarter interviewed in the National Health Interview Survey and race and number of eligible women living in the household Year and quarter interviewed Race and number of eligible women living in the household 1985 fourth quarter 1986 all quarters 1987 first quarter Households selected for the NSFG Black Atleast1................. All in 156 PSU's Allin 112 PSU's All in 156 PSU's All others EXACUY + ur: cor 5 ta weer 4 00 00 1 co None 30 weeks of every 52 in 112 PSU's! 1 week of 13 in 112 PSU's? EXBONY'D. ois von om von mie mont 3 om None Allin 112 PSU's 12 weeks of every 13 in 112 PSU's? Atleast3. ................ None Allin 112 PSU's Allin 112 PSU's The last 30 weeks of 1986. 2The last week of the first quarter of 1987. 3The last 12 weeks of the first quarter of 1987. NOTES: NSFG is National Survey of Family Growth. PSU is primary sampling unit. Chapter 2 | Components of variance To estimate the components of variance, the tech- nique of balanced repeated replications (BRR) was used. General information on this technique may be found in Wolter (4). The actual NSFG sampling strata were not set up in the way required for BRR. To fit the sample into the BRR model, the sample women were regrouped into variance strata and variance units, corresponding to the sampling strata and half-samples of classical statistical theory. Different sets of variance strata and variance units were formed to estimate the various components of variance. Total variance Two types of variance strata were established to estimate total variance. The first set consists of groups of NHIS second-stage units (segments) within SR PSU’s (self-representing primary sampling units). A total of 58 variance strata of this type were formed. Each such variance stratum is a systematic subsample of groups of four consecutive NHIS segments from SR PSU’s. The subsamples are exhaustive and mutually exclusive. Prior to the grouping and systematic subsampling, the segments were sorted by region, PSU, and order of selection (by the Bureau of the Census) for NHIS. The second set of variance strata consists of groups of NSR PSU’s. There were 42 of these strata. Each variance stratum of the second type consisted of between two and five NSR PSU’s from similar sampling strata. The two variance units for each of the first type of variance stratum consisted of a systematic subsample of the NHIS segments in the variance stratum and the complement of that subsample. The sort was the same as for the identification of the variance strata. The pattern of variance-stratum and variance-unit assignment in SR PSU’s can thus be illustrated as: 1A, 1B, 1A, 1B, 2A, 2B, 2A, 2B, .., 538A, 58B, 58A, 58B, 1A, 1B, 1A, 1B, ... . (Consecutive sets of four segments were grouped together into the same variance stratum rather than the more traditional consec- utive sets of two segments because, at the time, it was believed that this might improve the stability of the vari- ance estimator. Subsequent discussions have thrown this thinking into doubt. For future work, consecutive pairs are recommended.) The two variance units for each of the second type of variance stratum were formed by dividing the PSU’s in the variance stratum into two groups. Care was taken to 6 ensure that the exclusively black PSU’s were always in opposite variance units. The perturbation factors were the standard 0 and 2 of BRR for the variance units in the first type of variance stratum. The perturbation factors were 0 and roughly 2 for the variance units in the second type of variance stratum. The deviations from the standard 2 were made to allow for odd numbers of PSU’s and variation in the sizes of the sampling strata represented by the PSU’s in a variance stratum. More detail on the variance stratum-unit forma- tion and perturbation may be found in Judkins, Mosher, and Botman (5). Within-unit variance The variance strata for within-PSU variance were much simpler than those for total variance. All NHIS segments with any women designated for the NSFG IV sample were sorted by region, PSU, and order of selec- tion, grouped into consecutive preliminary variance strata of four segments each, and then systematically subsam- pled into 100 samples. Each sample of preliminary vari- ance strata constituted a final within-PSU variance stratum. Each variance stratum was systematically split into two variance units. The pattern of variance-stratum and variance-unit assignment within PSU’s can thus be illus- trated as: 1A, 1B, 1A, 1B, 2A, 2B, 2A, 2B, ..., 100A, 100B, 100A, 100B, 1A, 1B, 1A, 1B, .... Within-segment variance The variance strata for within-segment variance were formed in a manner similar to the within-PSU variance strata. All designated women were sorted by region, PSU, segment order of selection, and household identification number within segment, grouped into consecutive prelim- inary variance strata of four women each, and then sys- tematically subsampled into 100 samples. Each sample of preliminary variance strata constituted a final within-PSU variance stratum. Each variance stratum was systemati- cally split into two variance units. (As at the segment level, it is now believed that consecutive pairs might have yielded better results than consecutive sets of four women. Also, special modifications for segments with one or three sample women would have improved the within-segment variance estimates.) Replication All stages of adjustment were repeated on each of the 100 sets of replicated weights for each component of variance. This means that nonresponse adjustment and the iterative raking to Current Population Survey esti- mates and demographic control totals was repeated 300 times, i.e., 100 times for total variances, 100 times for within-PSU variances, and 100 times for within-segment variances. Smoothing Direct estimates of between-PSU variance and within- PSU-between-segment variance could have been com- puted for each characteristic of interest. Past experience has demonstrated, however, that these individual esti- mates of the “between” components are highly unstable. In fact, negative estimates of variance can occur. To counteract this, separate generalized variance functions (GVF’s) were fit for total, within-PSU and within-segment variances, and components of variance were estimated by subtracting b parameters rather than individual estimates. This smoothing of the variances prior to subtraction im- proves stability and usually results in better overall esti- mates of all components (6). As input to the fitting of GVF’s, variances were estimated directly for a large num- ber of characteristics. The characteristics included cross-tabulations of age by marital status; education by type of current contracep- tive; usual source of family planning services by metropol- itan status; religious affiliation by ever had intercourse, by expectations for additional children; parity by fecundity; age by relative order of marriage and first birth; Hispanic origin by age at first marriage; age by education by age at first intercourse for never-married women; Hispanic origin by living arrangement at age 14 by current contraceptive use for never-married women; religion by metropolitan status by ever-usage of family planning services for never- married women; and education of mother by ever had intercourse for never-married women. There were other tables as well. All the tables were repeated for black women by themselves. The curves were fit to the formula V2=b(1/X -1/T) (1) where V2 is the directly estimated relative variance for a statistic, X is that statistic, and T is the upper bound on the possible value for X. The results are shown in table B. Table B. b parameters for women, by race and components of variance Components of variance Black All races TAL wos iin was ma wmani diese 4,906 10,192 WHRINPSUY.. . iso 2 vs 255 508 we 4,350 9,500 Within-segment. . . . ......... 4,258 9,453 1PSU is primary sampling unit. NOTES: These b parameters are slightly different from those in Judkins, Mosher, and Botman (5). A different model was used here with a different set of items to improve comparability between total, within-PSU, and within-segment variances. Caveats These estimates imply a 7-percent between-PSU vari- ance for all races combined and an 11-percent between- PSU variance for black women. They also imply that between-segment variance was a trivial component of variance for most estimates. Although we believe that the general direction of these estimates is correct, several caveats are in order. A number of approximations were made in the vari- ance computations, almost all of which tended to exagger- ate the between-PSU component. The estimates of between-PSU variances are thus upper bounds. First, there are some very large sampling fractions at the first stage (some as large as 0.75). These have a strong dimin- ishing effect on between-PSU variance that was not re- flected in the variance estimates. Second, even though some small adjustment was made to compensate for the bias in the collapsed-stratum variance estimator, such an estimator is normally positively biased. This tends to exaggerate between-PSU variance. Third, the PSU’s of only black women lead to even worse positive biases because there is only one such PSU in every other stratum. Fourth, the variance strata for within-PSU variance prob- ably should have been assigned on the basis of all NHIS segments, not just those with a woman designated for an attempted NSFG interview. This has the effect of exagger- ating between-PSU variance and understating between- segment variance. (However, because poststratification by race and age essentially makes all inference conditional on the achieved national sample sizes, this is probably a very minor caveat.) Fifth, the within-PSU variance strata pair segments in different PSU’s when there are PSU’s with odd numbers of segments. This fifth caveat has the oppo- site effect of the fourth caveat but is probably smaller. Sixth, the within-segment variance strata pair women in different segments or even PSU’s when there are segments with odd numbers of designated women. This has the effect of exaggerating within-segment variance, thereby understating between-segment variance. Chapter 3 Model for components of variance Theoretical form A fairly good model for the relative variance of a ratio-adjusted estimator from a three-stage design such as NSFG is V2=V2[1+Pp A +p(\-D)Es + Ex 1 +a +02 +c] (2) where V2, is the relative variance that would be obtained from a simple random sample of women of the same size, P is the percent of women in non-self-representing PSU’s, p; is the intraclass correlation at the PSU level, \,; is the average number of sample women per PSU, Pp); is the relative increase in variance due to sampling at the first stage, p, is the intraclass correlation at the segment level, \, is the average number of sample women per segment, &, is a complex term due to variation in the number of eligible women per segment that is usually greater than one and that persists even with poststratification but is difficult to estimate, p,(A,—1)&, +&,-1 is the relative in- crease in the variance due to sampling at the second stage, a’ is the relative variance in within-household inverse probabilities of selection, b* is the relative variance of inverse probabilities of selection across second-stage strata, and c¢? is the increase in relative variance of the final weights due to subsampling of nonrespondents, adjust- ment for nonresponse, and other adjustments. Following common practice, the term §, was not directly estimated for this study. This type of factor is discussed at greater length in Hansen, Hurwitz, and Ma- dow (8) volume 1, chapter 6, section 8; volume 1, chapter 8, sections 1 and 11; and volume 2, chapter 8, section 4. There it is written as VV %/V2 There is also more discussion in an appendix to this report. The formula used by Waksberg and Northrup was slightly different. They used Vi=V2_ [1+Pp A +p(\-1)+ Vi, + a? +b?) (3) srs where V3, is the relative variance in the number of eligible women per segment, also not directly estimated. It might be argued that the omission of c? was an oversight, but other studies have indicated that poststratification can have variance-reducing properties that directly counteract some of this effect. Empirical studies in the last decade have indicated that V2, is usually on the order of 0.1 or 0.2 for area samples of persons through their housing units. Such values are too large to be reasonable. A substitution was made for the &,-1 term, which is expected to be considerably smaller than V3,. Note that Hansen, Hurwitz, and Madow also has a formula using V'%,, but that the form is slightly different. Adapting from that source to the problem at hand, the approximation from Hansen, Hurwitz, and Madow would appear to be V2=V2 [1+Pp\& + pa(\-1)E, + E-1 +a? +b% +? srs +V2,/m, 4) where m, is the number of segments. It is also important to note that the right bracket has moved, leaving the term involving variation in segment size as an absolute rather than relative term and that the term declines in impor- tance as the number of segments increases. Based on the discussion in Hansen, Hurwitz, and Madow, this formula is more appropriate for modeling the relative variance of an estimated total that was obtained with simple inverse- probability sampling weights than for modeling the rela- tive variance of an estimated total that was obtained by poststratified sampling weights. Poststratification largely eliminates the additive V},” effect, leaving only the multi- plicative factor &,. To maintain consistency with Waksberg and Northrup, a value was picked for £&, that had the same effect on the design effect as the IZ, that was assumed. A value of £,=0.042 is comparable to the V/2,=0.05 that they as- sumed. A strategy for estimating &, in future studies is given in Appendix II. Fitting the model for components of variance The intraclass correlations implied by these variance components are not very consistent with those found by Waksberg and Northrup when examining Cycle II. This was true even though there was considerable overlap in the table structures that served as the basis for inference about components of variance. (For example, both studies estimated components of variance for the estimated num- ber of black women who used contraception.) Of course, in addition to all the caveats mentioned previously, there are problems in translating the b parameters into compo- nents of variance. First-stage components In this section, values for P, \;, and p, are obtained. There were a total of 156 PSU’s with one or more designated sample women for Cycle IV. Of these, 130 were NSR PSU’s. However, no attempt was made to interview in several of these PSU’s because the sample sizes simply did not warrant the tremendous travel ex- penses associated with the PSU’s. The number of NSR PSU’s with at least one completed interview of a black woman is 88. The number of NSR PSU’s for all-race estimates is 112, and the number for other-than-black estimates is 88. Table C shows, by racial grouping, the numbers of NSR PSU's, the complete sample sizes in NSR PSU's, the proportions of the eligible populations in NSR PSU's, and the resulting values of \;. Substitution of these parameters into the formula b 0 WP P1™ por PN, 5) where by is the b parameter for total variance and byp for within-PSU variance, gives the estimates of intraclass correlation, py, at the PSU level shown in table C. These do not line up very well with the Waksberg and Northrup estimates also shown in table C. Intraclass cor- relation should be about the same for all races as for all races other than black by themselves. It is thus particularly troubling that the estimated correlation for black women in Cycle II was much lower than that for persons other than black but that this relationship is reversed in Cycle IV with the estimated correlation for black women being considerably larger than that for all races. The difference in smoothing techniques is part of the reason for the disparity. The Cycle IV smoothing technique was applied to the Cycle II variances rather than averaging direct- difference estimators of between-PSU variance. This re- duced the discrepancy somewhat in favor of the Cycle IV estimates, particularly for black women; hence the lean toward the Cycle IV estimate for black women. Finally, the Cycle IV estimates of between-PSU variance were positively biased for the reasons given earlier in this section (whereas the Cycle II estimates were not) and that the Cycle IV estimates were prepared with more PSU’s than the Cycle II estimates and should thus have better stability. However, there was concern that there was still a Table C. First-stage intraclass correlation for Design C large variance in our variance-component estimate, and thus it was assumed that the single intraclass correlation of 0.005 shown in the “composite” column of table C for all races combined, races other than black by themselves, and black women by themselves. Second-stage components Similar problems and inconsistencies beset the deter- mination of the intraclass correlation at the second stage (within segments). Table D gives, by racial grouping, the numbers of NHIS segments with one or more completed interviews, repeats the completed sample sizes, and gives the average segment sizes (\,). The standard formula for intraclass correlation at the second stage is p2= (bwp/ bws —1-&)/ [(A-1)&;] (6) Substitution gives the estimates shown in table D. The estimates obtained by Waksberg and Northrup are also shown. It is apparent that the estimate of &, is too large for this study or that byp/bws has not been well estimated, it is simply not reasonable to assume a negative intraclass correlation. The very small number of completed inter- views per segment probably led to substantial contamina- tion of bys by between-segment differences. Partially counteracting this was the fact that women other than black selected from the 1985 and 1987 NHIS samples were in segments adjacent to those selected from the 1986 NHIS sample. This was reflected in the direct estimates of variance components and smoothing but was not reflected in the computation of the number of nonzero segments. If allowance had been made for the neighboring segments, this would have reduced the effective number of segments, thereby boosting the average cluster size and bringing the estimated intraclass correlation closer to zero. In trying to reconcile these two sets of estimates, it must be kept in mind that the average number of completed interviews per segment for Cycle II was much larger than for Cycle IV, even though the typical land area was similar. The larger number of completed interviews means that segments consisting of just a single woman were rare. That rareness allows for more accurate estimation of the intraclass correlation. Thus, the Cycle II estimates were weighted more heavily in coming up with a composite intraclass correlation. Non-self-representing primary sampling unit Average Number non-self- with Number of Proportion representing nonzero interviewed of workload Cycle IV Cycle Il Composite Race sample women population’ (A) (p1) (1) (1) Alraces ......cupmsn~ 112 4,693 0.69 41.9 .0025 a .0050 Blagk c we.wio wns di mesmo 88 1,362 0.58 15.5 .0142 .0007 .0050 Other than black . . . . . .. 88 3,331 0.71 37.9 a .0066 .0050 Not actually estimated from the National Survey of Family Growth. Estimated from 1991 Westat Master Sample. Table D. Second-stage intraclass correlation for Design C Number Segment size Effect of Effective Interviewed variation Cycle IV Cycle Il Composite Race segments women Average (A;) (&) (p2) (p2) (p2) ARYBEES oo vs mip woes 3,143 8,450 2.69 1.042 -0.02 -—— 0.03 BISck «ows 218 we paz ak 1,056 2,811 2.66 1.042 -0.01 0.042 0.03 Other than black . . . . . .. 2,382 5,639 2.37 1.042 -— 0.046 0.03 - Table E. Components of within-segment design effects for Cycle IV, by race Other All than Component races Black black Within-segmentb . .......... 9,453 4,258 -—— Samplinginterval. . .......... 6,852 2,732 —-_— Within-segment design effect . . . . 1.38 1.56 —— BAR o sp sate nn me sot mE Br 0.27 0.55 0.07 dv be 4 ole ho BE Bak Tp 0.12 0.00 0.09 Relative variance in weights Before nonresponse sampling . . 0.27 0.55 0.07 After nonresponse sampling . . . 0.39 0.64 0.17 Finalwelghts., ... « «aarp 5s —— 0.51 0.16 Effects of variation in weights The b parameters for within-segment variance in table B are still larger than what would be expected from a simple random sample. Using the sampling intervals shown in table E gives the estimated within-segment design ef- fects also shown in table E (deff=>5/SI). (The design effect [deff] is the ratio of the actual variance of a sample to the variance of a simple random sample of the same number of elements.) These within-segment design effects are explained by the variation in weights as is theoretically to be expected. As explained above, the effect of the varia- tion in weights can be decomposed into three terms: a? is for the subsampling of women within households contain- ing multiple eligible women; b* is for the oversampling in the NHIS of neighborhoods with a high percent of black persons; and ¢” is the subsampling of nonrespondents at the close of regular interviewing for the special followup effort and the adjustments for nonresponse and poststrat- ification. The terms a? and b* were not separately esti- mated for Cycle IV. The term ¢” is equal to zero for black women because the relative variance in the weights of black women was actually smaller after subsampling for nonresponse, adjusting for nonresponse, trimming, and poststratification than it was prior to these steps. This is shown in more detail in the final rows of the table. Comparison of fitted model with empirical design effects The fitted model was used to recompute the achieved precision for Cycle IV. The purpose of this was not to 10 provide an improved estimate of the precision because the direct empirical measures of precision are probably better. Instead, the purpose was to provide a more even playing field for the comparison of linked and unlinked designs. Table F shows, by racial grouping, all the parameters of the fitted model for design effects along with the direct design effects. It also shows the effective sample sizes (nominal sample sizes divided by their respective design effects). Table F. Comparison of modeled and observed design effects for Cycle IV, by parameter and race Other All than Parameter races Black black Boi: dais sn 4565s aaE mE REDS 0.69 0.58 0.71 Number of NSR PSU's . ....... 100 85 88 I pp 58.3 19.2 455 Bf 4 5 S1h EG 3 So 8 BE RE AE a 0.005 0.005 0.005 Ret als Pid fe le AT wants 0.20 0.06 0.16 Number of nonempty segments . . 3,143 1,056 2,382 Xizis 2 60% a 60 4 Binh 9 BBA s 2.69 2.66 2.37 Da) 5 5 at ws 5 BE Fos 4 be Hd OE BE BE 0.03 0.03 0.03 Bao x 00k fs 5 Bre mk BRE 1.042 1.042 1.042 palA-NEa +E... ohh 0.09 0.09 0.08 BEEBE. wiv vo is ow 4 wari +» 0.27 0.55 0.07 OF, vine vr evi ws 40515 0 5 Tn 0.12 0.00 0.09 Modeled design effect . ....... 1.68 1.70 1.40 Observed design effect. . . ..... 1.49 1.80 -— Number of interviewed women . . . 8,450 2,811 5,639 Modeled effective sample size . . . 5,029 1,654 4,027 NOTE: NSR PSU's are non-self-representing primary sampling units. Chapter 4 Design specification Various aspects of required sample sizes are worked out for each design in this section. The goal was to keep precision at the level of the modeled effective sample sizes shown in table F and then to determine the cost penalties or savings associated with the alternate designs. To deter- mine the cost of an alternate design, it was necessary to know the number of PSU’s, the number of segments, the number of designated households, the number of desig- nated women from those households, and the number of interviewed women. These critical statistics are shown in table O for each of the designs. Design C—Most of the parameters for Design C (the actual Cycle IV design) were already set in table F. The response rate (among those who responded to the NHIS) was 82 percent. Even though original design specifications called for a sample mix nearer to parity between black women and women other than black, this will serve as the baseline reliability for the evaluation. (Specifications for future designs may repeat this, establish requirements for women of Hispanic origin, expand the eligible universe to women 54 years of age, or utilize other criteria.) The other designs have been specified in a manner to provide approx- imately the same reliability. Design B—This design is a slight variation on design C. Going to the field within 7 months of the close of NHIS interviewing for 1986 would have improved response rates a little and would have lowered costs. Projections of these improvements are given in the final section. From the point of view of statistical design, the only difference between Designs B and C is that B requires a slightly smaller designated sample size than C. There is no reason to think that design effects would have been affected at all. Design D—This design is similar to Design B, but some reduction in design effects could be expected. First, the sample could be spread out across 146 NSR PSU’s instead of just 100 or so and across 5,000 segments (of the total 8,200 in the full NHIS), instead of just 3,200. Second, a greater proportion of the population would be covered by SR PSU’s. Third, because there is a larger sample frame available, it would be possible to oversample multi- eligible households even more sharply than in Cycle IV. It should be easy to reduce the total design effect for statistics for women other than black women from the modeled 1.37 to 1.3. A design effect of 1.15 is probably a lower limit, given subsampling of nonrespondents, but 1.3 is conservative. Similarly, the increase in the numbers of sample PSU’s and segments and the reduction in the need to take all households containing just one eligible woman will serve to reduce the design effect for statistics for black women from the modeled 1.67 to 1.6. A further consider- ation making these reductions in design effects plausible is that there is variation in NHIS baseweights resulting from differential sampling by minority density stratum. Much of this variation for single-eligible households could be re- moved during the subsampling of NSFG from a full NHIS. These design-effect reductions imply that the sample sizes in terms of completed interviews could be reduced from 5,639 and 2,811 to 5,351 and 2,693 for women other than black and black women, respectively. A reasonable rounded total is thus 8,000 completed interviews instead of 8,450. Design A—The greatest uncertainties of design speci- fication pertain to the unlinked design. It is not appropri- ate to simply compare effective sample sizes from Cycle II or Cycle III with those from Cycle IV because the require- ments for sample allocation were sharply different. Nor can the estimates of design effects for the unlinked design in Waksberg and Northrup be used because there have been changes in reliability requirements. Rather, it is necessary to start with the components of variance derived above to build a new design. The same models are used as described in the section “Theoretical form,” but new parameters are obtained. The parameter derivation fol- lows the same structure as described under “Fitting the model for components of variance.” First-stage components The Cycle III model had 80 PSU’s, 60 of which were non-self-representing. Some of those contained no or very few interviewed black women. A rough projection of the effective number of PSU’s for black women is 30. The number of interviewed women was obtained as the result of a directed-iteration approach that gave the desired level of precision taking into account all contributions to design effects. Derivation of the effect is shown in table G. “Between-stratum” is an approximation of the size of a component that would exist in an unlinked design but did not exist in the Cycle IV linked design. In Westat’s typical unlinked design for NSFG, there are 60 NSR PSU’s selected from 40 NSR strata. This 1.5 PSU’s per stratum design arises from Westat’s desire to have a 11 Table G. First-stage contribution to design effect for Design A Average Contribution Effective Number of Proportion NSR to design Between- number of interviewed in NSR workload Composite effect stratum Race NSR PSU's women PSU's! A) p) P(A=1)p, component ATACES » 5 1 Sao s 5 & 60 11,763 0.73 143.1 0.005 52 .02 BIBER + 5 nam 5 seams 4 are 30 3,180 0.61 64.7 0.005 .20 .02 Otherthanblack . . . ...... 60 8,582 0.75 107.3 0.005 .40 .02 From Waksberg and Northrup, table 13 (1). NOTE: NSR is non-self-representing. PSU is primary sampling unit. flexible design that can be fielded with either 40, 60, or 80 NSR PSU’s in addition to a constant 20 SR PSU's. However, this feature introduces an additional component of variance known as between-stratum variance. It has never been measured for NSFG characteristics but is assumed to be small. For purposes of this evaluation, Westat’s estimate is that it adds 0.02 to the typical design effect. Second-stage components Black women were oversampled in Cycle III by over- sampling block groups that had a fairly high percent of black persons in the prior decennial census and then by screening the listed households to identify those with black female occupants. To determine the within-PSU design effects, it is first necessary to work out the distribu- tion of the sample across the second-stage strata. The within-PSU design starts with the assumption that four strata would have been used, defined by density of the black population within the block. These strata are re- ferred to as “old construction strata” because the area sample would have been used mainly for residents of buildings constructed prior to the previous census. Addi- tionally, because the survey took place near the end of the decade, a separate stratum would have been established for new construction built in localities that issue building permits. The housing units in this stratum would have been sampled by sampling the permits. Also, because the interviewing took place mainly in the spring, it would have been more efficient to interview college women in their dormitories rather than at home. (NHIS also has a dormi- tory sample.) The proportions of black, other than black, and all race women in each stratum shown in table H are based on 1980 decennial census counts with some ad hoc adjustments. The distribution of the population across these strata is subject to seasonal, decennial, and long- term fluctuation. The oversampling rates are relative to the base sampling rate. They were determined following rules that have been demonstrated in the past to yield near-optimal allocation. It was assumed that screening would be used in the other strata to create a uniform probability of selection for all households containing only eligible women who were not black. Thus, for example, two-thirds of the households with only white and Asian women in old construction stratum three would be dropped from the sample after screening. Although it would be possible to use screening to sample black and other women at different rates in all the strata, it was assumed that all households containing eligible women in strata with oversampling rates of one would be retained in the sample. Screening for black women in these strata is a very expensive proposition. Having fixed the allocation across the strata, it is then necessary to fix the overall number of sample segments. The choice of 1,718 segments was made using the follow- ing considerations. About 1,550 area segments and 150 permit segments were selected in Cycle II (8). About 18 dormitory segments were selected in Cycle III. (There were no dormitory segments in Cycle II because unmar- ried nulliparous women were not eligible. There were no permit segments in Cycle III because it was so soon after a decennial census. There were only one-half as many area segments in Cycle III as in Cycle II because of greater attention to small domains for which intraclass correlation is not as troubling.) Table J shows some projections of Table H. Percent distribution and oversampling rate of eligible women by stratum, according to race Percent Total Black Other Percent Oversampling Stratum population population population other rate Total iv coi ssp ams 2b be 100.0 100.0 100.0 Old construction: Less than 10 percent black . . . . 66.0 14.7 74.0 97.3 1 10-30 percent black . . ...... 5.0 11.0 4.5 78.1 1.4 30-60 percent black . . ...... 4.0 16.0 2.4 52.0 3 60 percent or more black . . . . . 7.0 44.0 0.7 87 4 New construction. . . ......... 15.7 12.0 16.1 88.9 1 College dormitories . . . ....... 23 23 23 86.7 1 Total population in millions. . . . . . 57.9 7.7 50.2 12 Table J. Second-stage contribution to design effect for Design A Effect of Effective Number of Average variation in Contribution to number of interviewed segment segment Composite design effect Race segments women size (Ap) size (&) (pa) (Aa=1)Ax&, + &—1 ANIACES worms snes 1,714 11,763 6.9 1.28 0.03 0.23 Black : ws snsnie amcws 2 809 3,180 3.9 1.28 0.03 0.11 Other than black . . . . . . . 1,329 8,582 6.5 1.28 0.03 0.21 how many of each of these types of segments would have at least one interviewed woman by race. The table also shows the contribution to the design effect that is implied by the resulting average segment sizes. Effects of variation in weights The design effect resulting from the oversampling of strata with high percentages of black persons was calcu- lated by the standard formula: b’=(2p;k;) (Sp; /k;)-1 (7) where p; is the proportion of domain population in a stratum and k; is the oversampling rate for the stratum. This formula yielded values for b* of 0.43 for black households and 0.15 for total households. The value of b* for households is 0.0 because subsampling of households other than black discovered in screening would return the other-than-black household sample to a self-weighting sample. Table K gives the population distributions across house- hold sizes where size is defined as the number of age- eligible women living in the house. These distributions come from Cycle III estimates (derived from table 2-7 in the National Survey of Family Growth Cycle III, final report (9). The retention rate is the proportion of eligible women who would be designated for extended interviews. Given the NSFG rule of one respondent per household, the rates are obvious. The parameter a’ reflects the increase in design effect due to subsampling within house- holds as before. The total design effects are obtained by summing the components derived in this section. Table L gathers all the numbers together in the same format as table F for ease of comparison. Note that the projected effective sample sizes do not quite line up between tables F and L. It is considerably more difficult to achieve the desired preci- sion for black women in an unlinked design than for women other than black or total women. The compromise sample size for the unlinked design yields lower precision than the linked design for black women and higher preci- sion for women other than black and total women. The only way to have matched precision by race would have been to use screening even in the stratum with the lowest concentration of black population. Comparing these projected design effects to other NSFG design effects in table M, it is interesting to note that the unlinked design effect for all-race estimates is projected to be larger than that for black estimates, which is the reverse of what was observed with the actual Cycle IV design. This reversal is mainly the result of the fact that with the linked design, it was possible to oversample multi-eligible households and thus reduce the a” to near zero for persons other than black. The design effects projected in Waksberg and Northrup for an unlinked Cycle IV are not strictly comparable because they as- sumed different within-household sampling rules, total sample sizes, and analysis domains. Nonetheless, they are not too far off. (The smaller design effects projected for never-married women other than black are the result of the smaller cluster sizes that are naturally obtained for smaller domains.) Also interesting to compare, the Cycle IIT design effects are much larger, particularly for all races. The main reason for this is the extensive and inefficient oversampling of white teenagers that was con- ducted in Cycle III. Also a factor was the differential sampling of ever-married and never-married women. Projected response rates Having calculated projected design effects, it is possi- ble to calculate the number of women that would have to be interviewed in an unlinked design to attain precision comparable to the linked Cycle IV design. The next step is to work backward to determine the numbers of housing units that would have to be selected from area-listing worksheets. Table K. Percent distribution of population and retention rate by number of eligible women within household, according to race Retention Number of eligible women within household Total Black Other rate JO si ioam sd niMtim mpm Es sma R a oa 100.0 100.0 100.0 FE Tih RAN LAE RI BEER IB RAMS SHAR tHE 66.8 56.9 70.5 1 LIT IE I TT 23.5 28.8 21.5 0.5 Be IE rR AEA PER SE ES ER SR AR ae 8.0 11.0 6.9 0.33 AOTINOIS +. ssn smaswembemsionvws vasa 1.7 3.3 13 0.2 BR 3 rn 6 £80 0 BBE 8 RR 8 BE RE BHR 0.19 0.24 0.17 13 Table L. Components of design effects for Design A, by parameter and race Parameter All races Black All others Priests sain nmin mn ame mn bie ae aes 0.73 0.61 0.75 Number of NSRPSU's . . ................. 60 30 60 NT 0 ai vs te 4 se eee re 1 tet Bh in Bo 143.1 64.7 107.1 DOF oo 4 sare oni von om ov 0m som on Bo eR es 3 oh a 8 0.005 0.005 0.005 PRAY 00 4 1 8 00 100 8 10 ip 541 dom 50k 5 cs le 2 01 3 v0 0 cd 0 +0 Be 0.52 0.20 0.40 Between first-stage stratum . . .............. 0.02 0.02 0.02 Number of nonempty segments . ............ 1,714 809 1,329 N55 20% hh G5 GF BD EE a 6% 0 3 BE 8 10 BR ok Tg 6.9 3.9 6.5 Bar 2 5 od Bh OES 50 2 B08 08 BH of ck BE ok 8 0.03 0.03 0.03 E85 de oo BS BR 5 Ee a RE ele ESTE hes BEE Lal 1.28 1.28 1.28 Pale Dearly sss animssnsnosms vas 0.23 0.11 0.21 Bs 0 Eh EE BE RE LEE ho 0 ER Ed 0.19 0.24 0.17 Bc ui ions on PG 0 055 0 RIE A 0 BLE HS 0.15 0.43 0.00 HEHDE rs mar oie BRIE Gk ET WE AE PE EE 0.34 0.67 0.17 Sn TRE FETE Be Tre AEE ETE AE he 0.12 0.00 0.09 Modeled design effect . .................. 2.23 2.00 1.89 Number of interviewed women . . . . .......... 11,763 3,180 8,582 Modeled effective sample size . ............. 5,276 1,590 4,541 NOTE: NSR PSU is non-self-representing primary sampling unit. Table M. Comparison of National Survey of Family Growth design effects, by race and various designs Unlinked Waksberg-Northrup Linked Cycle IV projections Ever- Never- New Actual Race Original! Revised? Modeled? married married projections Cycle lll ATACES + + wr vs sna nwa vi ws 1.56 1.49 1.68 va ves 2.23 3.00 BAK « x0 svn anions wenn 1.90 1.80 1.70 2.00 1.98 2.02 2.83 Other than black 1.40 1.79 1.43 1.89 1As reported in Judkins, Mosher, and Botman (5). 2s implied by table A of this report. 3As constructed in table F of this report. Experience from Cycle II has shown that about 16 per- cent of listed units will turn out to be vacant, built since the last decennial census, or not currently intended, ready, and fit for human habitation (7). Of the housing units that are occupied, a certain percent will resist being inter- viewed (either passively through not answering the door or actively by refusing). In Cycle III, 5 percent resisted the screening interview that was used to determine race and sex of occupants, and 16 percent of the remainder resisted answering the detailed questions (8). Having fixed design effects and response rates, the entries in table N illustrate the designated number of addresses that would have to be visited, the numbers of these that could be expected to be ineligible for screening, the number of occupied eligible households, the number of screener nonrespondents, the number of successfully screened households, the number of women to be desig- nated from these households for the extended interview, the number of actually interviewed women, and the effec- tive sample sizes. Table N. Sample sizes for Design A, by race and type of household or respondent Household or respondent All races Black All others Designated addresses . ................ 40,800 -—— -_— Less vacant or uninhabitable addresses . . . . . -6,500 —-—— p—— Households . . « sco rnsmo rm smmens wy cme 34,300 —-——— —— Less screener nonrespondents . . ........ -1,700 -——— —-—— Screened households. . . ............... 32,600 —-— ie Number of designated women . . . ......... 14,000 3,800 10,200 Less nonrespondents . . .............. -2,200 -600 -1,600 Number of interviewed women . . . ......... 11,800 3,200 8,600 Effective sample size’. . . ............... 5,300 1,600 4,500 Taking design effects into account. Chapter 5 Cost estimates and reconciliation with earlier projections Table O summarizes the design specifications that were developed in the previous section. These features served as the basis for the cost estimates. Table O. Summary design specifications for Cycle IV of the National Family Growth Survey Design D Design B linked to linked ~~ Design C full NHIS' Design A without Cycle IV without Specification unlinked delays linked delays Primary sampling units . . . . . . 80 156 156 198 Segments . ............. 1,700 3,100 3,100 5,000 Designated households. . . . . . 34,300 awit wars — Designated women . . ...... 14,000 10,000 10,300 9,500 Interviewed women . . ...... 11,800 8,450 8,450 8,000 National Health Interview Survey. Table P shows cost estimates for four possible designs for the NSFG that would approximate the same precision. Because the purpose of this cost analysis is to show a comparison of costs for the different designs, it did not seem necessary to show indirect cost markups. Therefore, all cost estimates shown consist of direct costs only. All cost estimates assume that the main study data collection would be conducted in 1992. Pretest costs have not been included for any of the designs because it is assumed that the cost of a pretest would be the same regardless of the design chosen. Costs Table P. Direct costs for design options in 1992 dollars for data handling for all designs were calculated assuming that the NCHS computer would be used. For Designs A and C, the response rates are those actually experienced for Cycles III and IV, respectively. The response rates shown for Designs B and D were calculated by estimating the number of women who had moved in Cycle IV but would not have moved prior to the interview had it not been for the delay; it is also assumed that these women would have cooperated at the same rate as the rest of the sample. In Cycle IV, 35 percent of women had moved between the NHIS interview and the first attempt at the NSFG interview. According to infor- mation gathered about those who moved during Cycle IV, only 30 percent would have moved by the time of interview had it not been for the 8-month delay. Waksberg and Northrup predicted a cost savings of 28-35 percent (relative to the Cycle III model) for a linked design with one-time interviewing of designated persons (with tracking), depending upon whether 200 or 100 NHIS PSU’s were used. The new report indicates a cost savings of 22 percent (table P, Design C versus Design A). The reasons for the difference are complex, involving changes in the objectives for Cycle IV, the procedure for oversampling black women in the 1985 redesign of the NHIS, cuts in the sample size for the 1985 and 1986 panels of the NHIS, the lag between NHIS and NSFG IV interviews, and improvements in the estimation of vari- ance components. Unavoidably, variance on the variance and cost estimates also plays some role. Design B Design D linked linked Design C to full Design A without Cycle IV NHIS without Cost item unlinked delays linked delays Total css nsnssminnnms sos 4,082,000 2,972,400 3,192,200 2,945,400 Professional labor . . . ......... 565,200 483,000 510,700 483,000 Clorical 1abor : +z 55 5 ss ems ws u9 470,800 350,600 373,300 315,400 Field BBO; » + ws vie wae ws as mE 1,125,200 808,300 874,400 787,300 Travel, « svms we ga: us SEs os os 838,800 497,100 537,600 527,900 Other direct Costs ++ «ov vw vow ve 1,082,000 833,400 896,200 831,800 Number of screeners . . . ....... 32,600 0 0 0 Number of interviews . . . ....... 11,800 8,450 8,450 8,000 Estimated response rate (percent) Screener . ............... 95 Go) ® ® IBHBIVIBW.. 1. « wo. « xt 1 i ivice io: on ws wz 0 om 84 84 82 84 National Health Interview Survey. 2Screener response rate comes from NHIS. 15 The impact of every factor was not quantified. Some were quantified and others were simply listed with an indication of the rough order of importance. Major questionnaire revisions late in the survey plan- ning process, resulting in postponement of the start of data collection, were the main cause of the downward revisions in the cost savings of a linked design. If a comparison is made between Designs B and A instead of C and A, there is a cost savings for the linked design of 27.2 percent instead of 21.8 percent. NHIS sample cuts also contributed. If the 1986 NHIS had been 100 percent and if the questionnaire had been approved in a timely fashion, the savings would have been an even better 27.8 percent (comparing Designs D and A). Another major factor was a suboptimal procedure for oversampling black women in the redesigned NHIS (1985-94). At the time that Waksberg and Northrup prepared projections of design effects for the linked de- sign, this factor was not anticipated. Waksberg and Northrup projected a design effect for black women of around 1.5 for a 2-year NHIS sample with 100 to 200 PSUs. The actual design effect for black women (table M) was 1.8. At least some of the difference is the result of the artificial limitations placed on the NHIS oversampling procedure (10), although improvements in variance esti- mation methodology in the current report or variance on both sets of variance estimates may also play a role. Finally, when Waksberg and Northrup were writing their report, the then-current objectives for Cycle IV were much more stringent than final objectives. Specific reliabil- ity targets were set for ever-married and never-married women by race (black and other). The final objectives had specific reliability targets only by race, and these targets were considerably more relaxed. Having a single reliability target by race reduces the amount of screening necessary in an unlinked design and thereby reduces the savings of a linked design. Waksberg and Northrup called for a total sample size of roughly 10,500 interviewed women to achieve the then-current reliability targets. When the linked sam- ple was finally selected, the goal was to obtain 8,500 16 interviews. Waksberg and Northrup thought that their results were fairly robust with respect to overall sample size, but in retrospect, it seems likely that the smaller sample size caused some inefficiencies of scale in the 156 NHIS PSU’s that would not have affected an unlinked design as severely. Working in the opposite direction, the relaxation of the goals allowed much more efficient sampling of women other than black. Originally, there had been a plan to oversample never-married women at twice the rate of ever-married women. When this plan was dropped, Waks- berg had the idea to oversample households with multiple eligible women other than black (5). This allowed a considerable reduction in the variation in the probabilities of selection for women other than black, thereby reducing the value of a? considerably below that projected origi- nally. (In reviewing this report, an error in Waksberg and Northrup was detected. In their table 13, the terms a? and b* are not defined but are of magnitudes that make it clear that the first is the amount to be added to the design effect due to differential sampling within households, and the second is the amount to be added due to differential sampling across block groups with different black popula- tion densities. Yet in chapter 5 of Waksberg and Northrup, a’ and b? are discussed as if they had the reversed meanings. It is suggested that a® be substituted for b? and vice versa within chapter 5. The discussion will then be consistent with the numbers in their table 16 and with usage in this report.) The reduction in a® was possible only with the linked design because oversampling of multi- eligible households would have necessitated doubling the screening sample —a very expensive proposition in an unlinked design. The fact that this effect is important and in the wrong direction to explain the overprojection of savings with a linked design probably puts added emphasis on the projected effective sample size for black women in a linked design as the cause for the overestimation of the cost savings with a linked design. It remains to be empha- sized, however, that the cost savings were still quite substantial. References Waksberg J, Northrup DR. Integration of sample design for the National Survey of Family Growth, Cycle IV, with the National Health Interview Survey. National Center for Health Statistics. Vital Health Stat 2(96). 1985. Mathiowetz N, Northrup D, Sperry S, Waksberg J. Linking the National Survey of Family Growth with the National Health Interview Survey: Analysis of field trials. National Center for Health Statistics. Vital Health Stat 2(103). 1987. Massey JT, Moore TF, Parsons VL, Tadros W. Design and estimation for the National Health Interview Survey, 1985- 94. National Center for Health Statistics. Vital Health Stat 2(110). 1989. Wolter KM. Introduction to variance estimation. New York: Springer-Verlag. 1985. Judkins DR, Mosher WM, Botman S. National Survey of Family Growth, Sample Design, Estimation, and Inference. 10. National Center for Health Statistics. Vital Health Stat 2(109). 1991. Valliant R. Generalized variance functions in stratified two-stage sampling. J Am Stat Assoc 82:499-508. 1987. Krug DN, Slobasky RF, Hendriks SK, Waksberg J. National Survey of Family Growth Cycle II, final report. Rockville, Maryland: Westat, Inc. 1977. Hansen MH, Hurwitz WN, Madow WG. Sample survey methods and theory. New York: John Wiley & Sons, Inc. 1962. Westat. National Survey of Family Growth Cycle III, final report. Rockville, Maryland: Westat, Inc. 1984. Waksberg J. Oversampling of blacks in NHIS. Memoran- dum for Monroe Sirken, Andrew White, and Steve Cohen. Rockville, Maryland: Westat, Inc. 1985. 17 Appendixes Contents I. Notes on cost estimation. ovine n nner nennnnnn. II. Improvements to measurement of components of variance 18 Se ee seas eee aes ese eee ese eee eee eee Appendix | Notes on cost estimation Professional labor is: Highest in Design A because there is additional field staff to be supervised during data collection, there is a listing to be organized and supervised, and there are more completed questionnaires to be processed. Next highest in Design C because the sample must be redrawn because of delay and additional tracking must be organized and supervised. Clerical labor is: Ordered proportional to the number of questionnaires to be processed, except that Design C has higher clerical labor costs than Design B because of tracking. Field labor is: Ordered proportional to the number of questionnaires to be completed, except that Design C has higher field labor costs than Design B because of tracking. Travel costs are: ® Highest for Design A because there are more inter- viewers to travel to training and more field work that must be supported by out-of-town travel to build response rates. Next highest for Design C because there are more people who have moved to be tracked and interviewed at their new locations. Next highest for Design D because it is less clustered than Design B (i.e., has more PSU’s as well as more segments). Other direct costs are: ® Proportional to the amount of field labor because this category includes costs for local travel, postage, tele- phone, and supplies. 19 Appendix II Improvements to measurement of components of variance The measurement of the components of variance could be improved for future studies. The first area for improvement concerns measurement of within-segment variance when the number of women in the segment was odd. The second concerns the measurement of the effect of variation in segment size. Segments with just one interviewed woman each should be dropped from the file when calculating within-segment variance. Of course, the balance of the sample would have to be reweighted to get the correct totals. The resulting within-segment variances would then have to be adjusted to compensate for the fact that the sample was larger than it appeared and that the weights were less variable. Also, special treatment should be given to the segments with three women each. For example, when half of a segment to be dropped consists of one woman, the remaining two should have their weights perturbed upward by 50 percent instead of 100 percent; similarly, when the half consists of two women, the remaining one should have her weight perturbed upward by 200 percent instead of 100 percent. Measurement of the effect of variation in segment size on between-segment variance for substantive characteris- tics is more difficult. It would require the calculation of an additional set of replicate weights. This set would repre- 20 sent the between-segment variance if all the segments were equal in size (number of interviewed women). Vari- ances could be calculated with this set of replicate weights and generalized. Label the new b parameter bypg for within-PSU equal-sized segments. Formula (5) would re- main unchanged. However, formula (6) would be replaced by p2= (bwpe/bws-1)/ (A-1) (8) and a new formula would be available to estimate {,: €2=bwp/bwre 9) Deriving the replicate weights for bypg would be fairly complicated. For simplicity of explanation, assume that two segments have been paired and that one contains two interviewed women and the other, three interviewed women. When the segment containing two women is dropped, the weights for the three women in the other segment would have their weights perturbed upward by 67 percent instead of 100 percent. When the segment containing three women is dropped, the weights for the two women in the other segment would have their weights perturbed upward by 150 percent instead of 100 percent. “U.S. Government Printing Office: 1993 — 342-327/80009 Vital and Health Statistics From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics Comparability of the Death Certificate and the 1986 National Mortality Followback Survey November 1993 rd 4 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES : yy Ve] foRa Ite lig IN =o) r Centers for Disease Control and Prevention kA National Center for Health Statistics SW CENTERS FOR DISEASE CONTROL AND PREVENTION Copyright information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested citation Poe GS, Powell-Griner E, McLaughlin JK, et al. Comparability of the death certificate and the 1986 National Mortality Followback Survey. National Center for Health Statistics. Vital Health Stat 2(118). 1993. Library of Congress Cataloging-in-Publication Data Comparability of the death certificate and the 1986 National Mortality Followback Survey. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 118) (DHHS publication ; no. (PHS) 94-1392) “November 1993." “A comparision of information of demographic items on death certificates with responses obtained in the 1986 National Mortality Followback Survey. This survey included a national sample of persons 25 years of age and over who died in 1986.” ISBN 0-8406-0484-X 1. Mortality — United States — Information services. 2. Mortality — United States — Data processing. 3. Mortality — United States — Statistics. 4. Death — Proof and certification — United States. I. National Center for Health Statistics (U.S.) Il. Series. lil. Series: DHHS publication ; no. (PHS) 93-1392. RA409.U45 no. 118 [HB1335] 362.1'0723 s—dc20 [304.6'4'097309048] 93-14607 CIP For sale by the U.S. Government Printing Office Superintendent of Documents Mail Stop: SSOP Washington, DC 20402-9328 Vital and Health Statistics Comparability of the Death Certificate and the 1986 National Mortality Followback Survey r 6 ¢ ,A FEB Zo 1994 Series 2: Data Evaluation and Methods Research No. 118 A comparison of information on demographic items on death certificates with responses obtained in the 1986 National Mortality Followback Survey. This survey included a national sample of person 25 years of age and over who died in 1986. EE ET oa, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics Hyattsville, Maryland November 1993 DHHS Publication No. (PHS) 94-1392 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jack R. Anderson, Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Office of Vital and Health Statistics Peter L. Hurley, Associate Director Division of Vital Statistics John E. Patterson, Director James A. Weed, Ph.D., Deputy Director James A. Weed, Ph.D., Acting Chief, Followback Survey Branch Mabel G. Smith, Chief, Statistical Resources Branch Joseph D. Farrell, Chief, Systems and Programming Branch Contents BEB ivr nvinsmnenr neni n sa msn a BID EN IE 08% 43 050 453 50H hors 5 nok a 4 40h 4 00 HIS 00 HER BF S08 RAR Rok wi WIATILE] SEAS) vi veer 10 win emosn wie 0070 wie 5050 06 rod B38 sp ie ms tig tr Fr Aa 0 1 0 A 00 305300 i 0 hr ier ram on a a 2 930 OCCUPATION 5. vr sr 0s wri ior i 0 A010 0 101 Re 0 00 0 1 0,1 1000 0 0 si INOUSIEY o0.05.0 54.0 50 5 0m 505 ss Shhh Bo FN FRE EES H6 EMRE SR 03 RE Ae HS RARE 0 4 H08 R HE S RR ERE VCTOTOTN SLATS. oo 50 6 107 5000 S000 500i 5 05 0 48305 0 HT 0 TR 0 Te By i aw a YE 00 PLACE OL BOOT. 00510 10 01x Eo RTT whims si 3 i Re RL B10 0 inet i me iw pS 1 DD IBCUSREOT, 4.0 emer: resrimens sem eh 0 ng RE TE EE (eho 0.6 EA W331 I so esti n miserencm essen cr AD 9 Es 0 EAT Fi Uo Sr tes sormcimsomlii ws 13 1 2 5 5 ; 6 HRSPANIC OLIN 05. vs 0 em min mov 5 wie we win wih 0 ix 010 90 10 950 0 910 910.0 B10 95 01 18 BIH P30 010 0 510 win 0 wie hw 0 970 win 0 Hip 0050 00 1 2.80 930 91318 913 7 8 8 9 9 0 RETCTOMICES .. 415 804 10mm vo vow smn version wr five wh ari ete ct 0 Be Bs PUES of Bw ok ofl ro ee 14 List Of detailed tables. . «oo vv vette ete eee eee 15 Appendixes I. US. Standard Certificate Of DEAN ..ivuuvimenmimsn sin sms smam ms mss sans mame stn srmsme sass ssn vsnss 48 II. Instructions for Alling death certificate MEMS. . cov vrvvrmrvmrmennrmsmn cari m smn vie vow suinn ens veo’ 49 ITI. Selected questionnaire items, 1986 National Mortality Followback Survey. ....................ooooii.t. 52 Text tables A. Registration areas reporting age, race, Hispanic origin, marital status, occupation, industry, place of death, and veteran status on the death certificate: United States, 1986 . o.oo iin irit iii ie anne 3 B. Number of death certificates and completed National Mortality Followback Survey questionnaires for reporting States and response rates by selected variables: United States, 1986. ...........ccoviiiiiiiiiiiiiinn., 4 C. Percent of informant questionnaires in agreement with corresponding death certificate with regard to age, by age of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986.................... 5 D. Percent of informant questionnaires in agreement with corresponding death certificate with regard to race, by race of decedent on death certificate and by age at death, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 ..............cooiiiiiiiiiiiinn.. 6 E. Percent of informant questionnaires in agreement with corresponding death certificate with regard to Hispanic origin, by Hispanic origin of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1080, sons arin stm ie £10 EERE ETE fs Emre mute men ve TR RE ER (fh FB fe ier vere pn 7 F. Percent of informant questionnaires in agreement with corresponding death certificate with regard to marital status, by marital status of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 . .. 8 G. Percent of informant questionnaires in agreement with corresponding death certificate with regard to occupation, by occupation of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, BOB cc cin 000 sims mavnms wr ois 313 0 0 500 9 im 0 ie 12 80m 0 rar ADE RE Be IR 1 RR wi 9 ili Percent of informant questionnaires in agreement with corresponding death certificate with regard to industry, by industry of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986............. Percent of informant questionnaires in agreement with corresponding death certificate with regard to veteran status, by veteran status of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 ... Number of responses by place of death on death certificate and on National Mortality Followback Survey questionnaire; LInHed States, 1086... ou ummm nis sown im sm aes sings m 1m 208 531s $00 £30 058 20.50 3010 £0 20808 3 0070 5 mr 1 10s Symbols --- Data not available . Category not applicable - Quantity zero 0.0 Quantity more than zero but less than 0.05 4 Quantity more than zero but less than 500 where numbers are rounded to thousands * Figure does not meet standard of reliability or precision # Figure suppressed to comply with confidentiality requirements 10 11 Comparability of the death certificate and the 1986 National Mortality Followback Survey by Gail S. Poe, M.P.H., Division of Vital Statistics; Eve Powell-Griner, Ph.D., formerly with the Office of Vital and Health Statistics; Joseph K. McLaughlin, Ph.D., National Cancer Institute; Paul J. Placek, Ph.D., Office of Vital and Health Statistics; Grey B. Thompson, Ph.D., formerly with the Division of Vital Statistics; and Kathy Robinson, formerly with Information Management Services Introduction The death certificate is the primary source of annual mortality data in the United States (See appendix I). The validity of cause-of-death information has been studied extensively (1,2), as has the accuracy of the occupation and industry items (3-16). Less information exists on the quality of the remaining information on the death certifi- cate. Two studies have compared Census Bureau Popula- tion Study interview responses with death certificate entries (17-20). In 1986, the National Mortality Followback Sur- vey (NMFS) was conducted by the National Center for Health Statistics (NCHS) to provide a large amount of information, most of which is not available elsewhere, on a NOTES: The data collection agent for the survey was the Bureau of the Census. Cosponsors of the survey included the Health Care Financing Administration; the National Cancer Institute; the Indian Health Ser- vice; the National Heart, Lung, and Blood Institute; the National Institute on Aging; the National Institute of Child Health and Human Development; the National Institute of Mental Health; the Veterans Administration; and the Office of the Assistant Secretary for Planning and Evaluation in the Office of the Secretary of the Department of Health and Human Services. This report was prepared in the Division of Vital Statistics of the National Center for Health Statistics. Isadore Seeman, formerly with the Office of Vital and Health Statistics Systems, provided overall project direction; Steven Botman, Office of Research and Methodology, pro- vided guidance in the design of the sampling procedure; Ruth Parsons, Information Management Services, provided guidance on computer programming; Betty Smith, Statistical Resources Branch, Division of Vital Statistics, provided content review. This report was edited by Margaret Avery and typeset by Annette F. Gaidurgis, Publications Branch, Division of Data Services. sample of deaths. These data are useful in assessing the reliability of demographic items reported on the death certificate. The purpose of this report is to assess the compara- bility of demographic information obtained from re- sponses on the death certificate with data from the 1986 NMES, which is an independent source using a different method of data collection, for those items common to both sources. Although it is not possible to discern which source of data is valid, the level of agreement sheds light on the quality of these information systems. Sources and limitations of data The data presented in this report are based on the 1986 NMFS conducted by the National Center for Health Statistics and on the death certificates filed with State registrars of vital statistics and compiled by NCHS. The 1986 NMFS comprised a nationally representative sample of adults aged 25 years or over who died in 1986. Oregon was not included in the survey because of the State’s respondent-consent requirements. The data are, there- fore, representative of deaths of adult residents in the United States excluding Oregon. A detailed description of the methods and procedures used in the NMFS has been published (21). The universe for the 1986 NMFS was composed of all death certificates of decedents 25 years of age or older filed in the United States. The sampling frame consisted of death certificates selected from the 1986 Current Mor- tality Sample (CMS). The CMS is a 10-percent systematic sample of death certificates received by the State vital statistics offices and transmitted to NCHS about 3 months after the deaths. CMS records were selected for each month of the year. The total sample was 18,733 decedents. This sample included 2,274 deaths selected with certainty (at a sampling rate of 100 percent within the CMS) to meet specific research needs. The groups for which all deaths in the CMS were selected included American Indian, Eskimo, and Aleut decedents; all deaths due to Asthma; déaths due to Ischemic heart disease for males 25-44 years of age and females 25-54 years of age; and deaths for selected cancer sites. Black decedents were oversampled 2.9 times, and decedents under age 55 were oversampled 3.1 times. The data presented in this report are not weighted. They reflect what actually occurred in the sample rather than estimates of the degree of comparability from an examination of all death certificates for U.S. residents 25 years of age and older dying in 1986. It is possible, if desired, to prepare weighted estimates of consistency because the public-use data tape contains a weight for each record (22). Because of the oversampling of some groups that generally had slightly lower agree- ment rates, weighted estimates would have produced slightly higher overall rates of agreement. In the tables, an aster- isk is shown for estimates of percents in which there are fewer than 30 cases in the denominator, because these figures do not meet standards of reliability or precision. An NMFS questionnaire was mailed to the death certificate informant, usually the decedent’s next of kin or another person familiar with the decedent. A followup questionnaire was mailed for nonresponding cases. Tele- phone and personal interviews were attempted for cases where there was no mail response. Following data collection, the questionnaire data and the CMS information were matched to the Multiple Cause of Death File. The primary matching criterion was that State of occurrence and death certificate number were identical; the secondary criterion was that demographic items such as sex, date of death, age, race, and underlying cause of death matched. The primary criterion could not be applied to Nebraska, Nevada, or New Mexico, because these States renumber the death certificates. Therefore, it is likely that for these three States there were some cases in which an incorrect Multiple Cause of Death File death certificate was matched to the questionnaire. Because inclusion of these States increases the likelihood that differences in data from the certificate and the question- naire may be due to matching errors, they are excluded from this report. The total number of cases excluded because they were from Nevada, New Mexico, and Ne- braska is 285. The overall response rate for the survey was 88.6 per- cent. In addition, there was item nonresponse for both the death certificate and the questionnaire. Also, not all States collect, code, and report all variables. Table A shows, for each variable included in this report, the States that report that variable. For each variable included in this report, table B shows: ® the number of sample cases for the reporting States ® the death certificate item completion rate ® the number of questionnaires completed ® the questionnaire item response rate e the questionnaire effective item response rate (the percent of cases in the reporting States for which there was a questionnaire entry for the item) ® the effective item response rate for both the question- naire and death certificate (the percent of all cases in the reporting States that have a response for the item for both the death certificate and the questionnaire) Table A. Registration areas reporting age, race, Hispanic origin, marital status, occupation, industry, place of death, and veteran status on the death certificate: United States, 1986 Hispanic origin Race > Q © Place Marital of status Occupation Industry death Veteran status AaDAMA. ...s soir sm emss mes Alaska BUZONRL «+s wives pow im ma wiv ATKONBAS. . vans sn inr evans Califomia. . «vce rev was nsw COIOIB00. + + +5 50 5 wna x sums Connecticut. . cos mssns sams DOIawarg. «ss csrssnes nisms District of Columbia . . . ....... FIONAR & isos 5 8 7.6% 5: 0 00 vivre Georgia Hawaii KortuCKY. « « « cove rvims exons LOUISIBAA. + = vse 20 0 iti 0 0 2th 30 4 MEIBAG.. . vcr i worm Sf Rig Massachusetts , « . .... +s ses + MICHIGAA. .. + vv ws © sobs nmad MIBBBSOIR . . + vv v4 sebaes v8 o2 MISBISSIOPL. « v.vsm snd «bene 010 MiSSOUN «+ + wv vss mawsmws ws MOMIBNA « &. «5 v0 ws oi 4 iowa 0% inde New Hampshire New Jersey . . . cscs vsowwens New York North Carolina North Dakota « ov cin n smeimn sms Ohio OKIBNOMA + c+ cn vies rv oims nme PennsylVBNEa: « « vs ws vb imsmmys Rhodelsland . . ............ VOIrmont © «cc views snaws en VOIR, +00 5 2 wi aw Biers os Bis Washington: : « « « vs «x mow vw 5 West VIrginia . « oc «+ «vx weve WISCONSHY «or cou 0 ¢ we win 98 ie er WYOMING: 2 os «6s wot i 8 on win HXXXXHXXXXAHXXXXXXXXXXXXXXXXXXXXXXXXNXXNXNXNXNXNXXNXNXNXNXNXNXNX HKXXXXHXAHXHXHXHXXXXXHXHXXXXXXXXXXXXXXXXXXXXXXXXXXNXNXXNXNXX xX X X X X DK MK XX X X xX X X x x XX XX XX XXX XX XX XX XX X x x XX X X X XXX XX XX XX x xX X XxX xX XX HX XXX XX XXX x x HXXXHXHXAEXAXHXHXXXHXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX XX XX XX X XX XX X X XX X Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. The effective response rate for both the questionnaire and the death certificate was between 82.3 and 86.6 percent for all items except veteran status, which was 75.7. For all variables in this report, with the exception of veteran status, the responses from the Multiple Cause of Death File are compared with those from the question- naire. Because veteran status is not included in the Mul- tiple Cause of Death File, this variable was taken from the CMS. In presenting the percents of responses agreeing in tables C-J, the percents are based on the number of cases in which there is a response to both the questionnaire and the death certificate. Cases in which entries for an item are blank, illegible, or otherwise unusable for either the questionnaire or the death certificate are excluded from both the numerator and denominator of the percents. In comparing the two data sources, information from death certificates was used as the denominator. That is, agreement levels reflect the degree to which next-of-kin information on the questionnaire matches that from the death certificate. Percent agreements shown are based on the groupings shown. For example, where percent Table B. Number of death certificates and completed National Mortality Followback Survey questionnaires for reporting States and response rates by selected variables: United States, 1986 Death certificate Questionniare Completed questionnaire and certificate Item Item Effective item Effective item Variable Total completion! Total response? response? response? Number Percent Number Percent Percent Percent AGB.:sinsrnsnsshsmn 18,448 99.8 16,339 97.9 86.7 86.5 REGE 4is mt aliens wm wis 18,448 599.9 16,339 97.9 86.7 86.6 Hispanic origin . . . ..... 8,356 98.0 7,568 94.1 85.2 83.7 Marital status . . ....... 18,448 99.3 16,339 97.6 86.5 86.0 Occupation . . « «vas 4 55 4,525 96.3 4,177 95.1 87.8 84.7 IBIUSHY: «iiss ana @e va 4,525 96.3 4177 92.4 85.3 82.3 Place ofdeath ........ 13,580 99.8 11,895 98.1 85.9 85.7 Veteran status . ....... 14,050 87.9 12,422 96.6 85.4 75.7 NOTES: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. The denominators of these rates exclude the number of cases in those States that did not code or collect this information on the death certificate. See table A for specific States. 1Cases for which a final death certificate was not matched were classified as nonresponses. 2percent of filled questionnaires in the reporting States for which there was a substantive questionnaire entry for the item. percent of all cases in the reporting States for which there was a substantive questionnaire entry for the item. 4percent of all cases in the reporting States for which there was a substantive response for the item for both the death certificate and the questionnaire. SFor 1986, the number of death certificates for which race was unknown, not stated, or not classifiable was 0.2 percent of the total deaths (for all States and registration areas). Death certificates with race entry not stated are assigned to a racial designation as follows: If the preceding record is coded “white,” the code assignment is “white”; if the code is “other than white,” the assignment is “black.” agreement is shown for the 25-29-year age group, this means the number of cases in which the age on both the death certificate and the questionnaire is in the range 25-29 divided by the number of cases in which the age on the death certificate is in the range 25-29. Similarly, where the percent agreement is shown for an occupation or industry category such as “managerial and professional,” this percent is for the group as shown—not for less aggregated levels. Sources of error for both the death certificate and questionnaire include reporting errors, coding errors, and processing errors. Except for occupation and industry, conceptually the variables are the same for both sources. The death certificate asked for the “usual” occupation and industry, and the questionnaire requested information on longest held occupation and industry in which the decedent worked for pay. “Usual occupation” on the death certificate is defined as the kind of work the dece- dent did during most of his or her working life. In addition, the place-of-death variables differ somewhat between the two sources. For the questionnaire, the re- spondent was simply asked, “Where did the person die?” For the death certificate, the place of death variable is based on the location of death, which may be at a hospital, en route to or on arrival at a hospital, or at some other place. If a hospital was cited, a distinction is made among decedents pronounced dead in the hospital or other insti- tution, those dead on arrival, outpatients or emergency room patients, and inpatients. With respect to age, Hispanic origin, marital status, occupation, industry, and veteran status, coding instruc- tions are essentially the same for both sources. Occupa- tion and industry were coded according to standard occupation and industry codes (23). There were differ- ences in coding race: On the death certificate, entries such as “Mexican,” “Cuban,” and “other Hispanic” were coded as “white”; on the questionnaire, such entries were coded as “other.” Moreover, responses that were not exactly one of the four major races were classified by coders in most cases as one of the four major races on the death certifi- cate, whereas they were left as “other” on the questionnaire. Copies of the U.S. Standard Death Certificate, the instructions for completing the certificate, and the respon- dent questionnaire items are included in this report as appendixes. Findings Age There was only 77.5-percent agreement on exact age of decedent (table C). The agreement was highest for decedents 25-29 years of age (85.9 percent) as reported on the death certificate, and lowest for decedents 70-79 years of age (74.0 percent). There was a strong relationship between percent agreeing on exact age in number of years and the interval between the death and the survey: There was 85.5-percent agreement for the shortest interval of 22-25 weeks, and only 67.0-percent agreement for the interval of 52 or more weeks. This relationship was observed for most 10-year age groups. There was greater agreement in age for white dece- dents (81.6 percent) than for black (67.1 percent) or Amer- ican Indian, Eskimo, and Aleut (70.8 percent) decedents. There was greater agreement for white decedents for each Table C. Percent of informant questionnaires in agreement with corresponding death certificate with regard to age, by age of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Age of decedent on certificate 25 years 25-29 30-39 40-49 50-59 60-69 70-79 80-89 90 years Case characteristics and over years years years years years years years and over ANCEBBS. « init wu vss ewe 77.5 85.9 81.0 81.0 775 772 74.0 75.4 78.2 Race WHHBL. «cv rviewn wwiiimy wmivian 81.6 87.7 83.6 83.6 81.9 83.4 78.7 79.2 82.1 BlaOK »:vws su sws smswara ems 67.1 80.9 75.1 73.6 67.7 66.5 62.5 62.4 64.9 American Indian . . .......... 70.8 *76.2 68.2 69.8 73.8 67.7 72.0 75.4 *59.3 Interval between death and survey D2-2BWEBKS ; «sis us sawn vn 85.5 74.4 87.4 93.7 90.5 88.0 84.4 81.5 82.4 26-29weeks . ............. 84.2 95.2 87.2 86.2 84.6 83.4 82.9 81.7 85.1 B0-32WBBKS . «vs snsmn vis 85.0 96.6 84.3 89.8 89.4 85.5 81.0 83.6 83.7 B3-3B5WEBKS . «vs us svi sux 81.0 89.6 85.6 83.5 79.1 82.7 75.4 80.9 81.8 B6-3BWeeKS . .. uk srams vas 75.4 87.8 80.5 84.0 76.4 73.8 71.8 70.3 74.2 39-41weeks . ............. 73.9 84.5 76.6 78.0 75.4 72.4 67.7 725 76.7 A2-AAWBBKS . «+ vo vv smn mus 70.5 79.5 80.0 71.8 72.6 68.9 64.9 67.6 71.0 48-47 WEBKS! : cs sw swe vin s nv 8 70.5 82.5 77.8 74.8 67.1 64.4 66.4 70.9 67.2 48-51weeks . ............. 69.5 81.8 74.3 71.3 61.5 68.6 68.1 68.8 69.6 52 weeks orlonger . ......... 67.0 80.9 721 68.6 68.0 70.1 65.0 56.7 62.5 Relationship Decedent was death certificate informant’s — BPOUSS «ov vs ww visi mi smn 83.0 86.2 84.9 85.9 83.2 83.7 81.1 79.1 80.0 Parent . «uw vars vn vm smw sw e 82.9 87.7 81.5 79.8 74.6 *95.2 *91.7 *85.7 *83.3 Childs: sv sings sma a a® 2% 73.1 *100.0 *65.0 75.6 65.2 69.9 69.2 77.4 78.0 Sibling. ................ 66.3 75.0 73.3 61.9 55.4 72.5 59.2 66.7 *90.9 Other relative. . . .......... 61.1 *50.0 53.3 *61.1 72.2 68.3 57.1 59.9 66.3 Nonrelative . « : sco swnenass 74.1 *84.6 71.0 73.8 76.2 84.4 72.5 65.4 81.3 Death certificate informant and survey respondent were — Both decedent's spouse. . . . .. 84.8 90.5 87.8 88.3 84.6 85.7 82.1 80.4 82.7 Not both decedent's spouse . . . 73.7 85.0 77.6 74.9 70.5 70.3 68.8 74.2 77.9 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon’s confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 10-year age group than for any other racial group. (Here- after in this report the category “American Indian, Es- kimo, and Aleut” will be referred to as “American Indian.” There was greater agreement (83.0 percent for all decedents) when the death certificate informant was the spouse, as compared with other relatives or nonrelatives. When the decedent’s spouse was both the death certificate informant and the respondent to the questionnaire, the agreement was far higher (84.8 percent) than when this was not the case (73.7 percent). This greater correspon- dence was observed for all 10-year decedent age groups examined. For 92.7 percent of the cases, the age was either the same or only 1 year different on the death certificate and questionnaire (data not shown). There was a slight ten- dency for the questionnaire age response to be older than the age on the death certificate. For 10.2 percent of the cases, the age was 1 year older on the questionnaire. For 5.0 percent of the cases, the age was 1 year younger on the questionnaire. Within 2 years there was 95.7-percent agree- ment, and within 5 years there was 98.2-percent agree- ment on decedent’s age. Mortality data are commonly tabulated by 5-year age groups for analytic purposes. An error of 1 year on the death certificate would result in a difference in the tabu- lations only when the correct age fell within another age interval. For 5-year age groups, there was 93.4-percent agreement (table 1). When the death certificate informant was the spouse, the agreement was 95.3 percent for 5-year age groups. When both the death certificate informant and the questionnaire respondent were the spouse, the agreement for 5-year age groups was 96.0 percent. Race Overall, there was a high level of agreement (97.9 per- cent) on race between the death certificate and the ques- Table D. Percent of informant questionnaires in agreement with corresponding death certificate with regard to race, by race of decedent on death certificate and by age at death, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Race of decedent on certificate American Case characteristics All races White Black Indian AVCABBS +: vrs 01x 010 1008 4 BLOW 5/4: WAR 210 8 0 ribat] th 97.9 98.2 98.0 92.9 Age UNOS BO. + «vows sais 55 ai ow 86H 00060 EH EHLE PERS 96.5 96.8 96.2 *90.5 BO=BYYBAIS. «vv vv ie bt Wins @ 4 lie 3h 0 000 0 ah 96.7 96.7 97.5 95.7 BO-00 YEAS. v.10) 2 3 2.58 55 Barbs ele fi seb aside aco che shone 97.3 97.3 98.5 91.3 BO-BOYOHIS 1 vss mene s Roms neta ts wll Finan» wee 97.7 98.1 98.5 88.9 BO-BIYBRB. os svn rumvs en dlns bRFIRE Danes Tan 98.3 98.6 98.2 95.4 TO-TOYBAIS : «cvs iw wiw iv) #1558 0) 0k wink V3. 18 imal wie WIE: #0 ke 98.3 98.7 98.1 92.9 BOB YORISS om 5. 55 50080 50 140 518 5: Sn 0:00 30 0 598 GH ds Tse 90a 98.5 98.8 98.4 925 OO YEAS AND OVBE | ic. vs = + 05% 06 Bab iv £00 308 0 480 wel: #300 98.3 98.8 97.0 *96.4 Interval between death and survey 22-2BWBBKS , ov: 55 1.0 0 vivo ® 3 3000 Bin wn ® BIE SHE 3 FRE 99.1 99.0 99.2 *100.0 DB-2OWOBKE «vv sv ® 510 iw #338 v.00 78 B16 Bik 36 % (8 #08 & 5208 98.7 99.0 98.8 95.0 BO-SIWOBKE «4 4 003 18) 0 #0) #50 ve 41080 1 00 ta: 058) 61cm 300 9p indy 0c 0 rie 98.8 99.2 98.3 89.5 B3-BBWEEKS . ...ocnnrnscns vi sn rn cn mentee 98.0 98.5 97.7 91.7 BB-BBWBBKS «civ oa 0 is wie vi 00 bin la) 900 600 0 57% Bott 16 430 0 97.7 98.2 97.5 91.5 BOAT WBBKS! . ov vu vis msm nani sna mb 89) 43 4.5 808 © Bini ad 97.8 97.5 98.9 96.8 AZ AA WEBKS + vi. 5 in: 3 05 2.05 3.5 0 2 8 59 05 4 5 50 0 5 0.50 100 (00 9 30 Bc 96.7 96.7 96.8 95.3 ABT WEBKS: , ie vv ves kk fv a 4 oo toy & 0 9 0 0 8 0 0 96.5 96.7 97.4 *80.8 ABB WBBKS: ; cv 7 voi emis iis hi 4 00 6 in 4 Sl 4 4 4 618 00 8 96.9 97.3 98.1 *86.2 B2 WOBKSIOF FONQIBK iv: «wv 515 5 9 5 104 50 1 4 16 0 3, 0 0.904 1 0 0 0 0 # 97.9 98.0 98.8 *100.0 Relationship Decedent was death certificate informant’s — SPOUSE is smemr emia dems sme ms tHe me ERIE 3 98.2 98.6 98.1 91.1 PABA. . oo cosnmsnmrps cmon vmsma snobs gnswss 97.4 97.0 98.5 *96.4 COI. «sien msm mss Saws meme amma knsase 98.0 98.1 98.4 96.2 BING: vis avs Rs AmB ERED a RE me Er RE aw 96.2 96.3 96.8 *88.5 OBI TBIBING. v1 ie: 5 vos 5 00 i 6 im 0a i 8 ton ws am 0 4 50 97.9 98.6 97.7 *90.9 NOIWBIBHVE . . . cco cnvmsvmems smsnmasmenmnnmenss 96.6 97.1 98.9 *86.7 Death certificate informant and survey respondent were — Both decedent's Spouse. . » «+ vs cvs vs saswrs sess 98.4 98.7 98.2 90.4 Not both decedent'sspouse . . . .................. 97.7 97.9 98.2 93.3 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. tionnaire (table D). However, for those reported to be American Indian on the death certificate, the level of agreement was lower (92.9 percent). Of the 7.1 percent of cases reported as American Indian on the death certificate but as another race on the questionnaire, most (80.0 per- cent) were identified as being white in the questionnaire (table 2). Unweighted data indicate that there were 92 more (21.8 percent more) American Indian decedents reported in the questionnaire than on the death certificate. Of the 122 cases identified as American Indian in the question- naire but not as American Indian on the death certificate, 70.5 percent were identified on the death certificate as white, and 27.9 percent as black. The increased reporting of American Indian on the questionnaire occurred for all of the intervals between death and survey and for all relationships examined be- tween the death certificate informant and the decedent. Even when both the informant and the questionnaire respondent were the decedent’s spouse, 21.3 percent more American Indian decedents were reported in the question- naire than on the death certificate. Hispanic origin There was 98.9-percent overall consistency in report- ing Hispanic origin between the death certificate and the questionnaire (table E). A high level of consistency was observed for both Hispanic origin (97.1 percent) and non- Hispanic origin (99.0), as well as for all races, intervals between death and survey, and relationships between informant and decedent examined. Of the 1.1 percent of cases in which there was disagree- ment, 88.5 percent were cases in which the origin on the death certificate was non-Hispanic and the origin in the questionnaire was Hispanic (table 3). This resulted in 19.6 percent more Hispanic decedents being reported in the survey, based on unweighted data. Higher reporting of Hispanic decedents in the questionnaire occurred for all races, intervals, and relationships of informant to dece- dent examined. When both the death certificate informant and the questionnaire respondent were the decedent’s spouse, there were 11.8 percent more Hispanic decedents reported in the questionnaire. Table E. Percent of informant questionnaires in agreement with corresponding death certificate with regard to Hispanic origin, by Hispanic origin of decedent on death certficate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Hispanic origin of decedent on certificate All Hispanic Non-Hispanic Case characteristics origins origin origin AUCBSES , «ovo wp va vinswmmms mis ws wns we ems w w 98.9 97.1 99.0 Race WHE St 5 4 co 508 50 6 8 0 8 i cop mnie oor 1 #0 ir in 3 em 7 me 99.0 97.7 99.1 BACK uv vis 00 5g dw eh rn BEBE ARE WE EW 99.4 *100.0 99.4 American Inoian . . c« cx ss suis ur sma RENE 2H w 94.3 *100.0 94.3 Interval between death and survey 22-25 WBBKE . ox 1 wip nweiwn nus ws cama SHEE fH EN 98.8 *90.9 99.0 26-20WBBKS . vos vss mt nb sme EEE Ap EET 99.3 100.0 99.3 BO-32 WOBKS! « 25. 2 i516 toc 8 0: 5 0 0 AIH 5 490 10 # 000 im 0 5: 82 om ov wim 99.1 *100.0 99.1 33-35WeeKS . . .... ee ee 99.6 *96.6 99.7 BE-SBWBBKS , ie viv von vin ia ww vw Ries wg ee 99.0 94.4 99.2 BO-4IWBBKE + 4 vo smb va ims 6 shim sFs is @aen 98.7 100.0 98.6 42-44 WEEKS . . . ii 98.9 95.1 99.1 AB-AT WBBKS 1, «iv i #1 #5 0 0s 3 6 6 4 9 0 5 RE RR 98.3 *100.0 98.2 ABB WBBKS . ics cvivin in wiv vine oi wim h Sk 50% Wi® ¥ IEEE 97.4 *91.7 97.8 B2WeBKS OTIONGSY + sss wh ss sma imiRaimsrodms ss 98.6 97.7 98.7 Relationship Decedent was death certificate informant’s — SPOUSE + v4 5 vw sds ¥ Beds 2EIEE H DELS IE WEE SED 99.2 97.3 99.2 PAINT or nm lp 2 5 0le) Sr) SR) Fe ones mane simi: sin 98.5 100.0 98.4 CRAIC 6] Gn ons 0 sen we i co pent) cm ot oo rt io Ht et maa 98.8 100.0 98.8 SIPING. + sors a eras sr ENR HEEL E NE HE VEHET $ 98.2 *96.4 98.4 Otherrelative. . . ccs sssnsvssss senna vasinss 99.3 *100.0 99.3 Nonrelative . . . ............ coins 98.2 *94.4 98.4 Death certificate informant and survey respondent were — Both decedent’s SPOUSE. « cous ws sv sv suv swan ow 99.4 97.6 99.4 Not both decedent's spouse . . . ................ 98.7 96.6 98.8 NOTES: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. Table F. Percent of informant questionnaires in agreement with corresponding death certificate with regard to marital status, by marital status of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Marital status of decedent on certificate All marital Never Case characteristics statuses Married Widowed Divorced married ACASES , us vrs mvry srw ws wn sma 95.0 98.4 93.1 87.1 92.9 Race WHILE. . . vvvv vmivn a waimnis vain smn 96.6 99.0 95.6 90.7 94.0 BIaOK «co 4 wv www mm ws www www wa we 90.6 96.4 87.3 76.6 91.1 Amsrican indian ; .. . sss ses ves arse 92.9 98.2 92.5 83.0 87.7 Interval between death and survey 22 PB WOKS v5 son vs mv mwas wwe saw 96.1 99.1 94.1 88.8 94.3 26-20 WEBKS , nx wuss we SEE BERT 5 97.2 99.4 95.6 92.2 96.1 30-32WeekS . . ...... iii 96.1 99.1 92.9 90.6 96.7 33-35weeks . .......... ii 96.5 99.3 94.2 89.7 95.8 BB-BEBWBBKS . vv iv vow v 0m 50 7% ws wih 9% 94.5 97.9 92.1 89.2 91.5 BOAT WEBKS .....c. 20 von 2 15030 3 i 3008: Bi SR 93.7 97.3 92.4 89.2 87.8 42-44weeks . . .. Li 93.0 97.0 91.3 82.8 92.5 ASAT WEBKS 5 ous sn vin s wm sms wo wasn 92.6 97.0 92.3 80.7 89.0 AB-BIWBBKS « «5: vvvnsentms enews sn 92.8 98.4 90.9 79.2 89.7 52weeksorlonger .:.c.cssmverensnn 94.0 98.6 93.3 77.4 93.0 Relationship Decedent was death certificate informant’s — SPOUSE vin sins wb ii bn iF OWE EEE 99.3 99.4 Fg *87.5 ve Parent. . ..covimuewrwnenmemassins 92.1 93.9 80.7 87.9 94.8 OHA. vv vivre ow ma oe ww whe 93.6 95.6 95.6 84.6 78.9 SIbING +: «3 swims swswnsmems swres 91.8 89.6 86.0 93.3 94.7 Other relative, «: vs ccs es ens swims 90.4 93.9 92.0 83.3 88.0 Nonrelative . . ................... 88.0 93.2 86.2 84.3 89.2 Death certificate informant and survey respondent were — Both decedent's spouse. . . .......... 99.6 99.7 ce *50.0 wine Not both decedent's spouse . . . ....... 92.6 94.8 93.2 87.4 93.1 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. Marital status There was also a high level of consistency of reporting between the death certificate and the questionnaire on marital status of the decedent (95.0 percent) (table F). There was agreement in 98.4 percent of the cases for “married” marital status, but only 87.1 percent agreement for “divorced” marital status. The agreement on marital status was 96.6 percent for white decedents, 90.6 percent for black decedents, and 92.9 percent for American Indian decedents. There was a slight decline in agreement on marital status as the inter- val between death and survey increased. There was almost total agreement (99.3 percent) when the death certificate informant was the decedent’s spouse. When the decedent’s spouse was both the death certificate informant and the questionnaire respondent, the agree- ment rate was 99.6 percent. When this was not the case, the agreement rate was 92.6 percent. Among the inconsistent cases, 12.9 percent had “di- vorced” reported on the death certificate. Of these 207 8 cases, 124 cases (59.9 percent) reported questionnaire marital status as “married,” 52 cases (25.1 percent) “wid- owed,” and 31 cases (15.0 percent) “never married” (table 4). Occupation The overall percent agreement for occupation based on the major occupation groups shown was only 71.0 per- cent (table G). As reported on the death certificate, the rate was lowest for managerial and professional occupa- tions (57.6 percent) and highest for farming occupations (81.9 percent). The consistency of reporting was not ap- preciably affected by race of decedent, interval between death and survey, or relationship of informant to decedent. For all occupational categories except managerial and professional, the percent of decedents in the category was about the same or higher for the questionnaire than for the death certificate. Based on unweighted data, compar- isons showed 6.1 percent more technical, sales, and ad- ministrative; 1.3 percent more service; 16.3 percent more Table G. Percent of informant questionnaires in agreement with corresponding death certificate with regard to occupation, by occupation of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Occupation of decedent on certificate Precision Operators, Managerial Technical, production, fabricators, All and sales, and craft, and and Armed Case characteristics occupations professional administrative Service Farming repair laborers Forces ACESS... + 5k wks le msn 719 57.6 71.0 75.8 81.9 69.5 74.2 69.4 Race While... v. vavwmssasmsi 70.4 57.4 72.0 69.3 82.5 70.2 76.7 73.3 Blog, sco sora mus uip inns 73.6 61.3 69.4 82.6 79.7 68.8 70.4 *40.0 American Indian . ........ 65.6 *62.5 *28.6 *71.4 *88.9 *60.0 *69.6 *— Interval between death and survey 22-25weeks . .......... 69.6 55.6 65.6 *73.9 *87.5 66.7 75.5 *75.0 26-29weeks . .......... 72.5 57.1 76.6 76.1 76.3 61.2 81.1 *85.7 30-32weeks . .......... 66.8 56.8 61.1 65.1 90.6 7.7 67.0 *80.0 33-35weeks . .......... 73.9 60.5 75.6 70.5 *79.3 79.4 80.6 *100.0 36-38weeks . .......... 70.4 51.9 81.7 84.6 73.7 66.7 67.3 *66.7 39-41weeks ........... 67.8 53.5 68.3 72.2 *95.2 65.8 68.1 *42.9 42-44 weeks . ....... 77.3 71.0 69.8 81.1 *81.0 87.1 77.6 *66.7 45-47weeks . .......... 70.7 *53.6 *57.1 *75.0 *91.7 *79.3 74.5 *50.0 48-51weeks . .......... 70.0 *54.5 *53.8 *94.4 *100.0 *30.0 76.9 *100.0 52 weeks or longer . . ..... 60.0 *100.0 *60.0 *83.3 *33.3 *50.0 *55.6 *— Relationship Decedent was death certificate informant’s — Spouse... 69.6 57.9 67.0 69.3 83.0 68.0 75.4 *90.0 Parent. ..ossvmmswrns 67.0 43.3 *66.7 *69.0 *69.2 68.8 75.9 *— Child. . . cox vinsnnsnes 67.3 *57.1 75.0 69.4 *85.7 *471 69.7 *25.0 BENING. nvr snp = Rew 71.8 *62.5 *82.5 *78.9 *42.9 *83.3 *75.0 * Other relative. . . ....... 81.1 *66.7 *88.9 *100.0 *90.0 *66.7 *66.7 *100.0 Nonrelative . . . ........ *65.5 *25.0 *80.0 *100.0 *— *50.0 *77.8 *100.0 Death certificate informant and survey respondent were — Both decedent's spouse. . . 69.7 58.2 66.9 67.0 83.8 67.4 76.8 *89.5 Not both decedent's SPOUSE, vv + vis wv was ies 72.0 56.2 73.5 79.5 81.6 71.7 73.0 *53.3 NOTES: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. farming; 5.2 percent more production, craft, and repair; 1.1 percent more operators, fabricators, and laborers; and 80.6 percent more members of the Armed Forces on the questionnaire than on the death certificate (table 5). How- ever, there were 26.8 percent fewer decedents recorded as “managerial and professional” on the questionnaire. When the decedent’s spouse was both the death certificate informant and the questionnaire respondent, there were 23.6 percent fewer decedents recorded as “managerial and professional” on the questionnaire. Industry The rate of agreement between the death certificate and the questionnaire based on the major groupings shown was about the same for industry (74.4 percent) as for occupation (table H). The agreement rate was highest for the mining industry (79.5 percent) and lowest for the public administration industry (62.3 percent). There was no essential difference in consistency of reporting by race of decedent or by whether the spouse of the decedent was both the death certificate informant and the questionnaire respondent. The number of sample cases is too small to assess differences across intervals between death and survey, or by relationship of informant to decedent (table 6). In spite of the overall relatively low level of agreement between the questionnaire and the death certificate on industry, the marginal distributions of industries for the questionnaire and death certificate were very similar (table 6). Veteran status The agreement between the death certificate and the questionnaire on veteran status was high (96.7 percent) 9 Table H. Percent of informant questionnaires in agreement with corresponding death certificate with regard to industry, by industry of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Industry of decedent on certificate Transpor- tation, Agriculture, communica- Finance, forestry, Con- tions, and insurance, Public All and struc- Manu- other public and real administra- ~~ Armed Case characteristics industries fisheries Mining tion facturing utilities Trade estate Services tion Forces Allcases. ......... 74.4 78.4 79.5 74.3 77.0 75.2 69.6 73.6 75.0 62.3 68.4 Race White. . .......... 74.3 77.1 78.0 78.7 771 76.4 705 77.1 71:1 64.9 68.8 BACK. «xv 4 vc 4c #00 mi a 75.2 80.7 *100.0 64.1 76.2 717 65.6 *57.1 82.7 *61.5 *60.0 American Indian . . . .. 72.6 *87.5 *100.0 *66.7 *85.7 *66.7 *75.0 *50.0 *75.0 *33.3 *— Interval between death and survey 22-25weeks . ...... 75.6 *86.4 *100.0 *80.0 75.0 *78.3 *69.0 *75.0 69.6 *66.7 *100.0 26-29 weeks . ...... 77.7 75.7 *57.1 66.7 83.3 73.9 69.1 *83.3 81.7 *79.2 *83.3 30-32weeks . ...... 75.3 86.7 %75.0 *72.4 78.0 *74.1 774 *76.9 72.0 *58.8 *66.7 33-35weeks . ...... 75.1 *77.8 *88.9 83.8 82.4 75.0 65.2 *76.9 69.4 *60.0 *100.0 36-38weeks . ...... 70.8 68.4 *87.5 62.5 73.6 79.4 67.6 *84.6 70.6 *52.9 *50.0 39-41weeks . ...... 73.4 *89.5 *40.0 *80.8 74.6 *68.4 77.1 *60.0 721 *73.3 *42.9 42-44 weeks . . . .... 70.7 *77.3 *100.0 *79.3 82.7 *64.3 65.8 *63.6 76.8 *64.3 *66.7 45-47 weeks . . ..... 74.9 *01.7 *100.0 *94.4 70.3 *83.3 *52.4 *50.0 84.6 *40.0 *50.0 48-51 weeks . ...... 76.6 *100.0 *— *50.0 80.0 *66.7 *71.4 *80.0 90.3 *25.0 *100.0 52 weeks or longer . . . 68.3 *28.6 *— *40.0 *75.0 *100.0 *100.0 *— *81.8 *— *— Relationship Decedent was death certificate informant’'s — SPOUSE vv vs vans 75.2 777 *70.8 77.9 76.8 77.8 68.5 82.4 74.4 65.4 *89.5 Parent. vu venues 66.3 *70.0 *50.0 *70.4 72.2 *66.7 *59.3 *60.0 64.4 *60.0 *33.3 Child, sax amass 77.3 *90.0 *100.0 *66.7 83.3 *75.0 *77.8 *60.0 75.5 *62.5 *50.0 Sibling. . ........ 75.6 *42.9 *100.0 *83.3 *75.0 *714 *62.5 *100.0 *83.3 *100.0 *— Other relative. . . . . . 80.0 *88.9 *100.0 *50.0 *75.0 *60.0 *77.8 *100.0 *85.7 *— *100.0 Nonrelative . . . .... 81.5 *— *— *50.0 *83.3 *100.0 *100.0 *100.0 *75.0 *100.0 *100.0 Death certificate informant and survey respondent were — Both decedent's SPOMSE: «+ viv 5 sw 75.4 77.5 *71.4 77.6 78.4 78.6 66.4 84.4 72.3 67.6 *88.9 Not both decedent's spouse. . ....... 74.0 79.7 *85.0 72.1 75.8 73.0 7.7 61.5 76.4 51.9 *58.8 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. (table J). However, the rate of agreement for nonveterans was higher than for veterans (98.3 percent versus 90.7 per- cent). There was no essential difference in the rate of agreement by race of decedent, interval between death and survey, relationship of death certificate informant to decedent, or whether the spouse was both the death certificate informant and the questionnaire respondent. The percent reported as veteran was about the same for both the death certificate and the questionnaire (20.5 percent and 19.9 percent, respectively) (table 7). Of the 349 cases in disagreement on veteran status, 202 (57.9 percent) classified the decedent as a veteran on the death certificate but as a nonveteran in the questionnaire, 10 and 147 (42.1 percent) classified the decedent as a nonvet- eran on the death certificate but as a veteran on the questionnaire. Place of death The consistency rate for hospital deaths (including inpatient, outpatient, and emergency room patient) was 88.3 percent (table K). Among those with “hospital inpa- tient” reported on the death certificate as place of death, questionnaire responses reported approximately 87 percent died in the hospital excluding the emergency room, and 8.5 percent died in the hospital emergency Table J. Percent of informant questionnaires in agreement with corresponding death certificate with regard to veteran status, by veteran status of decedent on death certificate and by race on survey questionnaire, interval between death and survey, and relationship of informant to decedent: National Mortality Followback Survey, 1986 Veteran status of decedent on certificate Both Case characteristics statuses Veteran Nonveteran ANCASES . 1c vw sins smi EI A PERI MARE ARIMA MEFS 96.7 90.7 98.3 Race WHHB . ..vois wv nmima sasha ssa s@s@s ims mens 96.6 91.2 98.1 BACK oo. cs vas msins sme his emit nud dma mh dmg 97.1 90.0 98.5 American Indian . . ... LL. 96.9 84.6 99.1 Interval between death and survey DOIE WEEE vv 50 5: 3. 5 17 lh oc 50 eth 300) 0 i) 4 BR hs 0h 97.8 94.8 98.8 SO-20WBEKS , «/a% 55 38 2505 ads ale Bin weston » 50 2miaimrin = 96.3 90.5 97.8 BO-32 WEBKS 4.0 0 ie 5 4 014 81 3130 07080 91 BF ip 0) 0) 80 6 0 6 97.0 91.1 98.6 B3-B38 WOBKS io. + voit sw view v 021% 3 8 @ S103 0 @ Ba R 8 RON 96.2 88.4 98.3 BE=38WREKS , 4 11a 5 waa 3a els £5 SHE A sud ald mmm 96.4 92.2 97.4 BOAT WEEKS i 1 yw 5 5 wt wine hth Simi rw 201 a wh wit 0s 96.8 86.3 99.6 ABA WBBKE i. 4: 00 coin 3 0 500 eH BR HF EE BE 97.2 92.6 98.2 AS-ATWEBKS: . ous vs ha nia ie sr eRe sHaRe CYTE E HE 97.2 93.9 98.0 ABST WEOKE: oq vss 2 0 500 Tole IEE 31 B00 01 0h isms ser sence 97.1 91.5 98.4 B2WBBKES OF IONGBE + + vv wv va vis wns ww ww win wor v ws 94.2 88.1 96.2 Relationship Decedent was death certificate informant’s — SPOUSE 515 ir: iv 200 573 0 00% 5 8 106 600 918 5% whe 69 0. 0k 5 aint 95.9 91.2 97.8 PRION coi vo 0 0 is i 5 001 0 0 # (m5e 0% 05 Ee 0 TEE ES 96.8 91.1 98.3 CII o: o= 2k 0 50 5 0a 1 B00 50 4 0 9 J 7 oh Bos HE ch ee RT ome 97.7 89.5 98.8 SING «cima Emam bs misms nm ma ews ma nurs 96.9 91.7 98.1 Other 18IatIV8., «. s serra vas ws smsmur mame Rag ws 98.1 90.3 99.1 NOTWEIZHVE cvs cms wr nn ams smams Enaks ames a 95.3 84.8 98.2 Death certificate informant and survey respondent were — Bolh0eCoUBNIS SPOUSE. «vi» isis eis sie sists ins aiminonn « 95.8 91.2 97.7 Not both decedent’S SPOUSE: . ..: wav wa a sis ms wx ts 97.3 90.2 98.5 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. Table K. Number of responses by place of death on death certificate and on National Mortality Followback survey questionnaire: United States, 1986 Place of death on death certificate Hospital Dead on Outpatient Hospital arrival — or status not Other All other hospital Response to questionnaire item emergency Dead on Status on care reported name not “Where did the person die?" Total Inpatient room arrival unknown certificate institutions entries given AILDIBCES : oak 55 + Bw wie puss mmm 11,639 5,661 1,101 800 189 1 1,367 2,519 1 Hospital emergency room . . . ....... 1,148 480 503 99 35 w— 11 20 - Hospital (excluding emergency room). . . 5,190 4,902 84 16 110 - 56 22 - Onwayto hospital. . ............. 235 28 101 76 6 - 2 22 - Nursing or personal care home. . ..... 1,480 17 11 40 4 - 1,270 38 - OWNNOMB. «cv sie ms smoms saisms 2,564 79 267 368 17 - 5 1,828 - Other place (undefined) ........... 794 52 106 162 17 1 22 433 1 Other's NOME: ws cv:os smrws cosine 228 3 29 39 - - 1 156 - NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 11 room. Among those classified on the death certificate as “hospital outpatient or in the emergency room,” over one-third (36.5 percent) were recorded on the question- naire as having died at their own home, another’s home, or another place. For those classified according to the death certificate as dead on arrival (DOA) at the hospital, questionnaire responses showed 46.0 percent as having died in their own home, 20.3 percent in another place, and 4.9 percent in another’s home. 12 Of those decedents whose death certificates cited their dying in another care institution, 92.9 percent died in a nursing or personal-care home according to the question- naire, and 4.9 percent died in the hospital. Among those classified as “all other entries” on the death certificate, 72.6 percent were reported as dying in their own home, 17.2 percent in another place, and 6.2 per- cent in another’s home. Discussion Consistency in reporting between the death certificate and the followup questionnaire was excellent for race, Hispanic origin, marital status, and veteran status. How- ever, in spite of overall high correspondence, there were some areas of lesser agreement for these variables. For example, based on unweighted data, there were 21.8 per- cent more American Indian decedents reported on the questionnaire than there were on the death certificate. Similarly, while the overall level of agreement on marital status was 95.0 percent, for those classified as divorced on the death certificate, there was only 87.1-percent agree- ment with the questionnaire. In addition, in spite of an overall agreement rate of 98.9 percent on Hispanic origin, 19.6 percent more Hispanic decedents were reported on the questionnaire than on the death certificate. Although the agreement rate for exact age in years was only 77.5 percent, the agreement rose to 92.7 percent for ages within 1 year and to 95.7 percent for ages within 2 years. There seems to be a small bias in the direction of the questionnaire age being older. This might be due to some questionnaire respondents reporting what the dece- dent’s age would have been at the time of the survey rather than what it was at the time of death. Levels of agreement on age, race, Hispanic origin, marital status, and veteran status were similar to those found in two studies in which Census Bureau Population Study interview responses were compared to death certif- icate entries (17-20). The consistency rates in reporting on occupation and industry were 71.0 percent and 74.4 percent, respectively. These low levels are consistent with prior research (3-16). The disagreements were not random for occupation: For all occupational categories except managerial and profes- sional, the percent of decedents in the category was the same or higher for the questionnaire than for the death certificate. However, there were 26.8 percent fewer man- agers and professionals on the questionnaire. In contrast to occupation, marginal distributions for industry were very similar for the death certificate and the questionnaire. It is possible that coding differences may have been a significant factor in the lack of correspondence in occupa- tion and industry between the two sources. Coding many occupation and industry entries that were very general such as “telephone” and “farm” was difficult. The source documents were not reviewed to determine whether dif- ferences were due to respondent reporting or to coding. There was good correspondence when the death cer- tificate place of death was “hospital inpatient,” but less consistency for entries reported on the death certificate as “hospital outpatient” or “emergency room.” There was very good correspondence for entries of health care insti- tutions other than hospitals on the death certificate. Overall, high rates of consistency between the ques- tionnaire and death certificate should add confidence in the interpretation and use of mortality statistics. However, even when marginal distributions are very similar, lower rates of agreement raise concern about possible biases in the mortality data. For example, American Indian dece- dents unidentified as such on the death certificate may have different characteristics from those identified as American Indian on the questionnaire. On the other hand, differences in marginal distributions do not neces- sarily lead to biases in assessing relationships among specific variables. If the data were weighted to produce national estimates of the degree of overall comparability, these rates would be slightly higher in general because there was oversampling of some groups that had lower rates of agreement. Through the use of the 1986 NMFS, it is possible to explore further the types and possible directions of poten- tial biases in the relationships among variables. Additional analyses could also include examining comparability ac- cording to other important control variables including age, sex, and cause of death. The standard death certificate was revised for use starting in 1989. It will be important for the next NMFS, planned for 1993, to investigate whether there are any changes in the levels of consistency in reporting. 13 References 10. 11. 12. Gittelsohn A, Royston PN. Annotated bibliography of cause- of-death validation studies, 1958-80. National Center for Health Statistics. Vital Health Stat 2(89). 1982. Rosenberg HM. The nature and accuracy of cause-of-death data report of the Workshop on Improving Cause-of-Death Statistics. National Committee on Vital and Health Statis- tics. National Center for Health Statistics. 1989. Buechley R, Dunn JE, Linden G, Breslow L. Death certifi- cate statement of occupation: its usefulness in comparing mortalities. Public Health Rep 71:1101-11. 1956. Alderson A. Some sources of error in British occupational mortality data. Br J Ind Med 29:245-54. 1972. Wegman DH, Peters JM. Oat cell lung cancer in selected occupations. J Occup Med 20:793-96. 1978. Frazier TM, Wegman DH. Exploring the use of death certificates as a component of an occupational health sur- veillance system. Am J Public Health 69:718-20. 1979. Rosenberg HM, Burnham D, Spirtas R, Valdisera V. Infor- mation from the death certificate: assessment of the com- pleteness of reporting. In DelBene L, Scheuren F, eds. Statistical uses of administrative records with emphasis on mortality and disability research, pp. 83-9. Washington: Social Security Administration, Office of Research and Statistics. 1979. Rousch GC, Meigs JW, Kelley J, et al. Sinonasal cancer and occupation: a case control study. Am J Epidemiol 111:183- 93. 1980. Swanson GM, Schwartz AG, Burrows RW. An assessment of occupation and industry data from death certificates and hospital records for population-based cancer surveillance. Am J Public Health 74:464-67. 1984. Balarajan. Comparison of occupations recorded at cancer registration and death. Public Health 99:169-73. 1985. Gute, Fulton JP. Agreement of occupation and industry data on Rhode Island death certificates with two alternate sources of information. Public Health Rep 100:65-72. 1985. Steenland K, Beaumont J. The accuracy of occupation and industry on death certificates. J Occup Med 26:288-96. 1984. 13. Schumacher MC. Comparison of occupation and industry information from death certificates and interviews. Am J Public Health 76:635-37. 1986. 14. Turner DW, Schumacher MC, West DW. Comparison of occupational interview data to death certificate data in Utah. Am J Ind Med 12:145-51. 1987. 15. Davis H. The accuracy of industry data from death certifi- cates for workplace homicide victims. Am J Public Health 78(12):1579-81. 1988. 16. Schade WIJ, Swanson GM. Comparison of death certificate occupation and industry data with lifetime occupational histories obtained by interview: variations in the accuracy of death certificate entries. Am J Ind Med 14(2):121-36. 1988. 17. Sorlie PD, Rogot E, Johnson NJ. Validity of demographic characteristics on the death certificate. Epidemiology 3(2):181-84. 1992. 18. Hambright TZ. Comparability of age on the death certifi- cate and matching census record, United States, May- August 1960. National Center for Health Statistics. Vital Health Stat 2(29). 1968. 19. Hambright TZ. Comparability of marital status, race, nativ- ity, and country of origin on the death certificate and matching census record, United States, May-August 1960. National Center for Health Statistics. Vital Health Stat 2(34). 1969. 20. McCarthy MA. Comparison of the classification of place of residence on death certificates and matching census records, United States, May-August 1960. National Center for Health Statistics. Vital Health Stat 2(30). 1969. 21. Seeman I, Poe GS, Powell-Griner E. Development, meth- ods, and response characteristics of the National Mortality Followback Survey, 1986. National Center for Health Statis- tics. Vital Health Stat 1(29). 1993. 22. National Center for Health Statistics. Public-use data tape documentation: National Mortality Followback Survey, 1986. Hyattsville, Maryland: Public Health Service. 1988. 23. U.S. Bureau of the Census. 1980 census of population: alphabetical index of industries and occupations. Washing- ton: U.S. Department of Commerce. 1982. List of detailed tables . Number of responses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986. .................. . Number of responses by race of decedent on National Mortality Followback Survey questionnaire and by race on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986... coi iiiiiinn.. . Number of responses by Hispanic origin of decedent on National Mortality Followback Survey question- naire and by race and Hispanic origin on death certif- icate, interval between death and survey, and relationship of informant to decedent: United States, T0800 5.5 me 0: win m0 0 m0 0 1 mt os we 48 A . Number of responses by marital status of decedent on National Mortality Followback Survey questionnaire and by race and marital status on death certificate, 16 25 28 interval between death and survey, and relationship of informant to decedent: United States, 1986 ........ . Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, inter- val between death and survey, and relationship of informant to decedent: United States, 1986 ........ . Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of infor- mant to decedent: United States, 1986 ............ . Number of responses by veteran status of decedent on National Mortality Followback Survey questionnaire and by race and veteran status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 ........ 30 33 38 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over Race and age All races: ABOLS vows ew nnw 15,964 709 732 922 1,131 641 1,070 795 1,132 1,322 1,627 1,735 1,629 1,315 Under 30 years ...... 696 690 6 - - - - - - - - - - - 30-34years ........ 756 17 703 30 1 - 1 3 - - - - - 1 35-30years ........ 915 - 19 876 15 3 - 1 - 1 - - - - 40-44 years ........ 1,141 1 1 15 1,095 27 1 - 1 - - - - - 45-49vyears ........ 635 - 2 - 15 600 17 1 - - - - - - 50-54years ........ 1,096 - - 1 1 10 1,026 53 3 1 1 - - 55-59years ........ 765 - - - 1 - 23 710 20 7 3 1 - - 60-64vyears ........ 1,124 - - - 2 - 2 18 1,088 41 4 4 - - 65-69years ........ 1,344 - - - 1 - - 7 48 1,233 42 10 1 1 70-74 years ........ 1,656 1 1 - - - - — 5 29 1,518 90 6 4 75-79years ........ 1,743 - - - - - - 1 2 8 50 1,597 67 16 80-84years ........ 1,647 - —- - —- - - 1 —- 1 8 28 1,512 85 85-89years ........ 1,285 - - - - 1 - - - 1 2 3 37 1,193 90-94 years ........ 791 - - - - - - - - - - 1 3 14 95 years and over. . . . . 370 - - - - - - - - - - - 3 1 White: AlBges ...cccvevvrns 11,303 515 516 678 863 435 781 474 716 864 1,124 1,224 1,217 991 Under30vyears ...... 506 503 3 - - - - - - - - - - - 30-34years ........ 533 1 503 16 - - - 2 - - - — - 1 35-39years ........ 670 - 9 651 8 — — 1 - 1 — - — - 40-44 years ........ 883 - 1 11 853 17 - - 1 - - - - - 45-49vyears ........ 423 - # - 1 414 8 wo - - - - - - 50-54years ........ 801 - - - 1 4 762 33 1 — - - - - 55-59years ........ 450 - - - - - 10 430 8 2 - - - - 60-64years ........ 702 - - - - - 1 4 676 20 1 - - - 65-69years ........ 874 - - - - - - 2 27 826 14 5 - — 70-74 years ........ 1,137 1 - - - - - - 2 11 1,078 44 1 - 75-79years ........ 1,219 - - - - - - 1 1 2 25 1,156 32 2 80-84years ........ 1,226 - —- - —- - - 3 - 1 6 16 1,161 38 85-89years ........ 986 - - - E& so - — - 1 - 2 21 945 90-94years ........ 617 - - - - - - - - - - 1 1 4 95 years and over. . . . . 276 - - - - - - - - - - - 1 1 Black: Allages ............ 4,117 165 191 203 232 177 246 282 375 421 456 457 359 292 Under 30vyears ...... 162 159 3 - - - - - - - - - - - 30-34years ........ 199 5 180 1 1 - 1 1 - - - - - - 35-39years ........ 203 - 6 188 6 3 - - - - - - - - 40-44 years ........ 223 1 - 3 208 10 1 - -— - - - — 45-49 years ........ 179 - 1 - 13 157 7 1 - - - - - - 50-54years ........ 250 - - 1 - 6 224 17 1 - - 1 - - 55-59years ........ 279 - - - 1 - 12 246 12 4 3 1 - - 60-64vyears ........ 381 - - - 2 - 1 13 338 20 3 4 - - 65-69vyears ........ 432 - = - 1 - - 4 20 375 24 5 1 1 70-74years ........ 472 - 1 - - - - - 3 17 398 42 5 4 75-79years ........ 467 — - — — - - - 1 5 24 391 30 14 80-84 years ........ 373 C= E - = or o = - gs 2 12 305 46 85-89years ........ 266 - - - - 1 - - - -— 2 1 15 217 90-94years ........ 152 - - - - - - - - - - - 1 10 95 years and over. . . . . 79 - - - - - - - - - - - 2 - 16 798 406 1 - 1 1 1 1 8 4 36 12 741 32 10 356 623 282 3 — 16 1 597 14 7 267 154 107 1 - 1 1 1 1 4 4 19 11 125 16 3 74 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Race and age on death certificate, interval, and Age at death on questionnaire 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over American Indian: ANBEEBI . «con vi cee mi 408 21 21 24 21 19 29 34 33 31 36 42 47 23 14 13 Under 30 years ...... 21 21 —- - - - - - - = ” wr ” - -— — 30-34years ........ 18 - 16 2 - - — - - i - - - - ” “ 35-39years ........ 26 - 4 21 1 - - - - - - - - - = - 40-44 years ........ 20 - - 1 19 - - - - - - - - — — — 45-49 years ........ 23 - 1 — 1 19 2 - - - - = — i a ~ 50-54years ........ 30 - - - - - 26 3 - 1 - - i - = - 55-59years ........ 31 - - - - - 1 29 - 1 - - = = = = 60-64 years ........ 34 - - - - - - 1 32 1 - = — — — — 65-69years ........ 31 - - - - - - 1 1 26 3 - - ws on - 70-74years ........ 36 - — - - - — - - 1 32 3 = = = ~~ 75-79 years ........ 46 - - - - - —- - - 1 1 39 5 = — 80-84years ........ 41 - - - - - - - = - — 41 ~ — — 85-89years ........ 24 - - - —- - - - - - ” - oe 23 1 = 90-94years ........ 16 - - - — - — - - w- ~ ~ 1 = 13 2 95 years and over. . . . . 1 - - - - - - - = ~ = = = - - 1 Other: Alages swiss mes 136 8 4 17 15 10 14 5 8 6 19 12 6 9 7 4 Under 30 years ...... 7 7 - - ~ - - se oe = = = - = —- az 30-34vyears ........ 6 1 4 1 - - - - = - = — — — — — 35-39years ........ 16 - - 16 - - - - - > - - - ; i” = 40-44 years ........ 15 - - - 15 - - - - - - = = ~ =~ — 45-49years ........ 10 - - - - 10 - - - —- = po — — — 50-54years ........ 15 - - - - - 14 - 3 - - - — — - - 55-59years ........ 5 - - - - - - 5 - pn” - - ~ = - 60-64years ........ 7 — - - - - - - 7 = = = = - - - 65-69years ........ 7 - - - - - - - - 6 1 - - a ” - 70-74 years ........ 11 - - - - - - - - - 10 3 - - - we 75-79years ........ Nn - - - - - - - - - - 11 = - = ” 80-84 years ........ 7 - - - - - = - = £= - 5 1 - 85-89years ........ 9 - - - - - - - - - a 1 8 - - 90-94years ........ 6 - - - = = = > = = i: = = 6 - 95 years and over. . . . . 4 - - - - - - - = - - ” - - - 4 Interval and age 22-25 weeks: Allages ............ 1,079 45 42 50 58 39 66 50 80 94 121 112 115 119 63 25 Under 30years ...... 43 43 - - - i ~ ~ - - - » - - - . 30-34years ........ 45 2 42 1 - = = ry - -~ ~ Ho = = — 35-39years ........ 50 - - 49 1 - —- - - — - - - ” we - 40-44 years ........ 59 — so — 57 2 - - - - - - - - - - 45-49years ........ 36 - - - - 36 - - - - on = - - = 50-54years ........ 70 - - - - 1 66 3 - - - - n = = — 55-59years ........ 46 - - —- - - - 46 -~ — — — — ~ — — 60-64years ........ 80 - - - - - - 1 78 1 - - ~ - i - 65-69years ........ 95 - wo — - - wr - 2 91 1 1 - - wi - 70-74 years ........ 124 - - - - - - - — 1 118 4 1 - —- aw 75-79years ........ 113 - - - - - —- - - 1 2 105 4 1 ~~ - 80-84years ........ 118 - - - - - - - - - - 2 109 5 2 - 85-89years ........ 115 - - - - - - - - - - - 1 111 2 1 90-94 years ........ 62 - - - - - - - - - ” = 2 58 2 95 years and over. . . . . 23 - - - - - - - - - = = = - 1 22 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over 26-29 weeks: AlBges uo ivsmrvnsns 2,200 63 85 121 150 89 139 105 143 197 204 276 225 200 127 76 Under30years ...... 63 63 - - = = ~ a - - = 5 oy . = _ 30-34vyears ........ 93 - 85 5 - - 1 2 - - - = - ~ i = 35-39years ........ 118 - - 115 2 - - - - 1 - - - - — - 40-44 years ........ 150 - - 1 146 3 - - - - - — - = - oe 45-49vyears ........ 89 - - - 1 86 2 - - - - - = i = = 50-54vyears ........ 136 - - - - - 133 3 - - - - = - = - 55-59years ........ 104 - - - - - 2 100 1 — 1 - - - - - 60-64years ........ 146 - - - 1 - 1 - 141 3 - - - — pov = 65-69years ........ - 197 - - - - - - - 1 189 4 2 - - 1 - 70-74 years ........ 208 - - - - - - - - 3 197 6 2 - - 75-79years ........ 276 - - - - - - - - 1 2 265 4 3 — 1 80-84years ........ 230 — — - - - - - - — - 3 215 9 3 - 85-89years ........ 196 - - - - - - - - - - - 5 184 3 4 90-94 years ........ 124 - - - - = - - - - ~ - - 2 119 3 95 years and over. . . . . 70 — - - - - ~ - - = - - 1 a 1 68 30-32 weeks: Al BGS uv svn sian ue 1,919 59 73 102 132 75 113 94 144 152 210 228 207 153 124 53 Under 30 years ...... 58 58 - - ~ = > - - ~ - = = = fe 30-34years ........ 77 1 71 4 - - - 1 - - - - - - - - 35-39years ........ 101 - 2 98 1 - - - - - = - = - = = 40-44 years ........ 134 - - - 131 3 - - - = = — = = — = 45-49years ........ 71 - = - - 71 —- - - - - - - = — = 50-54years ........ 115 - - - - 1 112 2 —- —- - - - - = = 55-59years ........ 93 - - - - - 1 91 1 - - - — = — i 60-64years ........ 142 - - - - - - - 140 2 = = or — = - 65-69years ........ 1565 - - - - - - - 3 148 3 - - 1 — = 70-74 years ........ 220 - - - - - - — - 2 203 15 - - - = 75-79years ........ 222 - - - - - - - - - 4 211 6 1 - - 80-84years ........ 205 - - - - - - - - - - 2 199 3 — 1 85-89years ........ 154 - - - = - = 4a - = = ' 2 146 5 1 90-94vyears ........ 123 - - - - - = - - - - ~ = 2 117 4 95 years and over. . . . . 49 - - - - - - - - ~ = = = - 2 47 33-35 weeks: AlAGeE «usw usinn an 2,343 115 88 115 167 76 139 111 172 216 255 238 240 218 121 72 Under 30 years ...... 115 114 1 — - - - - - = = - = - - = 30-34years ........ 86 1 84 1 - - = tt ~ ~ , — ~ - = 35-39years ........ 115 - 2 112 1 - - — - - - - - - — - 40-44 years ........ 167 - - 2 164 1 - - — - - - ~ - - a 45-49vyears ........ 76 - 1 - 4 72 2 - - - - - =; - - = 50-54years ........ 149 - - - 1 2 137 9 - - - - - - - - 55-59years ........ 105 - - - - - - 100 5 - ~ — — — - = 60-64years ........ 166 - —- - - - - - 159 5 1 1 - = - - 65-69years ........ 222 - - - - - - 2 7 207 5 1 - - = = 70-74 VBS ....« vou 264 - - - - - —- - 1 4 242 14 2 — 1 — 75-79years .....\. 240 - - - - - - - - - 7 221 10 2 - - 80-84years ........ 239 - - - - - - - - - - 1 220 15 3 85-89 years ........ 212 - - - - 1 - - - - - - 5 200 4 2 90-94 years ........ 117 - - - - - - - - - - - 1 1 111 4 95 years and over. . . . . 70 - - - - - - - - - - - 2 - 2 66 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Race and age on death certificate, interval, and Age at death on questionnaire 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over 36-38 weeks: Allages ............ 2,061 93 96 119 146 81 121 102 164 149 223 251 217 165 90 44 Under 30 years ...... 90 90 - - - - - - wi — i” - - 5 - — 30-34years ........ 98 2 92 4 - - - — = a - i = - - 35-39years ........ 117 - 4 110 3 - — - - - - - - = ~ - 40-44 years ........ 162 1 - 142 3 1 - 1 - - - - w 3 os 45-49years ........ 79 - - - - 78 1 - - - - - - — 50-54years ........ 123 - - 1 - - 114 7 - - - = he - ” 55-59years ........ 102 - - - - - 5 91 2 1 2 1 - - - - 60-64years ........ 163 - - - — - - 2 152 6 1 2 - - - - 65-69vyears ........ 154 - - - 1 - - 1 8 136 7 1 - - - - 70-74years ........ 226 - - - - - - - - 5 207 11 1 1 =~ 1 75-79years ........ 245 - — - - - - — oe 1 4 231 5 4 - = 80-84years ........ 222 - —- - - - - 1 - - 1 4 204 11 - 1 85-89years ........ 162 - - - - - - - - - 1 1 7 147 5 1 90-94years ........ 94 - - - - - - = = = w= = - 2 85 7 95 years and over. . . . . 34 - - - - - - - - - - = we = 34 39-41 weeks: Alagss .....vse5:%3 1,850 87 100 111 141 75 156 88 123 171 179 174 190 122 83 50 Under 30 years ...... 84 84 - - - - = = = - - - - = = = 30-34years ........ 101 3 93 4 1 - - - - = S w - - - — 35-39years ........ 113 - 6 105 2 - - - - - = = ort - w< = 40-44 years ........ 140 — - 2 134 4 - - - - - a - = 0 ™ 45-49vyears ........ 78 - 1 - 4 70 2 1 - — a - = - = — 50-54vyears ........ 159 - - - 1 149 7 1 - - 1 - —- - - 55-59vyears ........ 81 - - - - - 4 74 2 1 - - - - - = 60-64years ........ 128 - - - - - 4 3 115 8 1 - - = = — 65-69years ........ 162 = = - = - = 3 153 4 ~ - - _ _ 70-74 years ........ 183 - - - - - - - 4 6 163 13 —- - - - 75-79 years ........ 179 - - - - - - 1 1 3 10 156 7 1 - = 80-84vyears ........ 197 - - - - - — — - - 1 3 178 14 - 1 85-89years ........ 112 - - - - - - - = - — — 5 105 1 1 90-94years ........ 88 — - — -— - — — — -— - 1 - 2 82 3 95 years and over. . . . . 45 - - - - - - - - - = - - go - 45 42-44 weeks: ANEgBS «crv ivniwis nn 1,635 79 69 108 119 70 116 100 110 130 156 169 168 128 84 29 Under30vyears ...... 78 76 2 - - - - — - ” i - = = — _ 30-34years ........ 73 3 66 4 - - - - - - a = i —- ” 35-39years ........ 102 - 1 99 - 1 - - - - -— - o. - = 40-44 years ........ 118 - - 5 112 1 - - - - - - i as = — 45-49years ........ 77 - - - 6 67 4 - - - = a“ - - - 50-54years ........ 113 - - - - 1 107 5 - = = - 4 = = 55-59years ........ 102 - - - 1 - 5 89 3 4 - — — - ot = 60-64 years 108 - - - - —- - 4 97 7 - - i - - — 65-69vyears ........ 136 - - - - - - 1 8 116 7 3 1 - - - 70-74years ........ 148 - - — — — — - 1 3 134 9 1 - - - 75-79years ........ 177 — - —- — — - - 1 oe 10 151 11 3 1 - 80-84years ........ 168 - - - - - - - - 4 5 150 8 - 85-89vyears ........ 128 = — - ws — os = = an 1 1 4 116 6 - 90-94 years ........ 78 - - - - - - - = = - = 1 = 75 2 95 years and over. . . . . 29 - = = - = - - = _ _ _ _ 1 2 26 19 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —-Con. Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over 45-47 weeks: AVBGBE viv rmrvwrme 1,182 67 78 85 91 53 94 58 86 84 11 111 106 92 40 26 Under 30 years ...... 63 63 - - - - ~- = oo a = ~ = = -~ — 30-34years ........ 82 3 77 2 - - - - - - - - - = = a” 35-39years ........ 85 - - 83 2 - - - - - - - So » - - 40-44 years ........ 92 - or - 87 5 - - - - - - - - - 45-49years ........ 51 - - - 2 47 2 - - - - - - — - - 50-54years ........ 99 - - - - 9 91 6 1 - - - - vs ” = 55-59years ........ 53 - - - - 1 49 3 - - - - - we - 60-64years ........ 86 - - - - - - 3 75 6 1 1 - - - —- 65-69years ........ 88 - - - - - - - 7 75 5 1 - - - - 70-74 years ........ 110 1 1 - - - - - - 1 929 8 - — ~ - 75-79years ........ 13 - - - - - - - - - 5 99 8 1 - - 80-84years ........ 103 - KX - - - - - - 1 1 2 96 3 - - 85-89years ........ 96 - - - - - - - - 1 - i 1 86 6 2 90-94 years ........ 39 - - - - - - = - - we ~ 1 2 34 2 95 years and over. . . . . 22 - - - - - - - - - = = -~ a - 22 48-51 weeks: ANOGES 7 suits sm sins 856 54 47 60 59 32 65 43 56 65 84 97 93 55 33 13 Under 30 years ...... 55 54 1 - - = - = Pe i - - - ne - 30-34years ........ 48 - 45 3 - - - - - 52 i = ~ ~ - - 35-39years ........ 61 - 1 57 2 1 - - —- - - - - - - = 40-44 years ........ 56 - - - 56 - - - - - - - vit - - a 45-49vyears ........ 31 - - - 1 28 2 - - - - = = i: -~ o 50-54years ........ 73 - — —- - 3 61 8 - 1 - - - oe — 55-59years ........ 36 - - - - - 2 31 2 1 - - - = wt — 60-64years ........ 53 - - - - - - 3 47 3 a - - = = - 65-69years ........ 68 - - - - - - 1 6 57 4 - - - - - 70-74 years ........ 85 - - = - - - - 1 3 76 3 1 1 - a 75-79 years ........ 100 - - - - - - - - - 4 90 6 - a - 80-84years ........ 92 - - - - - - - - - - 3 82 7 - = 85-89years ........ 52 - - - - - - - - - - 1 4 46 1 90-94 years ........ 34 - - - - - - - - a = @ = 1 31 95 years and over. . . . . 12 - - - - - - - - - ” a ” os 1 11 52 weeks and longer: AIBEES ovina waiin ae 824 47 52 51 67 50 59 44 54 62 81 79 66 61 33 18 Under 30years ...... 47 45 2 - - - = - - - = & " “ — - 30-34years ........ 51 2 46 2 - - - - - - - - - 1 = - 35-39years ........ 53 - 3 48 1 1 gs - - - - - - - - we 40-44 years ........ 72 - 1 1 65 5 - - - - - - - wt = 45-49years ........ 46 - - - - 44 2 - - - - - - i = - 50-54years ........ 57 - - —- - - 54 3 - - - - — - = — 65-59 years ........ 43 —- - - - - 3 39 1 - - - ” - = - 60-64years ........ 52 - —- - 1 - - 2 49 - - - - = - o 65-69years ........ 65 - - - - - - - 3 59 2 1 - - - - 70-74 years ........ 86 - - - - - - - 1 1 77 7 - - - - 75-79Y0I8 . . vu vv 77 - - - - - - - - 2 1 68 6 - - - 80-84years ........ 71 - - - - - - - - - 1 3 57 10 - = 85-89years ........ 56 - — - - - - - - - ps - 3 50 3 90-94 years ........ 32 - - - - - - - - - - - po 29 3 95 years and over. . . . . 16 - - - - - - - = = - a or ” 1 15 20 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Race and age on death certificate, interval, and Age at death on questionnaire 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over Missing: Alages ..v.unvvimee 15 - 2 - 1 1 2 - - 2 3 - 2 2 - - 30-34years ........ 2 - 2 - - - - - ~ = - - = = _ 35-39years ........ - - - - - - - - —- ~ = - = = - - 40-44 years ........ 1 - - - 1 —- - - -~ - - - - - - —- 45-49vyears ........ 1 - - - - 1 - ~ - ” - ve = - - 50-54years ........ 2 - - - - - 2 - - - - = oe? oo = - 55-59years ........ - - - - - - - - - - ” - = - - od 60-64years ........ - - - - - - - - - - - we - - - - 65-69years ........ 2 - - - - - - - - 2 = = - - - = 70-74 years ........ 2 - - - - - - - - - 2 - we Le. - — 75-79Y08I8 . « «vv x vs 1 - - - —- - - - - - 1 - - wi ” - 80-84years ........ 2 - - - - - = - = - - = 2 - - - 85-89years ........ 2 - - - - - - - - = Pe = 2 = - Relationship and age Decedent was death certificate informant's — Spouse: ALBOBE «ivr tin iin mw 5,997 164 245 388 552 318 579 429 561 640 726 663 440 218 56 18 Under 30 years ...... 160 158 2 - - - - - = P= ho. - ” - i at 30-34years ........ 252 6 234 11 - - - 1 - - - = a - po = 35-39years ........ 382 - 7 370 3 1 - 1 - - - - - - ” w 40-44 years ........ 564 - - 6 543 14 - - 1 - - - = - — — 45-49years ........ 309 - 1 - 4 300 4 - - - - - - - I” - 50-54years ........ 599 - - 1 - 3 567 27 1 - - - oe - po = 55-59years ........ 410 - - - - - 6 395 7 2 - - - - wr - 60-64years ........ 564 - - - 1 - 2 3 541 15 2 - - —~ — - 65-69years ........ 645 - - - 1 - - 2 11 613 14 4 - - - - 70-74 years ........ 732 - 1 - - - - - - 6 693 28 4 - - = 75-79 years ........ 659 - - - - - - - - 4 12 625 18 - - - 80-84 years ........ 438 - - - — - Co i - - 4 5 413 15 - 1 85-89years ........ 213 - - - — = - - - - 1 1 5 201 4 1 90-94 years ........ 59 - - - - - - - - - - = 2 52 5 95 years and over. . . . . 7 - - - - - - - - - - = - - I 11 Parent: AVages cams mvaman 1,172 335 257 214 188 57 51 16 16 6 7 5 6 8 5 1 Under 30 years ...... 332 330 2 - - - = - - wi - = = — — 30-34years ........ 265 4 251 8 - - 1 1 - - - - i - - = 35-39years ........ 21 - 4 202 4 - - - - 1 - - i ” = 40-44 years ........ 188 1 - 4 180 3 - - - - - - i Go = - 45-49 years ........ 60 - - - 3 53 4 - - - ~ - — = 50-54years ........ 49 - - - 1 1 45 2 - - - - - - ~ - 55-59years ........ 14 - - —- - - 1 13 - = - - = -r - 60-64years ........ 16 - - - - - - - 16 - - - —- a. - — 65-69years ........ 5 - - - - - - - - 5 = - a — — - 70-74 years ........ 7 - - - - - - - - - 7 = = - - = 75-79years . ....... 5 - - - - - - —- - - - 5 - - - - 80-84years ........ 8 - - - - - - - - - = = 6 2 — _ 85-89years ........ 6 —- -. - - - - - - - - - - 6 - — 90-94 years ........ 5 - - - - - - - = - - - pe - 5 a= 95 years and over. . . . . 1 - - - - - - - - = = - = = = 1 21 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over Child: AN BEES « wivu ssw sw he 1,759 3 3 16 37 41 88 67 125 148 194 223 268 259 187 100 Under 30 years ...... 2 2 - - - - — - = - = _- z= _ _ _ 30-34years ........ 4 8 3 = = oe - - - - - - - 1 = ot 35-39years ........ 16 - - 16 - - = 2 a - = ~ = - y — 40-44 years ........ 39 - - - 36 3 - - - - - - - - — 45-49years ........ 39 - - - 1 37 1 - - - - - - - we = 50-54years ........ 94 - - - - 1 83 7 2 - - 1 - - - - 55-59 years ........ 61 - - - - - 4 85 1 1 - - - - ~ — 60-64years ........ 126 - - - - - - 3 112 1 - - - oe = i 65-69years ........ 150 - - - - - - 2 8 132 5 3 - - - - 70-74years ........ 197 1 - - - - - - 2 4 179 9 - 1 1 - 75-79years ........ 228 - - - - - - - - - 9 207 10 2 - - 80-84years ........ 280 - - - - - - - - - 1 3 256 18 1 1 85-89years ........ 246 - - - - - — — — - - - 2 233 7 4 90-94years ........ 188 - - — - — — = oe = - re 4 174 10 95 years and over. . . . . 89 - - - - - - - - - - i = = 4 85 Sibling: AIBgeSs . «ios uvsvimsun 608 36 48 57 60 45 40 30 54 49 55 52 49 22 9 2 Under 30 years ...... 36 36 - - - - ~ - ” wi a - = — oe 30-34years ........ 48 - 45 2 - - - 4 - - - ~ w - - = 35-39years ........ 57 - 2 53 1 1 - - - - - - - = - — 40-44 years ........ 61 - 1 2 56 1 1 - - - - - - ” - -— 45-49vyears ........ 44 - - - 3 39 2 - - - - ~ - jo = = 50-54years ........ 41 - - - - 4 34 3 - - ~ ~ wr - = 55-59years ........ 33 - —- - - - 3 26 2 2 - - = = — — 60-64years ........ 50 - - - — — oe — 48 - 1 1 - - we — 65-69years ........ 52 - - - - - - - 3 46 3 - - - - - 70-74 years ........ 50 - - - - - - - - 1 47 2 = = = = 75-79years ........ 53 - - - - = - - ¥ w- 3 46 2 4 — - 80-84years ........ 46 - — — - - - - - - 1 2 43 - - — 85-89years ........ 26 - - - - - - - - - - 1 4 21 ” a 90-94 years ........ 9 - - - - - - - - = = - & 9 - 95 years and over. . . . . 2 - - - - - - - - = = — = - — 2 Other relative: Alages sus vwsnnims 463 1 12 16 10 7 13 6 14 29 37 56 80 80 64 28 Under30 years ...... 10 9 1 - = - - - i - - ~ - = = 30-34years ........ 15 2 19 2 - - - - ~ i - = - =~ = 35-39years ........ 15 - - 14 1 - - —- - - - - - — — - 40-44 years ........ 8 - = - 8 - a - - - = _ = _ _ 45-49vyears ........ 10 - - - 1 7 1 1 - - - - —- a - - 50-54years ........ 1 - - - - - 11 - - - - - -— - —- — 55-59years ........ 7 - - - — - 1 5 1 - - - — — — — 60-64years ........ 14 - - - - - - - 1 2 - 1 - = i - 65-69years ........ 27 - — - - - - — 1 25 1 - - = ~ - 70-74 years ........ 43 — - - — - - - - 2 35 5 - 1 - = 75-79years ........ 55 — - - - - - - 1 - - 47 6 1 - - 80-84years ........ 85 - - - - - - - - 1 2 69 12 4 5 85-89years ........ 77 - - - - - - - - = ~ 1 4 65 6 1 90-94 years ........ 60 - - - - - - - = = - = - 1 56 3 95 years and over. . . . . 26 - - - - - - - - w- = - 1 - 1 24 22 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over Nonrelative: AlBOBS «ox wena maa 378 14 9 22 39 24 20 21 26 24 33 37 43 35 19 12 Under30 years ...... 13 13 —- - - - - - ” = = a = = _ — 30-34years ........ 10 1 8 1 - - ~ - - - = - - - _ - 35-39years ........ 21 - 1 20 - - - - - - - - - = = 40-44 years ........ 40 - - 1 39 - - - - - - - = = = = 45-49 years ........ 25 - - —- - 24 1 - - — = i - - = — 50-54years ........ 19 - - - - — 18 1 - - - —- - - - - 55-59years ........ 23 - — — - - 1 20 2 - = = & = = = 60-64years ........ 24 - - - - - - - 24 = rr — - = — — 65-69years ........ 21 - - - - - - - - 20 1 - - pe i. - 70-74 years ........ 34 - - - - - — - — 1 31 2 - - - = 75-79years ........ 35 - - - - - - - - 1 1 32 1 - - 80-84years ........ 47 - - - - — - - - 1 - 3 40 3 - - 85-89years ........ 34 - - - - - - - — 1 - - 1 31 1 - 90-94 years ........ 19 - - - - - - - - - = = = = 18 1 95 years and over. . . . . 13 - - - - - - - - - = - 1 1 a 11 Not stated: AUBOBS 4 ws smnma rims 5,587 146 158 209 245 149 279 226 336 426 575 699 743 693 458 245 Under 30 years ...... 143 142 1 - - - ~ —- fie = = ~ = = _ _ 30-34years ........ 162 4 151 6 1 - - - - - - — - nt - - 35-39years ........ 213 - 5 201 6 1 - - - - - - - - - = 40-44years ........ 241 - - 2 233 6 - - - = = - we = = ~ 45-49years ........ 148 - 1 - 3 140 4 - - - - - — -— - - 50-54years ........ 283 - - - - 1 268 13 - 1 - - - - - - 55-59years ........ 217 - - - 1 - 7 196 7 2 3 1 - - -~ = 60-64years ........ 330 - - - 1 - - 12 301 13 1 2 - - - = 65-69years ........ 444 - - - - - - 3 25 392 18 3 1 1 - 70-74 years ........ 593 - - - - - - - 3 15 526 44 2 2 - 1 75-79years ........ 708 - - - - - - 1 - 3 25 635 30 12 1 1 80-84years ........ 743 - - - - - - 1 - 1 13 685 35 6 2 85-89years ........ 683 - - - - 1 - - - - 1 - 21 636 18 6 90-94 years ........ 451 - - - - - - - - - - 1 3 7 427 13 95 years and over. . . . . 228 - - - - - - - - - - - 1 - 5 222 Death certificate informant and survey respondent were — Both spouse: Alages ............ 5,083 116 193 317 483 252 494 372 505 578 632 559 360 167 40 15 Under 30 years ...... 116 115 1 - - - = = - a ” ”» ~ = = - 30-34years ........ 199 1 190 7 - - —- 4 - - - - - ~~ — 35-39years ........ 311 - 2 305 3 - - 1 - - - - = - - - 4044 years ........ 492 - - 4 478 9 - - 1 - - - - - - - 45-49years ........ 246 - —- - 1 241 4 - - = = = = - - 50-54years ........ 510 - - 1 - 2 484 23 - —- —- - = jo = — 55-59years ........ 354 - - - - - 4 343 5 2 - - - — - — 60-64years ........ 505 - - - - — 2 2 491 8 2 - - - - - 65-69years ........ 583 - - - 1 - - 2 8 559 10 3 - - - - 70-74years ........ 639 - - — - — — - - 5 608 23 3 - i - 75-79years ........ 556 - - - - — me - - 4 9 527 16 —- = - 80-84years ........ 357 - - - - - - - - - 3 5 337 11 1 85-89years ........ 163 - - - - - — — - as a 1 4 154 3 1 90-94years ........ 43 - - - - - — - or ie - - - 2 37 4 95 years and over. . . . . 9 - - - - - — - > > - = - = = 9 23 Table 1. Number of reponses by age of decedent on National Mortality Followback Survey questionnaire and by race and age on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Age at death on questionnaire Race and age on death certificate, interval, and 30 years 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95 years relationship Total and below years years years years years years years years years years years years years and over Not both spouse: ANBIBE ivy wimt ons 10,370 580 520 580 606 377 533 392 563 682 921 1,120 1,230 1,127 Under 30 years ...... 567 562 5 - - - - - = Po i - = 30-34years ........ 538 16 495 22 1 - 1 2 - - - - - 1 35-39years ........ 580 - 17 547 12 3 - - - 1 - - = - 40-44 years ........ 607 1 1 1 575 18 1 - - = ” = 5 45-49years ........ 377 - 1 — 14 348 13 1 - - - — - - 50-54years ........ 544 - - - 1 8 501 29 3 1 1 - - 55-59years ........ 379 - - - 1 - 17 337 15 5 3 1 - - 60-64years ........ 556 - - - 2 - - 16 502 30 2 4 - - 65-69years ........ 695 - - - - - - 5 36 615 29 7 1 1 70-74 years ........ 945 1 1 - - - - - 5 24 841 65 3 3 75-79years ........ 1,131 - - - - - - 1 2 4 39 1,016 51 16 80-84 years ........ 1,248 - - - - - - 1 - 1 5 23 1,136 72 85-89years ........ 1,104 - - - - —- - - - 1 2 2 33 1,022 90-94 years ........ 741 - - - - - - - - - = 1 3 11 95 years and over. . . . . 358 - - - - - - - = - i” ” 3 1 Not stated: AIRGSS «ius runansivs 511 13 19 25 42 12 43 31 64 62 74 56 39 21 Under 30 years . ..... 13 13 - - - = P= ? = = = p = o 30-34years ........ 19 - 18 1 - - - - - - - se - - 35-39years ........ 24 - - 24 - - - - - - or - ~ ne 40-44 years ........ 42 - - - 42 - - - - - - = - = 45-49years ........ 12 - 1 - - 1 - - - ~ > - - > 50-54years ........ 42 - - - - - 41 1 - - - - = = 55-59years ........ 32 - - - — - 2 30 - - - - a ~ 60-64years ........ 63 - - - - - - - 60 3 i ge = = 65-69years ........ 66 - - - - - - - 4 59 3 - - = 70-74 years ........ 72 - - - - - - - - = 69 2 = 1 75-79years ........ 56 - - - - - - - - = 2 54 wo = 80-84years ........ 42 - - - - - - = = - e a 39 2 85-89years ........ 18 - - - - 1 - - - - - - - 7 90-94 years ........ 7 - - - - - - - ~ - oi - - 1 95 years and over. . . . . 3 - - - - - - = 5 = = f= — 751 388 1 - 1 1 1 1 7 3 33 11 698 28 10 344 7 3 1 = 6 = - 3 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 24 Table 2. Number of responses by race of decedent on National Mortality Followback Survey questionnaire and by race on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Race on questionnaire Race on death certificate, interval American and relationship Total White Black Indian Asian Other Race AVIEEOR «oi 5s vs min 6 0 Wale 0 15,972 11,167 4,060 514 147 84 WHHS. o: + vie oi 3 0 #00 ip 4 0 win iw 4 11,316 11,109 36 86 23 61 BIAOK. «v.05 % biveiis wm pa pew 4,101 28 4,021 34 5 13 American Indian. . . ........ 422 24 2 392 1 3 ASIN. 4 omen dive "HE 131 5 1 2 116 7 OBE. «ov 0k iw 00000 iw 00 0 3 1 - - 2 - Interval and race 22-25 weeks: AICO8 «15 vw 55 ww sb vl iva 1,084 812 236 27 9 - WIR. «iv + nan wins po mien 4 819 811 1 5 2 - BIaCK. + «+ «oe Arey 237 1 235 1 - - American indian. .......... 21 - - 21 - - Aslan. .... HE vo ot an 7 - - - 1 - Other. . ... Fron on aA ew ER - - - - - - 26-29 weeks: Araces . ..csesemsms sass 2,246 1,656 496 7 21 2 Whe: . vcocasensnramane 1,668 1,651 3 11 3 - BAK, «50 ie 00 1 0 we 0 5 498 1 492 2 1 2 American Indian. . ......... 60 2 1 57 - - ABIBIV: ov msn 5 mis mnie wane 20 2 - 1 17 - OBE, sums mgr sms mE WS - - - - - - 30-32 weeks: AUTA0OS vs ovis wisn sm bems 1,946 1,477 409 40 19 1 WHS. «ovo sms mmsimammpns 1,482 1,470 4 3 5 - BIBOK, vc civ cmsmmimsmn ims 412 3 405 3 - 1 American indian. .......... 38 4 - 34 - - BSB vss serra 13 - - - 13 - OBE. wi vv vv iso 000 08 (8 avin 1 - - - 1 - 33-35 weeks: AIIBOBS « ¢ vv vo vnsnnenrnms 2,353 1,662 588 73 23 7 WIRE, «vos smsmennsmvons 1,676 1,651 4 13 3 5 TE 598 6 584 4 2 2 American Indian. . . ........ 60 4 - 55 1 - ASI, « oe sens menu dns ne 19 1 - 1 17 - Other. ova ns musnsorsrs - - - - - - 36-38 weeks: AVIBCO8 «v4 v5 v.04 vin sion wh 2,053 1,392 556 72 21 12 WB, ovo 3 nvm 590 50 vd ve 1,409 1,384 2 12 4 7 BlaoK. «i. 5% 20 ows ine be 2 567 4 553 6 1 3 American Indian. . ......... 59 3 1 54 - 9 FL I 17 1 - - 15 1 OMVET. 4605 oc 0 ich 0rd 0 06 0 4 1 - - - 1 - 39-41 weeks: AIIaces . ...ovvcvvsnnvnsmn 1,832 1,189 535 77 20 11 WHS, vos snr vnsassasmn 1,216 1,186 7 11 2 10 Black: + «vs 500 94 % #1 4% BP Wa 533 1 527 5 - - American Indian, . ......... 63 1 - 61 - 1 LT A EM gn 20 1 - 18 - OMe. ws wrnsi ans eres me - - —- - - - 42-44 weeks: AUTBOBE , ov vwnivs vue " 1,623 1,066 470 63 8 16 WEIS. oo vv rms mums ms sos 1,095 1,059 9 13 - 14 Black: .:vconmisesniminny 476 5 461 9 - 1 American Indian. . ......... 43 2 - 41 - - ASI. i ¢ (0: 5 on iw 300 dos 02 4 J 9 - - - 8 1 OWNER: vcr emis saii mame - - - - —- - 25 Table 2. Number of responses by race of decedent on National Mortality Followback Survey questionnaire and by race on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Race on questionnaire Race on death certificate, interval American and relationship Total White Black Indian Asian Other 45-47 weeks: AIBOBS . x vs maiwmams na aie» 1,163 765 338 33 13 14 WHHB. . vv sms mmis ins smsma 4 782 756 3 10 2 11 BIACK: sw swan mams nmsma 344 4 335 2 1 2 American Indian. .......... 26 5 - 21 - - A818: ny sme mE EERE 11 —- - - 10 1 OST. s ws vwsmns ms saswan - - - - - - 48-51 weeks: Allraces . . ............... 846 541 260 32 4 9 WWIAHE. & ou mn oii i005 00 ew 3 550 535 3 5 1 6 BIACK:: » 4 5: ovis iw 0 5 302 4 00 5 262 3 237 2 - - American Indian. . . ........ 29 3 - 25 - 1 ABBA. . vss vrs ms Ba dH 5 - - - 3 2 OBE. + ovo wie ow 50k 5 0 i 0 92 4 - - - - - 52 weeks and longer: Allraces . . . .............. 811 593 171 26 9 12 WIE, «wv vn iim Bis nn wore 604 592 - 3 1 8 Black: . 2 sons wsisma misono 173 - 171 - - 2 American indian. .......... 23 - - 23 - - ASIN 5 ir ve vn pms ae ae 10 - - - 8 2 ORBY:. « vein oi # ies ie 6 0 90d fo 1 1 - - - - Missing: AUHTBEEE vvvo0 0 0 08 wins 0b 15 14 1 - - - WHRHEB., «vivian rh wi 09 9 9 0 9 14 14 - — - - Black. + vcs vs wvm as manw vie 1 - 1 - - - American Indian. .......... - - wt - - - ASIEN. «cv vv ee em - - - - - - Oar, . vv vv prvi mwrrysns - - - - - - Relationship and race Decedent was death certificate informant’s — Spouse: AUTACES ov 5 vo mss 45 mwa om 6,002 4,614 1,153 148 62 25 WHS. 00 3. 000 & 20m 30 5 3s 4,661 4,595 13 27 9 17 BIABK: © us v 200s 0 v0 wis 0 ain 1,161 10 1,139 8 1 3 American Indian. . . ........ 124 7 1 113 1 2 ASIAN. «5 is swam ban a rms 53 1 - - 49 3 OBE: v4.0 5 4 0 re vo fe 5 09 009 3 1 - - 2 - Parent: Al1B0BS ow sv sm mens wmams bn 1,161 779 330 40 6 6 WIRD, 5.0 5: + oo 2 4 mt 3 9.3 5 be 801 777 7 10 1 6 BIACK. «vu rusnssmismmausnn 328 1 323 3 1 - American Indian. . . ........ 28 1 - 27 - - ASIBIY, « s iosnsemsmr svn 4 - - - 4 - OB: ns ninniminmnimime —- - - - - - Child: A TBCBS vv vows cis wane in 1,746 1,199 443 61 29 14 WHS. v3 roicwp rms simoms 1,217 1,194 2 9 2 10 BIAOK: m2 sms wa aman gris as 448 2 441 2 1 2 American Indian. . . ........ 52 2 - 50 - - BBIBIY: v5 cvs swismms ams 29 1 - - 26 2 ORY: cov mim ims vane wa - - - - - Sibling: ANYEOBS svi vu vans swiss 600 345 217 29 2 7 WIS, i on oi 5 iw 0 2 90 3 os & 8 (30 8 351 338 2 4 1 6 Blaok, .: cvcvsensanemomy 222 4 215 2 - 1 American Indian. . ......... 26 3 - 23 - - PBIB. «5 cov Hvis we 50s 9nd 1 - - - 1 - OIRBY, wo vmvo meme mmemsinn - - - - - - 26 Table 2. Number of responses by race of decedent on National Mortality Followback Survey questionnaire and by race on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Race on questionnaire Race on death certificate, interval American and relationship Total White Black Indian Asian Other Other relative: ANTACES vib 4 bie #0 6 400 45s 471 282 170 14 3 2 WARE. 12. i te ison ow oc ows me od 284 280 - 1 1 2 BECK. +4 v5 swiv a vmw no boas 173 1 169 3 - - American INQIAIY «vv tw 000 0 0 i 11 1 - 10 - - ASIAN, vi hnn sn vw gE 8 3 — 1 - 2 - OBE. + ov vn ds seme mr eam - - - - - - Nonrelative: AUYEOES i ooo von 4 5 mien wn ie 384 269 90 18 4 3 WAG, +. ot 0 oi oe wo coe 0s ve 274 266 - 4 2 2 BIaGK, + sw s a wan sue sn sn ns 91 1 90 - - - American Indian. . . ........ 15 2 - 13 - - ASIAN. + o 5 «0 0 win iw mow ol Bie 4 - - 1 2 1 Other. cc cosvssmvns nas - - - - - - Not stated: ANTACRE «. oui 53 vv 7 5 wb bin at vw 01 3 5,608 3,679 1,657 204 41 27 WHRE.. 7 & sss vias #5 0 08 3,727 3,659 12 31 7 18 BIaoK, . sc vcsnmv mint anes 1,678 9 1,644 16 2 7 American Indian. . ......... 166 8 1 156 - 1 PRIBY. wn nim 5rd vw md 8 eb 8 37 3 - 3 32 1 LT EE SEE Jr - - - - - - Death certificate informant and survey respondent were — Both spouse: AHYBOBE . sv v vs wn ewnmnsinss 5,075 3,951 943 114 50 17 WHS. ov a0 5 0 0 5 5 93 0 oe 3,988 3,936 9 23 8 12 BACK: + ios i voi 5 8 4 5s Bi 3 950 7 933 6 1 3 American Indian. . ......... 94 7 1 85 1 - Asian. ................. 43 1 - - 40 2 OINBE: 000 0 4 wi 0 3h 900 06 #4 - - - - - - Not both spouse: AILTBBOS ov viv thin ves 5 mm v3 10,346 6,837 2,972 383 89 65 WHS. ov vvm vs mndm es wan ees 6,947 6,798 26 61 13 49 Black: « «vis ws nmans nega 4 2,999 18 2,944 26 3 8 American Indian. . . ........ 315 17 1 294 - 3 ASIN. . vis iranian srry 82 3 1 2 71 5 OMBr, . sscvivsresvronay 3 1 - - 2 —- Not stated: ANTBOBEG +. 05 5. 500 00 0 900 00 Foie 551 379 145 17 8 2 WARS: vw vss vm sms inva smn 380 375 1 2 2 - Black: . vv vermnnwre esos 152 3 144 2 1 2 American Indian. . . ........ 13 —- - 13 - - BBIBN.. 5 5 vw vis bet Toman Ee Bom pow 6 1 - - 5 - OIABE. i «mon vim 5d iw 3 - - - - —- - NOTES: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. Multiple entries of race on the questionnaire were classified in the following way to correspond with the coding of similar entries on the death certificate: American Indian and white (two cases) were classified as American Indian. Asian and white (one case) was classified as Asian. Black and white (one case) was classified as black. There were five cases in which three or more races were given on the questionnaire; these were classified as other.” 27 Table 3. Number of responses by Hispanic origin of decedent on National Mortality Followback Survey questionnaire and by race and Hispanic origin on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Origin on questionnaire Hispanic Non- Race and Hispanic origin on death certificate, interval, and relationship Total origin Hispanic Race and Hispanic origin All races: AIONOITE + a4 vs slik ois silo wails Hh $3 8h SAREE ERE Had Hab 6,990 366 6,624 HIBPARIG: vs 4 wins # 5% wi B40 E 55 oF Bie niki Weck ok 191 Bhp oF pn HE AE 6 306 297 9 INONEIBPAMIG. +. ome wr #0 8 0 5 #972 0.5 2 5 906 80 #0 0.00 0 8 55 5.0 © 8 5% 08 0 6,684 69 6,615 White: PIEDHGIIS 1c: v v0) £1008 +1 3 505 5 50 Bk ne des 1 i of Wh 8 Heme ® 0610 0 050 oft ed) 5 6 5,129 338 4,791 HIBPANICONGIN. «wv ms swe ms rma pus HEAT EME CREMP BEDE 0 301 294 7 NOP-HUBEAINIG,, + «vs io 0500 «on 6 5 wi om oe 4 8 (0 6 a8 A 0 We Gs 4,828 44 4,784 Black: AITOIDINS + 5:5 vs wav wwe s wa 60s HF a eB rae 0s 08 EHOW Bh 8H 1,606 11 1,595 HISPOMIGIONOT. « 5s win vs ms 5 98 0% 40H BR EWE BS S009 Eb EWR S 1 1 - INOHEUSHANIC,, « 1 wv imivech 0 00 0 2k od 50 oi in ue Bim ln 0 fl #5 4 20 kB 0 ct 7 3 0 1,605 10 1,595 American indian: AUOHIING «+ sos vs vanes bss RE EW sr ERE mA SWARMS LMS 0S BE 159 10 149 PHEPANIBIONIEIN. is fh a wins 0 0 0 er 00 di BG RR Be RR 1 1 - INOAHUSDANICH + 5 viv wi vw a5 4 8 3 wb hind a ih foe Bows oh a ¥ 500 ie we 0 9 158 9 149 Other: AUOHDING. «vs v sv bm ram am ys ms Bn AME mR Em EER ERT BR SREB 96 7 89 HISPANIBIOABIN, wv 4 vo: 0 05 0 win 50 06 ov fy 4 00 Biot alow oF di 8 8 0) ways 0) 0 0 3 1 2 NONHIBPRAING. ix 5 + wis 5 2 bob arin dig 3 ie i bk wie 0 0m M0 0 mr 8 0 00 93 6 87 Interval and Hispanic origin 22-25 weeks: ALONGING 5s wv vm RT EF Ere BE Hae MES FRAG TES SWE ERAS 407 14 393 HISPANIC OMG. «5 «mis in 5 2 wa i 850 3 0R 8.6 3 Rin iho 0 R08 3 0 R05 BE HSE 11 10 1 INODY-FIBRAOICH: « vv: oi swt # uo 04 #60 80 680 co do 9 00 08) We 05 al fs 800 8 000 9000 396 4 392 26-29 weeks: AUOHOING ¢ yr vv ws eeu saamup tne smwemes we weens dss es 851 36 815 HISPANIC ONO. v2 sos mss bia Wa $2 RB LNA RF NF BEI TERE ES 30 30 - INOARIBOANIG: «ov sits in 40 Folie bio ode i 3 3) Holl 06 0 0 ch 00 9 0 HO WE 821 6 815 30-32 weeks: AIOHGING «5c prov smemprmume s ws wo BREE SHWE ¥ wwwx» ow 783 29 754 HBPANICONGIT. + sms wo twee ims NOIR RI GER R ERS 22 22 - NGIHIBIRANC.. « ov: + 50: 6 10 i510 0 4 0 5108 Hh 6 Bock i 0 hades 08 vi 0 i ht A ic 761 7 754 33-35 weeks: AIONOINS: + 5 + v5 iw 200 w 40 5 5 00% 50 4 40% 0 G04 9 R14 B13 08 000 0 wg 943 31 912 HISpAnio Ona. « : cosa snc usin ns sssme ids nebbs was vans 29 28 1 INOIHIBPBIAIG. : +. x 505. 51 # 3007 5 18 ots 90m 500 0 07 B35 oF coir ot 0 0 0 a B03 i om 914 3 911 36-38 weeks: AIOHDINS, «svi v www sms BEE ws WE ra As Es Sus % Ae ews wet 892 41 851 HIBPANUCONDMY, ov + cms mmemamp imine sms His HAW E I HIT I Basa 36 34 2 NORHHSBAMIC. « coins mais smi boa mm ams md ams n eres sess 856 7 849 39-41 weeks: ADHOINS, ov vi Wa bw aS A hs Beis SRS Ha EHR W SE AE Sh EE Re 823 51 772 FHSPANIBIRIIII. ov coi mi met doe ois wo 30k ow for coin 80 2m aml 0m 6 3 © rt 40 40 - NOBHIBDAING « 5 5 vt 5 ow ty 300 8 of 3) 8 me on Bk sos ood 3 16 ws 5 400 5 ed I 3 of ow 783 11 772 42-44 weeks: AIONIONIS 1:2 250 en vie 5 2 08 00 Bw 0 8 IK 3 00 Sh 00 00 eh REA 6 802 741 45 696 FBPATHGIOPIGI: .v 51c wv wt 20 6 5 42 1 ome 0 oF 0 od A 0 8 0 41 39 2 NORHIIEPRING sw: 515 018 don 5 51% tu 5 & le br ow 0108 0 teh 20 ow oh ev i 700 6 694 45-47 weeks: BIOHIITS! iv. 5. wre os vt wm is worm ln hs = 21 4-00) B00 100 ord Whim 0 ooh BL 8 205 Bl or 0 6 Brim 540 38 502 HISPAIGOHGUY, «or va vir in tot rtm edness mains ws 29 29 - NONHISDANIC, «vs 254 wm a 55 50 8 wi ah oe 508 9s 5 4% 0a @ 8 § wow waa 511 9 502 48-51 weeks: AILOAGIE ac oo 5 00 6 500 0 0 0 000 0 ea 0 60 60 0d E00 3 0 0 0 0 00 425 31 394 FISPARIBIONEIN. 4 + «chix oi 5s 45 ui 0 3 oh 00d S16 oF 1 00h Ek ch 0 0 ek bk hei 24 22 2 NONHIBPANIG: + x 5, 403 0 is ms vw 8 mab wR hw x 6 iw 0 NE wee 401 9 392 28 Table 3. Number of responses by Hispanic origin of decedent on National Mortality Followback Survey questionnaire and by race and Hispanic origin on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986—Con. Origin on questionnaire Hispanic Non- Race and Hispanic origin on death certificate, interval, and relationship Total origin Hispanic 52 weeks and longer: PUOTIGIE. ... vc 4.110 sion coicaians iss ent oe sacra cns oo 400 Rim 1 riot chest’ i 0 505 000 0 0 580 50 530 FIBPANIG ORGY, «+ vk 1s iv co ww ww ww iw wom ow wwin wae 9 wh 44 43 1 NOPHISPANIC, 0.4 ow 20h oH 00 S066 5 50 0 WES we 3s Bi Hie 536 7 529 Missing: PILGFIOIIE: wo. von: cova in ims smi imi cms cor wo coi, coe 30 90 Tt i # pio 5 - 5 HISPANIC. +c vss sresrsr spss sms urns nasa snvib anes vis - NOM-HISPANID. + + ss vs a VL ERT FTER EL AR TIE BEEBE MERE HEY 30 WBE 5 - Relationship and Hispanic origin Decedent was death certificate informant's — Spouse: AUOHAGING . cnr nv vr rr mrss ss cms vw rs ab ame ar Ems o8 a 2,625 129 2,496 HISPANIGIONIGIN. + viv 49 v0 ww 31% 60) 8 0 8 00 A BLE A we We ee 113 110 3 NORHISRANNIC: 3% #3 & 500 4 5 6 REE 00 BLE 0 008 G8 5 5005) 0 $8 00F 0 50 wy 0 2,512 19 2,493 Parent: ATORIBIITS, co. 110 crn svins rm v2 1 i 0 on 5 50 oni ww 50 oo 0 ee Gd 200 0 ot 0 #40 hed) 0 542 39 503 HISPANICIONGIN: ¢ vs vs ss vas swomn rms ms sas ma wa swo mes ohio 31 31 - NON-HISPANIC: + + sv wv vay we em ams dm e@ sep sms s Mensa Frwanmoe 511 8 503 Child: AIIOHBIE. o: 0: 0 1 15: wx 0955 100 is we 105 om sw 01 500 0 0 0 0 Dn 0 0 0 9000 0 9 1,087 58 1,029 HBPARIBIONGIA. vr or 1s wre wms ame ms mmemsmme Be mms PRA ERs E20 45 45 - NOFHIBPANNIC, ¢ iv wis tis tEs MB EES Res BBR EMS WB rH MEGS gS 1,042 13 1,029 Sibling: AVOTBINE 4: vx macs mum rm dm ms msm AE ap fH AMEE EA RES 333 32 301 HISPANIGONGIN,: = + cuss emsms ra smpe mms ems mums ams mass 28 27 1 INORVEUSORIMIC & 55.0 6 (5% 10 3 100 105 8 080 000 90 5 (600 6 (00% 0 81% foi 6 (90% 05 & 10d wi 3 0 305 5 300 Other relative: AOTIOING 4 sv 5 ws wmv orn om 6m mw we ow 0 30 0% ww 0 ws fo % fw ee & 0% we 293 11 282 HISPANIC ORIGIN: co: 2 iv 5 63.00 5: 8 5.0 50 4 0 2.5 81 7 8 3108 81 3 [#1 90k 0) 18 B08 6) % 10 iw 9 9 - INONHUSDBING, & +. oo 0: « 56 50 8 500 5d © 00 0 00h 5 Biome 00 300 0 00 ih 1 ot 284 2 282 Nonrelative: PIONS ov nviws ws ows 00 ww a5 a 0m www we sie vaso ws we wie ewe 271 21 250 HISPANIC ORIGIN iv 5.2 vr w 3s #180 0 0 sik at 8 060 0 5080 0 08 0 2 9000 i 60 #0 18 17 1 INDEEIBOANNC. 0. + co 000m 4 car tim; 4500000000 0 omarn (00 030) rw i0 ks 00 3 10 0 G0 0 253 4 249 Not stated: AVOHOING « vv um sry no swsmy s wows s was £m ess Svs dwvnss os 1,839 76 1,763 FISPEINE OO: =: 5 0ri0it 6: 5: omits Wi 1b sich 30 eh Wie) 8h Ta 5 0 Ld) 0 iy AME 9b 8 62 58 4 INGITFISTBIIC. ox 0:10.50: wives atm. caren mens ences senses co ess 28s 0 8 00 i iam ARR 0 ft 1,777 18 1,759 Death certificate informant and survey respondent were — Both spouse: BIGTIBIAE. ..co110x. +2oxccrssaim senor oes wn hoon oe vires oe gps rose om i: oh i 8 pr 0p 2,199 95 2,104 HISPANIC ONO: v5 3 vs 95 4 ETRE LET 0a FA PEE LEO WR 10 GE BEE 85 83 2 INOIISDRNGC, «5: iv 1000 00000 0 00500 0 te) 00 30 a 0 ie 90 eine wo 0 2,114 12 2,102 Not both spouse: ABIES: 0: 1010 coc iow cor warn. sos sos win wii n wie 8 0 0 OER BRIE Ae BBR 4,554 253 4,301 HBPANIS OHGING + vv wiv www 00 ww 000 0h 80 0 0008 0 oe Te 208 201 7 INORFISRANIC, & oem & 00% 3 RE RS Bo 8 0 00 8 38 i SLE ie 4,346 52 4,294 Not stated: ANOAGING + vv mon ma vn mms meas wer sg pees dems svbns ve 237 18 219 HISPONICTONONY: us 0 wv 3 0200 9 0% 5 8 80% 6 80 Tard kn 13 13 - NOEISPRINNC: 55 ww sais © Ws ME B04 MEME § WTAE £0 S000 480800 Bo 224 5 219 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 29 Table 4. Number of responses by marital status of decedent on National Mortality Followback Survey questionnaire and by race and marital status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Marital status on questionnaire Race and marital status on death certificate, Never interval, and relationship Total Married Widowed Divorced married Race and marital status All races: All marital statuses. . . ............... 15,865 7,776 4,522 1,689 1,878 MEG : v5 saisas amis smams aime ms 7,692 7,470 29 68 25 WILOWED 0. cvs 0b suis wb smiwn deme 4,753 125 4,426 163 39 Divorced. . ............ 1,600 124 52 1,393 31 Novermamied ... os cvs iws suv wn enews 1,920 57 15 65 1,783 White: All marital statuses. . . ............... 11,280 5,736 3,139 1,196 1,209 IMBIBE & i ve wre: soni we 0 + 0 ow wie 8 5,660 5,602 17 31 10 WIHOWED ios 0s ot os 55 10 rv rs io wots ale ie 3,229 50 3,088 80 1 DIVOICRY. . srsscnimbssmemb amas nn 1,144 62 28 1,038 16 Nevermarried . . ................. 1,247 22 6 47 1,172 Black: All marital SIAUSES. wu cas sr ss ns swe wa 4,040 1,780 1,229 427 604 MATHBE 1 3 oi5 ob 0 als ha of 505 905 Baie & hg 1,683 1,623 12 34 14 WHOOWBE © ov vv wav nin iw ws ww 0 dis iw 5 0 1,359 72 1,186 74 27 DIVOIced. vcs swnsms evn ms wes mses 394 57 22 302 13 NEVE MBIA «rar os ms wane ss 604 28 9 17 550 American Indian: Almansa) StalUSES. « v..ms wu 5 ws wwe we ns 411 176 126 56 53 MarriBd « snes sas ss sas as sar aE 167 164 - 3 - WILOWEE z5 cvs 255 news oh Sah £2 40 134 Y 124 8 1 DIVOICBU. vss nv mvs an tno nswn avs wa 53 5 2 44 2 Nevermaried . : cs cvs snes nvnsnres 57 6 - 3 50 Other: All marital statuses. . . ............... 134 84 28 10 12 MBEIBY x vo: wii 5: wimiwie wind vo wy iy 82 81 - — 1 Widowed . cu vssassrnas seems wens 31 2 28 1 - DIVOICBT. « 05 2 Bisa 6 G00 0E A045 EEE 9 - — 9 as NBVBFIMAMIBG vv vv vv vx vw mwmy wir 12 1 - - 1 Interval and marital status 22-25 weeks: All marital statuses. . . ............... 1,086 544 323 102 117 MBIT. oc viv ww vin in on 2s bw was vay 537 532 2 3 - WIdOWed uo x vain s vanes wove ae on 338 3 318 16 1 PIVOICEE,. oo 0 7 2408 shame paket vases 89 6 3 79 1 Nevermarried . . ................. 122 3 - 4 115 26-29 weeks: Allmarital statuses. . . . : cover varias a 2,235 1,182 659 214 230 MESH! 5 0 0. 5 18 5 100 4 Win ioe 5 00 F000 Bp 00 00 1,116 1,109 2 3 2 WIGOWBY + 40 onan 5 ww 00 ws 0 pie 682 7 652 20 3 DVOICBD. «cx x wmv ss winan vv wa mae ss 205 10 4 189 2 Nevermarfied : : » ss vwa ims onvns sas 232 6 i 2 223 30-32 weeks: All marital statuses. . . ............... 1,937 988 553 190 206 Marfied sv ssrvesn srs amrs ns wines 959 950 2 6 1 WIDOWED ovo vm sn vivo wie via 5 0 id 589 21 547 18 3 DIVOICRE, oo sth ha Ab BB a i # 180 13 4 163 - Nevermarried . . ................. 209 4 - 3 202 33-35 weeks: Al marital Sialuses. ; . ccs ne vermis ns 2,341 1,169 677 21 284 Married » curs imsa vans aman mes 1,144 1,136 3 4 3 Widowed . ..................... 708 16 667 20 5 DIVORCE. » . vx corms sv vmmmmyiorm ws 203 1 6 182 4 Novermaried . .. cos cnrws vo saneus 286 6 1 5 274 30 Table 4. Number of responses by marital status of decedent on National Mortality Followback Survey questionnaire and by race and marital status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Marital status on questionnaire Race and marital status on death certificate, Never interval, and relationship Total Married Widowed Divorced married 36-38 weeks: All marital statuses. . . ............... 2,043 1,013 560 232 238 Mamet . vies vps ninviss ens ms@y 990 969 8 8 5 WIOWB 2s. 555 + 6% ais Hoes in 3 bed oinon 593 20 546 22 5 Divorced. . . .......coiiiii 213 17 4 190 2 Nevermarried . . ................. 247 7 2 12 226 39-41 weeks: All marital Statuses. «. ws ws svn vn sms ins 1,819 854 518 222 225 Married . ............ 837 814 4 13 6 Widowed . .......... oo... 542 18 501 14 9 DVOread:. .» cvs wn swiss smsms ames 203 11 9 181 2 Nevermarried . . ................. 237 1 4 14 208 42-44 weeks: Ailmarital statuses. . . . ........: 0:04 1,605 749 473 193 190 Married ..: corns snows anim Eman 728 706 2 13 7 Widowed . ..................... 505 19 461 21 4 DIVOrced: ». s vs ses sms manmews snzms 186 18 7 154 7 Nevermarried . . ................. 186 6 3 5 172 45-47 weeks: All marital statuses. . . ............... 1,152 553 309 144 146 Married , .:vvsnivrinsns inser as 534 518 3 11 2 WILOWE sno so mvmn dsm smimm nme 323 12 298 1 2 Divorced. . ........ viii 140 16 7 113 4 Nevermarmied cv. vs swims ewsmsves 155 7 1 9 138 48-51 weeks: All marital statuses. . ................ 833 399 226 95 113 Matfed «+. us ansesns sma nv me wy 377 371 2 3 1 Widowed . susvssnsses ns 2m ams Ess 243 8 221 9 5 Divorced. . . ........ oii. 96 16 2 76 2 Nevermarried . . ................. 17 4 1 7 105 52 weeks and longer: All marital statuses. . . . .............. 799 366 219 85 129 Married . ........... 0... 361 356 1 4 - WIBOWBE. «5: ci00 5 wow + it win 2 0.00 west 4002 3150 225 1 210 12 2 DIVOTCE. + + vv toe eee eee 84 6 6 65 7 Nevermamed. . . vw: ss snicn swiss ans 129 3 2 4 120 Missing: ALA) SIBUSES .. vv wiv wv sin nw 15 9 5 1 - Married: ss sos sms m ss mama ssn 9 9 - - - Widowed . ..................... 5 - 5 - - Divorced. . . ....... iii 1 - - 9 - Relationship and marital status Decedent was death certificate informant’s — Spouse: Al marital statuses. . + 4 +. sve vw suis ua 5 6,007 5,939 26 29 13 Marmied : cv imw ens vn sna REL EE ww 5,964 5,930 8 22 4 Widowed . ..................... 23 5 18 - - Divorced. . . ......... 8 1 - 7 - Nevermamied . . c cron vos vw ams sw sm 12 3 - - 9 Parent: All marital statuses. . . ............... 1,146 185 51 307 603 Married ... covvvvvnnssnsms sr in 147 138 1 4 4 Widowed sworn snanm iv res snes 57 8 46 3 - Divorced. . . .......oiiiiii 323 24 3 284 12 Nevermarried . . ................. 619 15 1 16 587 31 Table 4. Number of responses by marital status of decedent on National Mortality Followback Survey questionnaire and by race and marital status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Marital status on questionnaire Race and marital status on death certificate, Never interval, and relationship Total Married Widowed Divorced married Child: Almarital statuses. . . .....cceevvvr es 1,755 289 1,168 260 38 Married , uc vn ns nese rvemy £ows 252 241 TZ 3 1 WIHOWEA cvs crsvasnananmenines 1,198 24 1,145 24 5 DINGIEBA, «v4 2.00 2: aon hin imion aor mdion 267 23 16 226 2 Nevermarried . . ................. 38 1 - 7 30 Sibling: All marital statuses. . . ............... 594 81 126 129 258 MBIAOH ovr v. co: +: 01000110 tomas aronss haseion corsets 67 60 - 4 3 WIdOWad os sav wns srs ns vue 143 10 123 7 3 DIVOroBE. + vv + ww vie wim s wie #08 00 8 8 Ew 120 5 1 112 2 INEBYBL IBIBO « i..v 5 5 vivo ow iih #85 ik 264 6 2 6 250 Other relative: All marital S1alUSeS. vv 4 «ws wn vse ws v8 469 63 249 52 105 MRA... ox ini a0 000 8 Basi BE SEE 49 46 - 3 - WICOWEH oo. 0 006 0 000) 0 hs hon) 5 co riiia 261 10 240 9 2 OIVOPBBA: «4:0 v.01 500 wearin mi 0 10 061m 158000 sn 42 2 5 35 - Nevermarried +c vo « cova unm r wens 117 5 4 5 103 Nonrelative: All marital statuses. . ................ 366 83 115 78 90 Mamied. ov vo vmams vie we wpewy sm 73 68 - 4 1 WIDOWED va ove a5 wns 45 DHE 95 D0 130 4 112 9 5 DIVOICBU. ss srs ns samme visins snes 70 7 3 59 1 Nevermarried . . ................. 93 4 — 6 83 Not stated: Al marital statuses. . . . cov vi vn ss nr ns 5,528 1,136 2,787 834 771 NMIBIHBA .o. 5.0 scsiccti ar dnicnninis whinessinioes damien 1,040 987 13 28 12 Widowed . ......... thal ER et Ei va 0 2,941 64 2,742 111 24 Divorced. « sv sins sw WE ER 770 62 24 670 14 Never married . . ....... LR EA 777 23 8 25 721 Death certificate informant and survey respondent were — Both spouse: All marital statuses. . . ............... 5,083 5,069 3 1 - METAB. + + 2.5 Biron Bite us 21 0 0 wie Sugino 5,073 5,060 3 10 - WIdOWed , ss cus vs prs sme s £5 5 5 - - - DIVISBd: + su ans an spas mmm s ns vines 2 1 - 1 - Nevermarried . . ......... cov 3 3 - - - Not both spouse: Al marital Staluses. . . sev s vrs serra 10,238 2,217 4,495 1,655 1,871 MEITIBL. ov vc 2 00 WR Bae eb @ RE % 8 2,042 1,935 26 56 25 WIGOWEH +o vom rans son sesmanms 4,721 117 4,402 163 39 DIVOIGBA. « «5 vv nv uw mms wm mms om ww uy 1,568 114 52 1,371 31 Nevermarfied . co vu von suse sn wey 1,907 51 15 65 1,776 Not stated: All marital statuses. . . . .............. 544 490 24 23 7 MABE ro vsmmom mamas mas enn swns 477 475 - 2 - Widowed + ss sine 19 90m v0 0% 0 F DEE 8 27 3 24 - - DIVOICRG. i 2.5 4 415 0d kn 40 56:5 509 @ 30 9 - 21 - Nevermarried . . . ................ 10 3 - - 7 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 32 Table 5. Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Occupation on questionnaire Operators, Managerial Technical, Production, fabricators, Race and occupation on death and sales, and craft, and Armed certificate, interval, and relationship Total professional administrative Service Farming and repair laborers Forces Race and occupation All races: All ocoupations . « +, «cvs ven 2,665 342 454 398 264 404 738 65 Managerial and professional . . . 467 269 91 35 9 29 26 8 Technical, sales, and administrative. . . ......... 428 37 304 22 8 20 33 4 SOIVICO + sv ns smims rms 393 12 13 298 15 11 38 6 Farming. «esis basins swismn 227 4 1 5 186 9 21 1 Production, craft, and repair . . . 384 7 10 8 9 267 76 7 Operators, fabricators, and BOOSIE ovis wie vim s onn nmi g wi 730 12 29 30 37 66 542 14 Armed Forces . . .......... 36 1 6 - - 2 2 25 White: AN OCOUPBHIONS «+ + ss 0 v0 50s is 1,923 282 388 192 172 340 499 50 Managerial and professional . . . 394 226 82 22 8 28 21 7 Technical, sales, and administraive . . . ......... 372 31 268 17 8 17 27 4 SOIVIOB. cs swiss nwa inn 189 7 8 131 5 9 25 4 Farming... cc canvasmsons 154 2 - 4 127 9 12 Production, craft, and repair . . . 325 7 10 6 8 228 60 6 Operators, fabricators, and 1aDOISIS os wns ome wsmn 459 8 16 12 16 48 352 7 Armed Forces . ........... 30 1 - - - 1 2 22 Black: AN oooUPEIONS . « vs wx eww 670 51 59 198 79 54 215 14 Managerial and professional . . . 62 38 6 12 - 1 4 1 Technical, sales, and administrative. . . . ........ 49 4 34 5 - 1 5 - BOVICE ovis uinuw sims swans» 195 5 5 161 8 2 12 2 Fang « os wv xwo www ews 64 2 1 1 51 - 8 1 Production, craft, and repair . . . 48 - - 1 1 33 12 1 Operators, fabricators, and ABOISIS oes sme s mse 247 2 11 18 19 16 174 7 ArMBO FOICBS +» vs ws vw von 5 - 2 - - 1 - 2 American Indian: All occupations . . . .......... 64 8 6 6 11 10 23 - Managerial and professional . . . 8 5 2 - - - 1 - Technical, sales, and administrative. . . ......... 7 2 2 - - 2 1 - SOIVIBE. i. +c 5 + 0 me 0 Homie Be 7 - - 5 1 - 1 - Famiing. co covomuvmama i 9 - - - 8 - 1 - Production, craft, and repair . . . 10 - - 1 - 6 3 - Operators, fabricators, and ODO + vo vi 2.70 wi 5 in 4 wine 23 1 2 - 2 2 16 Armed Forces .. .......... - - - - - - - - Other: AlocoupationS. . « . ..cocuun 8 1 1 2 2 - 1 1 Managerial and professional . . . 3 - 1 1 1 - - - Technical, sales, and administrative. . . . ........ - - —- - - fo I” - SOVICO ow. swv vivmen sms 2 - - 4 1 - - - FRITING s. 5.00: 5.50% v.00 ie 0 0.00 58 8 - — - - - - - - Production, craft, and repair . . . 1 - - - - - = Operators, fabricators, and 10088 vs meas Ey sm vs 1 1 - - - - - - AmedForces ............ 1 - - - - - - 1 33 Table 5. Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Occupation on questionnaire Operators, Managerial Technical, Production, fabricators, Race and occupation on death and sales, and craft, and Armed certificate, interval, and relationship Total professional administrative Service Farming and repair laborers Forces Interval and occupation 22-25 weeks: AllococupationS. . . +. vv vv 0 cn 214 32 37 26 27 29 56 7 Managerial and professional . . . 45 25 10 4 - 2 3 1 Technical, sales, and administrative. . . ......... 32 5 21 1 1 1 3 - SOIC: «a wa wv 4 0% 5 Was wad 23 1 - 17 1 1 2 1 Faming = . vs se vas sede 24 - - - 21 1 2 - Production, craft, and repair . . . 33 - 2 1 1 22 6 1 Operators, fabricators, and 18DOrers « « «cans vas ranss 53 5 3 3 3 2 40 1 Armed Forces . ........... 4 - 1 - - - - 3 26-29 weeks: Allocoupations., . « « «vx semis» 469 60 82 65 41 66 141 14 Managerial and professional . . . 77 44 13 4 2 8 3 3 Technical, sales, and administrative. . . ......... 77 8 59 2 2 4 2 - SEIVICE +. vs wus wow oi 2 matin s 71 2 3 54 1 4 5 2 Faming. cess vwnsarnsns 38 2 3 29 - 4 - Production, craft, and repair . . . 67 1 2 1 2 41 20 - Operators, fabricators, and YBBOFBIS + ivi s:is 0 4 9i le) mornin rie 132 3 4 1 5 9 107 3 Armed Forces ....cv vac vss 7 - 1 - - - - 6 30-32 weeks: All occupations. . . .......... 349 47 55 46 41 60 91 9 Managerial and professional . . . 74 42 12 8 1 6 4 1 Technical, sales, and administrative. . . ......... 54 3 33 7 2 3 6 - SEWICE: 116 403 Ak wma death o 43 2 2 28 1 1 8 1 FRING oo vivo 00 0 fb ¢ friitit os 32 - - - 29 1 2 - Production, craft, and repair . . . 53 — 2 1 - 38 12 - Operators, fabricators, and 1BDORSIS + suis vos sade 88 - 5 2 8 11 59 3 AMNBO FOICES «cov vv van mvs 5 - 3 - - - - 4 33-35 weeks: Alloccupations. . . .......... 426 65 88 56 36 63 113 5 Managerial and professional . . . 86 52 19 7 2 3 2 1 Technical, sales, and administrative. . .......... 82 6 62 1 1 3 9 - SOMVICB vs as unm rE 61 5 3 43 3 - 7 - FaIMING 5 4 4105 51% wi 5m Slain 5 29 - - 1 23 3 2 - Production, craft, and repair . . . 63 - - 2 - 50 10 1 Operators, fabricators, and [aDOBIB «5 2 ssw sme 2 ans R 103 2 4 2 7 4 83 1 AIBA FOIOBS +o + 4 v0 4 wwe io » 2 - - - - - - 2 36-38 weeks: A OCoUPBLIONS . , vv + vu wan ves 368 40 70 57 39 53 96 13 Managerial and professional . . . 54 28 14 3 1 3 3 2 Technical, sales, and administrative. . . ......... 60 4 49 2 - 1 2 2 BEIVICE: 10.4: 4m wlln 28% 3% 410 4 » 52 1 1 44 9 4 1 Fang. suis sn sme am ims s 38 3 - - 28 3 5 1 Production, craft, and repair . . . B1 3 2 - 3 34 7 2 Operators, fabricators, and 1BOTSIS oc sv ns ss sms sos 110 3 4 8 6 12 74 3 AMmEd FOICes . . .....omvsu. 3 - - - - - 1 2 34 Table 5. Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Occupation on questionnaire Operators, Managerial Technical, Production, fabricators, Race and occupation on death and sales, and craft, and Armed certificate, interval, and relationship Total professional administrative Service Farming and repair laborers Forces 39-41 weeks: All occupations . . . .......... 276 31 46 47 29 45 72 Managerial and professional . . . 43 23 10 3 2 1 4 Technical sales, and administrative. . . . ........ 41 6 28 3 1 2 1 Service ................ 54 - 2 39 2 3 7 Farming... vox vemaw ews 21 1 - - 20 - - Production, craft, and repair . . . 38 - 2 - - 25 1 Operators, fabricators, and laborers . .............. 72 - 3 2 4 12 49 AME FOICBS «vv vv sna warns 7 1 1 - - 2 - 42-44 weeks: All oooupatIONS . « « «ws mw ving 242 26 38 44 23 40 69 Managerial and professional . . . 31 22 3 3 - 2 4 Technical, sales, and administrative. . . . . ..... 43 2 30 4 1 3 3 SeIViCS : css mv rms sas 37 - - 30 3 1 3 FANNING; os cwsms swams uns 21 — 1 1 17 - 2 Production, craft, and repair . . . 31 - 1 1 27 9 Operators, fabricators and 1BDOrSIS «sv we ws wm wr mus 76 1 3 5 1 7 59 Armed Forces ........:«:: 3 - 1 - - - - 45-47 weeks: AlLOCOUPRLIONS + «vs x 4:5 5:5 0 1 + 167 21 20 23 16 34 50 Managerial and professional . . . 28 15 5 - 1 3 4 Technical, sales, and administrative. . . ......... 21 2 12 1 - 1 4 Service ................ 28 1 1 21 2 1 2 Farming. . cvs wnswaevsmen 12 - - - 1 - 1 Production, craft, and repair . . . 29 1 - - - 23 4 Operators, fabricators, and laborers . .............. 47 2 1 1 2 6 35 Armed FOIoeS . « «vos vases 2 - 1 - — —- - 48-51 weeks: All occupations. . . .......... 110 12 14 27 9 7 37 Managerial and professional . . . 22 12 5 3 - - 2 Technical, sales, and administrative. . . ......... 13 - 7 1 - 2 2 Service ........ i. 18 - - 17 1 —- - Faming. «5 sue saanm seme 6 - - - 6 - - Production, craft, and repair . . . 10 - - 2 1 3 3 Operators, fabricators, and JABOIBIS. , .« «iris ov von wie 39 - 2 4 Y 2 30 Armed FOrees «uv vv ssw vo sus 2 - - - - - - 52 weeks and longer: All occupations. . . .......... 40 7 4 7 3 6 11 Managerial and professional . . . 5 5 = — - - - Technical, sales, and administrative. . . ......... 5 1 3 - - - 1 Service ................ 6 - 1 5 - - - Farming. «cs «vows mvwwss 6 - - - 2 1 3 Production, craft, and repair . . . 8 - - 1 4 1 Operators, fabricators, and laborers . .............. 9 - - 2 - 1 8 Armed Forces . . .......... 1 — - - - - 1 Missing: Alloccupalions:. . « + «a cs ws ss 4 1 - - - 1 2 Managerial and professional . . . 2 1 - - - 1 - Technical, sales, and administrative. . . ......... 35 Table 5. Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Occupation on questionnaire Operators, Managerial Technical, Production, fabricators, Race and occupation on death and sales, and craft, and Armed certificate, interval, and relationship Total professional administrative Service Farming and repair laborers Forces SBIVICE a 5 wom v5 sree aes - - — - - - - Farming: sess csvwry raw - - - - - - - - Production, craft, and repair . . . 1 -“ - - nr — y —- Operators, fabricators, and IabOrers «sss vss vw sms wns 1 - - - - - 3 - ABO FOICES . : wins swans # - - - —- - - - - Relationship and occupation Decedent was death certificate informant’s — Spouse: All occupations. . . .......... 1,222 173 212 119 108 220 356 34 Managerial and professional . . . 228 132 49 1 3 18 12 3 Technical, sales, and administrative. . . ......... 200 19 134 6 6 11 21 3 Service . LL... 127 8 7 88 2 4 16 2 FaImMiNg . « « « «ces cmomws ans 94 2 1 2 78 4 7 - Production, craft, and repair . . . 219 4 6 4 4 149 48 4 Operators, fabricators, and laborers . .............. 334 8 14 8 15 33 252 4 ArMEAFOICES « ws vv ws wm 20 - 3 - - 1 - 18 Parent: All occupations. . . .......... 212 18 31 30 12 34 76 1 Managerial and professional . . . 30 13 6 2 1 4 2 2 Technical, sales, and administrative. . . ......... 27 4 18 2 - - 2 1 SBIVICE . cnc iiinsums 29 - 2 20 - 1 3 3 Farming. ............... 13 - - 1 9 - 3 - Production, craft, and repair . . . 32 - 1 1 1 22 6 1 Operators, fabricators, and IBBOTBIS + oo co monn ms Sas as 79 1 2 4 1 7 60 4 Armed Forces . . .......... 2 - 2 - - - - - Child: Aloooupalions » » «va vv ves ens 171 21 37 36 23 16 36 2 Managerial and professional . . . 28 16 6 4 - 1 1 - Technical, sales, and adminisirative.. . « v. «ov nw wt 32 2 24 4 - - 2 - Service iu vein vse as 36 1 2 25 . 4 2 2 - Faming. uns wren tvess en 21 - - - 18 1 2 - Production, craft, and repair . . . 17 1 - 4 1 8 6 - Operators, fabricators, and IBDOIBIS «v5 swvvs vw sun se 33 1 2 2 - 4 23 1 Armed FOrces . .: .cvovo unas 4 - 3 - - - - 1 Sibling: A ocCUPatIONS y+ + «+ wx sv vw ss 78 7 6 18 6 14 27 - Managerial and professional . . . 8 5 - + - 1 1 - Technical, sales, and administrative. . . ......... 8 1 5 1 - - 1 - BeIVIEB «ums nn sme mn vmss 19 1 - 15 1 - 2 - Faming: + cx snc mrmr ssw 7 - - — 3 - 4 - Production, craft, and repair . . . 12 - - 1 - 10 1 - Operators, fabricators, and 1BDOTBIS « «wv ssw simu row wns 24 - 3 - 2 3 18 - AIMBO FOICeS + «i vn vis ws + - - - - - - - —- Other relative: A OCCUPBIONS... .. + x 5 vn: 4 wi 4 1m we 53 7 11 10 12 2 10 1 Managerial and professional . . . 9 6 2 - 1 - - - Technical, sales, and administrative. . .......... 9 1 8 - — - - - Service ................ 9 - - 9 - - - - Faming.. . «cove nmewnmms 10 - - - 9 - 1 - 36 Table 5. Number of responses by occupation of decedent on National Mortality Followback Survey questionnaire and by race and occupation on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Occupation on questionnaire Operators, Managerial Technical, Production, fabricators, Race and occupation on death and sales, and craft, and Armed certificate, interval, and relationship Total professional administrative Service Farming and repair laborers Forces Production, craft, and repair . . . 3 - - - - 2 1 - Operators, fabricators, and 1aBONIS vs sows @ 20s 2m 12 - 1 1 2 - 8 - AIMBA FOICRS «vv 5c v.05 460s 1 - - - - - - 1 Nonrelative: All occupations. . . .......... 29 4 7 7 1 2 7 1 Managerial and professional . . . 8 2 3 2 1 - - - Technical, sales, and administrative. . . ......... 5 1 4 - - - - —- oT 4 - - 4 - - - —- FRING « «i 46 5.84 wi © ai sok - - — - - - - - Production, craft, and repair . . . 2 1 - - - 1 - —- Operators, fabricators, and 180OrerS «5 4 5 sin wv wm © 9 - - 1 - 1 7 2 Armed Forces . . .......... 1 - - - - - - 1 Not stated: Alloccupations . « « vss + ws wns s 900 112 150 178 102 116 226 16 Managerial and professional . . . 156 95 25 15 3 5 10 3 Technical, sales, and administrative. . . ......... 147 9 11 9 2 9 7 - SSIVIEE. » vos ux wavs vase 169 2 2 137 8 4 15 9 Farming « «ss ms wa wm 2 ewes 82 2 - 2 69 4 4 1 Production, craft, and repair . . . 99 1 3 1 3 75 14 2 Operators, fabricators, and laborers ............... 239 2 9 14 17 18 174 5 Armed FOrees: , « +x vu + wiv 5 5 8 1 - - —- 9 2 4 Death certificate informant and survey respondent were — Both spouse: Allocoupations . . . .:c:ces5s 1,058 159 185 94 93 183 313 31 Managerial and professional . . . 208 121 46 9 2 15 12 3 Technical, sales, and administrative. . . ......... 169 17 113 5 6 7 18 3 SOIVIBS: 4 vos vs wes mw 80 0 0 103 8 7 69 1 4 12 2 Farming. o + ws wwe ws wstvmis 80 2 1 2 67 3 5 - Production, craft, and repair . . . 190 4 5 3 4 128 44 2 Operators, fabricators, and laborers ............... 289 7 12 6 13 25 222 4 Ame FOICes . .:.q.: 0:04: 19 - 1 - - 1 - 17 Not both spouse: All occupations. . . .......... 1,637 175 260 298 163 202 407 32 Managerial and professional . . . 249 140 44 25 7 14 14 8 Technical, sales, and administrative. . . ......... 253 20 186 16 2 13 15 1 OBIVICE o.oo vs snsms smsms na 283 4 5 225 14 6 25 4 Farming. ............... 141 2 - 3 115 6 14 1 Production, craft, and repair . . . 173 3 4 5 4 124 29 4 Operators, fabricators, and 1aDOIBYS ; = + wi sis im x wisn oe i 423 5 16 24 21 39 309 9 Armed Forces ............ 15 1 5 - - - 1 8 Not stated: All oCCUPRIONS + «4 + nis «vis ws vs 70 8 9 6 8 19 18 2 Managerial and professional . . . 10 8 1 1 - - - - Technical, sales, and administrative. . . ......... 6 - 5 1 - - - ~ BOIVICB . «svn emammsnmems 7 - 1 4 - 1 1 - Farming «vv vv yw wn ames 6 - - - 4 - 2 - Production, craft, and repair . . . 21 - 1 - 1 15 3 1 Operators, fabricators, and 1BDOIGIS uc vs snus sms mn 18 — 1 - 3 2 11 1 Armed Forces . . .......... 2 - - - - 1 1 - NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 37 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces Race and industry All races: Alingustries . :.:: csv v0ss 2,602 233 51 254 637 234 349 81 573 124 66 Agriculture . ..... LL... 218 171 - 7 22 3 5 - 9 - 1 MII «0 5.0 ove iim ent ote iw 44 2 35 1 2 1 - - 3 - - Construction » + + viv a sv 5 245 9 1 182 21 1 4 - 6 5 6 Manufacturing . . «cv 00 200 4 648 13 8 16 499 15 44 4 32 7 10 Transportation, communication, ang UIIBS. + «+» wwii wine 234 5 1 7 15 176 10 1 8 4 7 Trade cucu isu svnen sens 329 9 3 12 32 8 229 2 27 3 4 Finance, insurance, and real BELA, «rinse i Titi 500 Bo 87 - - 2 3 2 7 64 4 4 1 SEIVICES.. + + wins 0 iti win 629 19 1 15 37 12 44 8 472 15 6 Public administration . . . . . .. 130 5 2 8 6 5 6 2 10 81 5 AMEd FOICES i... + av. ov v wisn 38 - - 4 - 1 - - 2 5 26 White: Al INQUSIIBS. + + cvs ww wi snare 1,893 158 46 191 491 176 276 68 346 90 51 AQHOMIUIG . : +5 355 swans 153 118 - 6 17 2 3 - 7 - - MINING. = oi scsi: 210 5) B08 58 3 4 41 2 32 1 2 9 - - 3 - - Construction . . .......... 169 4 - 133 14 6 3 - 3 1 5 Manufacturing . « « vo ve vues 502 10 7 10 387 13 36 4 24 4 7 Transportation, communication, andutiities . . ..c.cvv vu 178 4 1 3 11 136 8 - 6 4 5 Trade ................ 264 8 3 12 25 5 186 2 19 2 2 Finance, insurance, and real A 70 - - 2 2 1 5 54 3 2 1 SBIVICBS: » ns sas cs smiens 387 9 1 12 29 8 31 7 275 10 5 Public administration. . . . ... 97 3 2 8 4 3 4 1 5 63 4 Armed Forces . . ......... 32 - - 4 - 1 - - 1 4 22 Black: Allindustries . . ......«s 5+ 638 63 3 55 129 54 69 11 214 26 14 AQHOURUIE . ... v6 tn vimamirs 57 46 - 1 5 - 2 - 2 - MIDING , ws uv smsns smn ums 2 - 2 - - - - - - Construction » . «wv ses vos 64 5 - 41 6 5 1 - 3 2 1 Manufacturing . . . ........ 130 2 1 6 99 2 8 - 7 2 3 Transportation, communication, and utilities . cov vu cus. 53 1 - 4 3 38 2 1 2 - 2 Trade iusnvinssnsnn ins 61 1 - - 7 3 40 - 8 - 2 Finance, insurance, and real oslale, wn rn ir nay 14 - - - 1 1 1 8 1 2 - BOIVICBS. ++ vv wv vmv ms ome 226 8 - 3 7 3 13 1 187 3 1 Public administration. . . . ... 26 - - - 1 2 2 1 3 16 1 Armed Forces . .......... 5 - - - - - - - 1 1 3 American Indian: Allindustries ............. 62 10 2 8 15 3 4 1 12 7 - Agriculture . ............ 8 7 - - - 1 - - - - - MAING. cv wv wn cmivwin cme 1 - - - - - - - - Construction . . cs car vu 12 - 1 8 1 - - - - 2 - Manufacturing . . ......... 14 1 - - 12 - - - 1 - - Transportation, communication, andutilities . . .......... 3 - - - 1 2 - - - - - TBAB oon wns mv wwmrmesmy 4 - - - - - 3 - - 1 - Finance, insurance, and real OSE, 5 sans ures RE mn 2 - - - - - 1 1 = - - SOIVICES. «civ vn vir wmv min 12 1 - - - - - - 9 2 - Public administration . . . .... 6 1 — — 1 ao - - 2 2 - Armed Forces ........... - - —- - = i = - 38 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces Other: All industries . ............ 9 2 - - 2 1 - 1 1 1 1 AQHICURUIG . « «win www ons - - - - - - - - - - — MINING: :cermscasnsnng - - - - —- - - - —- - —- Construction . . .......... - - - - - - —- - —- - - Manufacturing . . ......... 2 - - - 1 - - = = 1 = Transportation, communication, and utilities . . .......... - - - - - - - - - - —- Trade ................ - - - - - - - - —- - - Finance, insurance, and real ostafe. oi; vs snp 1 - - - - - - 1 - - - BOIVICES. + us wmv inisinms nga 4 1 - - 1 1 - — 1 - - Public administration. . . . . . . 1 1 - - - - - - - - Armed Forces .. ... «vv: 1 - - - - - - - - - 1 Interval and industry 22-25 weeks: Allindustries .......cexxse 209 26 5 22 46 20 31 4 35 11 9 Agriculture. «.. «vows ev rus 22 19 - 1 1 - - - 1 - - yy Re 3 - 3 - - - - - - - - Construction . . .......... 20 1 - 16 2 1 - - - - - Manufacturing « « «sc ss 5 ws » 48 3 2 - 36 1 3 - 1 - 2 Transportation, communication, and utilities . ........... 23 - - - 3 18 1 - - 1 - TIBOB ins s me co ia iis mime 29 1 - 2 3 - 20 1 1 1 - Finance, insurance, and real ostale. sc anv raran 4 - - - - - 1 3 - - - BOIVICES: «sz vrs irms wan 46 - - 3 1 - 6 - 32 3 1 Public administration. . . . . . . 9 2 - - - - - - - 6 1 Armed Forces . .......... B —- —- - - - - - - - 5 26-29 weeks: Allindustries . ............ 452 38 8 32 123 42 55 13 104 24 13 Agriculture . ............ 37 28 - 1 3 1 w- — 4 - - Mining. ............... 7 1 4 - 2 - - - - - — Construction... c «sv esesss 36 1 - 24 5 1 2 - 2 - 1 Manufacturing . . ......... 120 2 1 2 100 4 4 1 4 1 1 Transportation, communication, and utilities . . .......... 46 9 - 2 3 34 3 - 2 1 - TEHE sv va vmsmnvy smn ug 55 2 2 2 6 1 38 - 2 - 2 Finance, insurance, and real BSI. sve x 12 - - - - - - 10 - 1 1 Services. ....... 109 3 - 1 4 1 7 1 89 1 2 Public administration. . . . . .. 24 —- 1 - - —- 1 1 1 19 1 AmedForces . .........: 6 - - - - - - - - 1 5 30-32 weeks: All industries . ............ 340 36 4 34 93 27 47 12 64 13 10 Agriculture . ............ 30 26 - 1 2 - 1 - - - - MINING ov c or emsnewnens 4 - 3 - - 1 - - - - - Construction « «cz: cvvsss 29 1 - 21 4 1 - - 1 - 1 Manufacturing . . ......... 91 5 1 2 71 1 5 1 3 1 1 Transportation, communication, and utilities . . .......... 27 1 - 1 2 20 — - 1 - 2 Trade cccsvssnrnnansws 48 2 - 1 3 1 37 - 4 - - Finance, insurance, and real oft. si wanna ms nnn 13 - - - - 2 10 - 1 - Services. . ............. 75 1 - 5 8 3 2 1 54 - 1 Public administration. . . . . . . 17 - - 2 3 - - - 1 10 1 Armed Forces . . + cvs vos wis 6 - - 1 - - — - - 5 4 39 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces 33-35 weeks: All industries . ............ 410 30 8 38 113 44 50 15 88 19 5 Agriculture . . ........... 27 21 - 1 4 1 - - - - - MINING. sa cns emsns swam 9 - 8 - — - - 1 - - Construction ii» «xs ws mas 37 - - 31 2 2 - - - 1 Manufacturing . . ......... 108 1 oe 1 89 3 9 2 3 - - Transportation, communication, and utilities . . .......... 40 2 - 2 1 30 2 - 2 - 1 THE oor mromamnimamh 46 - - 1 4 3 30 1 6 1 - Finance, insurance, and real BStalS, secs ms mam 13 - - - - 1 1 10 - 1 - BOIVICES . o.oo sm bn ae 108 5 - 1 11 3 6 2 75 4 1 Public administration. . . . . . . 20 1 - 1 2 1 2 - 1 12 - Amed Forces . .........: 2 - - - - - - - - - 2 36-38 weeks: All industries . ............ 359 34 10 34 88 32 42 14 76 16 13 Agriculture . ....... vacua 38 26 wa 1 7 - 2 - 1 - 1 MIBING. : s:asrasns emma 8 - 7 — - - - - 1 - - construction = « sews sms s 32 1 1 20 4 1 1 - 2 1 1 Manufacturing . . ......... 91 1 1 6 67 1 5 - 5 2 3 Transportation, communication, andutiities . . : ssa eww 34 - 1 - 1 27 1 - - 2 2 THOOG + svi sasmbemems 37 1 - 2 3 - 25 - 5 - 1 Finance, insurance, and real estate. . .............. 13 - - - 1 - 11 - 1 - DL 85 3 - 3 4 2 8 60 1 1 Public administration. . . . . . . 17 2 - 1 1 - - - 2 9 2 Armed Forces . .......... 4 - - 1 - 1 - - - - 2 39-41 weeks: All industries . ............ 267 22 5 32 62 16 38 6 62 19 5 AQHICURUIE ..«v vvivwvrwoms 19 17 - =~ 1 - - - 1 - - MINING: vi 2es ensme amamse 5 1 2 1 - - - - 1 - - Construction . . .......... 26 - - 21 2 1 - - - 2 - Manufacturing . . ......... 63 1 1 2 47 - 4 - 5 2 1 Transportation, communication, and utiles «= «sve sme 19 - - 1 2 13 1 - 1 - 1 Trade nies snsms smina 35 2 1 1 3 - 27 - 1 - - Finance, insurance, and real estate. . .............. 10 - - 2 - - 1 6 1 - - A 68 1 1 1 7 1 5 - 49 - Public administration. . . . . .. 15 - - 2 - 1 - - 1 11 Amed Forces ........... 7 - - 1 - - - - 2 1 3 42-44 weeks: Allindustries . ............ 242 23 5 30 41 17 43 9 58 14 2 Agriculture . ............ 22 17 r 1 1 - 1 - 2 - - MINING. : c:5: snsws sms 4 - 4 - - - - - - — - Construction . . .... xxx: 29 2 - 23 1 3 - - - - - Manufacturing . . ......... 51 wi 2 32 3 7 - 6 1 - Transportation, communication, andutiities . . . «cx: vv: 14 - - 1 1 9 2 - 1 - - TAUE ..ovcevrmrmnnmema 38 1 - 2 4 - 25 - 5 1 - Finance, insurance, and real OSHA. os vv inn wn 1 - - - 1 - 2 7 1 - - SOIVICBS. «+ vs sv sna smerny 56 3 - - 1 1 8 1 43 2 - Public administration. . . . ... 14 - 1 1 - 1 1 1 - 9 - Armed Forces . .......... 3 - - - - - - - - 1 2 40 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces 45-47 weeks: All industries . ............ 167 14 6 19 36 22 17 3 41 6 3 AGACURUIE: . « vv: «aie ivi on 3 0 12 1 - - 1 - - - po - = Mining. oc weewsuswssmss 4 - 4 - - - - - = = — Construction . «cu: sar aan 18 - - 17 1 - - - ~ — = Manufacturing . . ......... 37 - 2 - 26 1 3 - 4 - 4 Transportation, communication, and UHIBS « « «vis wv v ms a 18 1 —- - 1 15 - - - — 1 TraAG wus mamas wasn mss 21 —- —- — 6 3 11 - 1 - — Finance, insurance, and real estald. . vivre 6 - - - 1 1 - 3 1 — - SBIVICES: + vs vx nins ress 39 2 - 1 - - 2 - 33 1 - Public administration. . . . . . . 10 - - 1 — 2 1 - 2 4 — Armed Forces ........... 2 - - - - - - -= = 4 1 48-51 weeks: AlINAUSHABS « ; 5 mss ws mais 111 7 - 7 25 8 17 5 36 2 4 Agriculture . ............ 4 4 - - - - - wi ss - a Mining. ............... - - - - - - - = w= = Construction - ... vc. zx 45055 12 2 - 6 - - 1 - 1 1 1 Manufacturing . . ......... 30 - - - 24 1 4 - 1 - - Transportation, communication, and utilities . . .......... 9 —- — - 1 6 - 1 1 - - TS seven urs aznsas 14 - - 1 - - 10 = 2 = 1 Finance, insurance, and real OSIAB. ; is sums Ens EER 5 —- —- — - - - 4 1 = ow Services. . ............. 31 1 - - - 1 1 - 28 - - Public administration. . . . . .. 4 - - - - - 1 - 2 1 - Armed FOICeS + . «vv osu ov 2 - - - - - - - — = 2 52 weeks and longer: All industries . ............ 41 3 - 5 8 5 9 - 9 - 2 Agriculture . . ........... 7 2 — 9 2 1 1 - = a = MIBING + v5 co 0 5 ory win mie wee — - - - - - - - = = = Construction « : . «. «25:5: 5 1 - 2 - - - — 1 Manufacturing . . . ........ 8 - —- 1 6 — - ws - ~ 1 Transportation, communication, and utilities . . .......... 3 —- —- — - 3 = = - = = Tale «coves vosvsnniws 6 - - - - - 6 on = — — Finance, insurance, and real Stale, «arm vm ak ad me - - - — - - - a a ” = Services. . ............. 11 - - - - i 2 = 9 i = Public administration. . . . . . . - - - - - - = - = = = AMY FOICES . v «ss: 5:40 1 - - 1 —- - - ~ ~ - — Missing: All industries . ............ 4 - - 1 2 1 i ss = - - Agriculture . . ........... - - —- — - - - wo = - MIDING: us crevsrrsms nas - —- - —- - - - - = = — Construction . . wx vss sas 1 - - 1 - - - — - — — Manufacturing . . . ........ 1 - - - 1 — - ss si -~ — Transportation, communication, and UEHHES: & « «x +4 + v5 was 1 - —- — - 1 = = Pe = ~ Trade asus vmems ams nsw - - - - - - — — — oo — Finance, insurance, and real estate. . .............. — - - - - - i - - - - Services. . ............. 1 - - - 1 - - - - = we Public administration ...... - _ - - - - = -— = = - Armed Forces «+. xs sss 53 4 - - = - = = a — — _ - 41 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces Spouse: All industries . ............ 1,218 99 20 137 324 130 155 52 194 73 34 Agriculture . ............ 94 73 - 3 11 - 4 - 3 - - MINING. cc «vs vr xim snes 24 1 17 1 2 1 - - 2 - - Construction . = » «x sw vv 5 50 131 4 1 102 10 7 2 - 1 1 3 Manufacturing . . ......... 332 7 1 9 255 11 26 3 12 4 4 Transportation, communication, and utilities . . .......... 126 3 - 4 7 98 3 1 5 2 3 Tad « + cove swiss vwvan 149 5 1 5 17 7 102 — 9 3 - Finance, insurance, and real estate. . .............. 51 - - 2 - 1 3 42 - 2 1 Services. . ............. 211 3 - 6 17 3 13 4 157 7 1 Public administration. . . . . . . 81 3 - 5 5 2 2 2 4 53 5 Armed FOICBS « «vn wv awas 19 - - - - - - 1 1 7 Parent: AINAUSHIES +... « «wc vvn v van 196 9 1 27 50 18 26 5 40 8 12 AQTICURUI® «+ sss vn vv ens 10 7 - 1 - 2 - - - - - MINING: s seus sans s Bans 2 - 1 - - - - - 1 - - Construction . . .......... 27 2 - 19 1 2 - - 2 - 1 Manufacturing . . + «vv «ve 54 = - 3 39 1 2 1 4 1 3 Transportation, communication, ano utities .: susan vam 18 - - 1 3 12 2 - - - - Trade ................ 27 - - 2 1 - 16 — 4 - 4 Finance, insurance, and real osiale, cs severe sun 5 - - - 1 - - 3 - 1 - BOMVICES csv sas naan in 45 - - —- 5 1 5 1 29 1 3 Public administration. . . .. .. 5 - - 1 - - 1 - - 3 - Armed Forces . . ......... 3 - - - - - - - - 2 1 Child: AlINGUSHIBS , «cv cv si ssw ena 163 23 4 5 41 10 23 3 44 7 3 AQHCUIUIS . i. ccovsn sian 20 18 - - 2 - - - - - - Mining. ............... 1 - 1 - - - - - - - - Construction . . .......... 6 - - 4 2 - - - - - - Manufacturing « « «vv sow wns 36 - 3 - 30 - 2 - 1 - - Transportation, communication, and utiles . . ccs very vn 12 - - - - 9 2 - - - 1 Trade ................ 18 1 - - 2 - 14 - 1 - - Finance, insurance, and real cs 5 - - - 1 — 1 3 - - - BOIVICES. «vw va we nnaw 108 53 3 - 1 4 1 4 - 40 - - Public administration. . . .. .. 8 1 - - - - - - 2 5 Armed Forces . .......... 4 - “ - - - - - - 2 2 Sibling: AlINCUSHIBS . .. uc vnrnw ins 78 5 1 1 18 7 9 2 23 2 - Agriculture . ............ 7 3 - - 4 - - - - - - Mining. ............... 1 - 1 - — - - - — - — Construction +. «+ + v www wows 12 1 - 10 - - - - 1 - Manufacturing . . ......... 16 - - - 12 2 1 - 1 - - Transportation, communication, and utilities . . .......... 7 - - - - 5 1 - 1 - - TBAG + vv sv vvv iv wmv ww mw 8 - - 1 1 - 5 - 1 - - Finance, insurance, and real OSIAMG: 5 ov wwe 2 YE EE 2 - - - - - 2 - = - SOIVICES. « : «cuss riser ia 24 1 - - 1 - 2 - 20 - - Public administration. . . . . .. 1 - - - - - - — = 1 - Arad FOICBS + «vn - - - - - - - sm wr — — 42 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 — Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul Construc- ~~ Manufac- and and real adminis- ~~ Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces Other relative: Alindustries ...... ova vee 50 9 2 3 8 3 8 1 15 - 1 AQHCURUIg . . «vv vv ove us 9 8 - 1 - - - - - - - MIB «5 2 sae on dunk os mn 1 - 1 - - - - - - - - Construction . . .......... 2 - - 1 - - - Ke 1 - - Manufacturing . «+. cv sv 5 4 8 - - 9 6 - - - 1 - - Transportation, communication, and utilities . : cscs em san 5 - - - 1 3 - - 1 - - Trade ................ 9 - 1 - 1 - 7 - - - - Finance, insurance, and real OSIMG. vv vv awww ws we 1 - - - - - - 1 - - - BOVICES. +» vs swans a HE 14 1 - - - - 1 - 12 - - Public administration. . . . . .. - - - - - - - - - - - AImet) FOIBBS. «vx vs wiave vs 1 - - — - - —- - —- - 1 Nonrelative: Alindustries ...« iva: cuss 27 1 - 2 5 4 3 1 9 1 1 Agriculture . ............ — - - - - - - - - - - Mining. ............... - - - - —- - - - - - - Construction +: us ove ws ¥ ws 2 - - 1 1 - - - - - Manufacturing . . ......... 6 - - - 5 1 - - - - - Transportation, communication, ARO UIE: «+ ccov vi aviv cow 20 aos 2 - - - - 2 - - - - - Trade cvvovpninvnne wre 2 - - - - - 2 - — - - Finance, insurance, and real BIR cls tl incu ne rere in 1 - - - - - - 1 - - - BOVICEE. vv wi wevinn mms win 12 1 - q - - 1 - 9 - - Public administration. . . . . .. 1 - - - - - - - - 1 - Armed Forces . .......... 1 - - - - - - - - - 1 Not stated: AL IRGUBHIBE. vv 5 ve ws mmr mas 870 87 23 69 191 62 125 17 248 33 15 Agriculture « , ... sxsw nis 78 62 a 2 5 1 1 - 6 a MINING 5.5 55 ai 5 ars ie ims 15 1 14 - - - - - - - - Construction . . .......... 65 2 - 45 8 1 2 - 2 3 2 Manufacturing . . ......... 196 6 4 3 152 - 13 - 13 2 3 Transportation, communication, and utilities . . .......... 64 2 1 2 4 47 2 - 1 2 3 Trade ................ 116 3 1 4 10 1 83 2 12 - - Finance, insurance, and real OStale, . ws nw rna pes way 22 - - - 1 1 3 12 4 1 - Services. . ............. 270 10 1 7 10 7 18 3 205 7 2 Public administration. . . . . .. 34 1 2 2 1 3 3 - 4 18 - ABC FOES . . vs vu swans 10 - - 4 - 1 - - 1 - 4 Death certificate informant and survey respondent were — Both spouse: Allindustries . ............ 1,055 85 17 116 289 117 129 46 158 67 31 Agriculture «ovo ws via 80 62 - 2 10 — 3 - 3 - - MINING: «x sms uwsms swmsme 21 1 15 - 2 3 - - 2 - - Construction . . .......... 116 3 1 90 9 7 2 - 1 1 2 Manufacturing . . ......... 287 7 - 7 225 9 21 3 9 3 3 Transportation, communication, and utiles «: cv usm ws 112 3 - 4 6 88 2 - 4 2 3 Trade .uwinimvimsamems 128 5 1 4 17 7 85 - 6 3 - Finance, insurance, and real estate. . .............. 45 - - - - 1 38 - 2 1 CL 177 2 - 6 16 3 1 3 128 7 1 Public administration. . . . . .. 7 2 - 3 4 1 2 4 48 5 Amed Forces . .......... 18 - - - - - - - 1 1 16 43 Table 6. Number of responses by industry of decedent on National Mortality Followback Survey questionnaire and by race and industry on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Industry on questionnaire Transpor- tation, communi- Finance, Race and industry on death cation, insurance, Public certificate, interval and Agricul- Construc- ~~ Manufac- and and real adminis- Armed relationship Total ture Mining tion turing utilities Trade estate Services tration Forces Not both spouse: Allingustries . .. cv ccoen ius 1,476 144 29 129 326 109 216 33 407 50 33 AQHCUILIIG . «vv sis einis ims 133 106 - 5 11 2 2 — 6 - 1 BIRT vx: a mes oe 10s im 20 3 17 1 - - - - 1 Construction . . .......... 122 6 - 88 10 4 2 - 5 4 3 Manufacturing . . ......... 339 5 7 8 257 6 22 1 22 4 7 Transportation, communication, and utilities . . .......... 115 2 1 2 8 84 8 1 4 2 3 TAOS + cv vu mms mnnmsma 198 4 2 8 14 1 142 2 21 - 4 Finance, insurance, and real cg gg 39 - - 2 3 1 4 24 4 1 - SOVICES. oi sus umn ma 441 17 - 8 21 7 33 5 337 8 5 Public administration. . . . .. . 52 3 2 5 2 4 3 - 6 27 - AME FOICBS «+ « vw viv» ws V7 - 2 - - = wa 1 4 10 Not stated: All industries . ............ 7 4 5 9 22 8 4 2 8 7 2 AQriCURUIB . «vv vn nvm 5 3 - - 1 1 w- - — - - Mining: «scorns emsnmsws 3 - 3 - - - — - - — Construction .. . ... .os ssc 7 - - 4 2 - - - - - 1 Manufacturing . . ......... 22 1 1 1 17 - 1 - 1 - - Transportation, communication, and utilities « : ws cv ews sus 7 - - 1 1 4 - - - - 1 TIES «iv vrmmrmemnnms 3 - - - 1 - 2 - - - - Finance, insurance, and real estate. .......c.0nun 3 - - - - - - 2 - 1 - Services. . ...... 11 - 1 1 - 2 - - 7 - - Public administration. . . . . . . 7 - - - - - 1 - - 6 - Armed Forces . .......... 3 - - 2 - 1 - - - - - NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 44 Table 7. Number of responses by veteran status of decedent on National Mortality Followback Survey questionnaire and by race and veteran status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 Veteran status on questionnaire Race and veteran status on certificate, interval, and relationship Total Veteran Nonveteran Race and veteran status All races: AIVRIeran Statuses . ; : sc vi mew hi mew maw simmembmus 10,637 2,121 8,516 WOIBIAN viv vis ma sais win hmatis $ isime Rata 9a #0 a leians 2,176 1,974 202 INONWVBIBIARI. . viv vs sms md samen demas sHmaimb & 53 mns 8,461 147 8,314 White: AVelSran SIalUSES « s suis wis sais nin vite $i ame GRD 7,509 1,622 5,887 VSIBTAN + vv voivws smeims Emons toss MAME EWAWY 1,659 1,513 146 NORVBIBIAN . «uc ws rmsms rims ms em tms emamns sass + 5,850 109 5,741 Black: AIVeIeran StallUSe8 iz csc ws sn a Ms aims Wa sais we mma ws 2,781 450 2,331 WVBIBIBI «os ivn sins ma sms mim emiame swe me sms mm ans mw 461 415 46 NORYBIBIAN « + « + wom ss msmasmhmu rw smmen sma als 2,320 35 2,285 American Indian: Alveloran SIatuBes ,. , .. cas ws cmsmr sv IE NEB IRP EW EW 257 35 222 VBISTAN © vs viv nrssmswmamem nisms am ame sds mn male 39 33 6 INONIVBLBIAIY « is vo wv a 0 is fy mis a [ine 0 Ii 6 ob 0 3 35% 6 B00 G0 cat i 0 Gm 218 2 216 Other: Al VEISIBIY SIARIBES: «or o.oo 5 bi ss wi asi hws wa am 5 0 % 6 0 8 72 11 61 WBIBEAN cc. + i svi ws mn 0 i 9 WB Te A 4 I 9, 8 3 a 200 2 0 13 10 3 NONVEIBIAN . « . ccc nsmrsmsmanmsms bmrmwamsnmen 59 1 58 Missing: AUVEISran SIatUBe. » . vcs rm rnb sms sv sms RN rR 18 3 15 VEBIBIAR «oo: 550 250 5: 4105 9.00% 418 3.08% 3:0 900 5 +08 30% B45 304 iy 4 3 1 INONVBLBIAN . ov « sm rs mss a +0 8 0 io 8 0 400 8 0 0 ie 0 al ol 14 - 14 Interval and veteran status 22-25 weeks: Alvaloran Statuses . . . «sce sv rms rus va sas Rae w re 833 191 642 VBIBIaN «cc crn rata nsa ssn sd eins es ais © omens 193 183 10 NOAVBIBIAN . ohn stn ne sv msm 50 28 A SES 248 20 2am iy ib 640 8 632 26-29 weeks: Al VOIETaNSIAIUBES , ore io vos mini 3 0 he 10 300 1918) 038 Torii i 8 9 1,616 334 1,282 VERON 1s: iv: 0 300 000 30 18 oes i 0 38h SLR oi 0 338 306 32 INGIVBIBERIY... 5.05. ar tom om: 3150 high A 90 2 tor 00 Sold ot dh 9 1,278 28 1,250 30-32 weeks: A VIeran SIalUSES «+ ov vw vis w viv sin @ + 3 sine Ewe Few. 1,423 281 1,142 WVBIBIAIY «5a 00 50 (6 50 ih JE iE) 0 500 16 9100 500 es Eb 291 265 26 NOOVAIBIBN. +s 2% 2 mss sma ms s BESS S2TBS PBEBS » 1,132 16 1,116 33-35 weeks: AlVRISran SIAIUSES . «ovo vv vm in v vv on va xis Es sus 1,656 334 1,322 VBIBIAN. «0s + iid 40 oti 5 0 hl] HAIR) 0 it oe hm be 353 312 41 NONVBIBIAN . + + sr vas v Ess 3 4B EXE WE SERED S 5x Dwd # 1,303 22 1,281 36-38 weeks: ALVOISran SIatUSeS . . . . sv vv svi mmo ws bw wae we 1,416 290 1,126 VBIBIAN «css issss pasa ss mas a ness puss s neon 283 261 22 NONVEIBIEN . c+ sis so vu nes we Bens ws Ha ene bens 1,133 29 1,104 39-41 weeks: AN VASran StAIUSES . vv waa mis 4% $516 4h 9.54 we vw neo 1,236 225 1,011 VBIOIEN vv vas % aE 24 min & Sins #0 980 5 & £0 E04 Ww 256 221 35 NOOWELBIAI .. ..c v 2 vin s vans SHR FE REAL DPE T 39 980 4 976 42-44 weeks: Allveteranstatuses . . . ................. in 1,028 189 839 VOIBIAN + + vous u as vse vas ns pases ns gases wus 188 174 14 INOIIVBIBIBIV. vv. 515 0:0 or 0 t 80 (80 i 08 3 165 0 85 0 0 G0 HE 8 840 15 825 45-47 weeks: AN VBloran StatUBes . o.. crv ns vrais sa EAE TERETE Ba swe 720 135 585 VBLBIAN vv ier win mx tn How oh ors dhs OR 0 B14 Wi 5 Be 131 123 8 INOOVBISEBIN x wis v4 orp 3s # pie Sak EB % BEE + EE ah 589 12 577 45 Table 7. Number of responses by veteran status of decedent on National Mortality Followback Survey questionnaire and by race and veteran status on death certificate, interval between death and survey, and relationship of informant to decedent: United States, 1986 —Con. Veteran status on questionnaire Race and veteran status on certificate, interval, and relationship Total Veteran Nonveteran 48-51 weeks: Alveleran Statuses . . v: sec ss sss as su RRT LEE HME DET 454 81 373 NVEIBTAN co. ia vive sista siotion sonra wise Solinsks Bolin olds Site Bile, SU 82 75 7 Nonveteran. . . ......... iin 372 6 366 52 weeks and longer: AlVvaloran statuses ns ss srs AEs RETR ARATE ARES 241 59 182 Veteran . . . oo... 59 52 7 Nonveteran. . . c cor u swvenvvrsvrswres ss ewww rena 182 7 175 Missing: AN VOISIG IANS: cs sn sna dd REAR EIRENE SRR REE 14 2 12 Veteran . . o.oo. ea 2 2 - INONWVBIBIBIN... = 5 4 x vv sown vow aemian som win a wpm sms 12 - 12 Relationship Decedent was death certificate informant’s — Spouse: Al VOIBIaN SIAUSES «vs u s vv sp vss sms sur wo vu ow wn 4,015 1,079 2,936 WSHBFAN 5.100 00m otc 11 ives she masons us Boas 6) Fos ld 0 40 B00 #10 ts 1,112 1,014 98 Nonveteran. . . ......... iii 2,903 65 2,838 Parent: AlVleran SIaluBBS . «ova sn sR EEF EET wan 8s 0 8 HEH 718 143 575 VYOIOIBN i vvt vr rmta ti dRe se i ntis MIs A uate s Bas 146 133 13 Nonveteran. . . .......... iin 572 10 562 Child: AIvOleran StaluSES .. & cv sn tHe sR PET FRIES RH ARE 8 5 0 Y 1,013 113 900 NSIBTAN © ves coo snstamnn: 4 on wl 9008 Bundt i Bed 218 51 P00 Bh 80 5 4 114 102 12 Nonveteran. . ............ iin. 899 11 888 Sibling: Alveloran Sialuses . . . c. uses sss swarms pew ww ws 326 60 266 LL EE TE TT ITI ITTY, 60 55 5 Nonveteran. . ............ inn nnnn. 266 5 261 Other relative: Alvetoran Statuses . . curbs insmr sR ims eH IHD 265 30 235 VOIBIAN wis in ies SHAR SRAM MEME A ARE ARE a 31 28 3 Nonveteran. . . ....... iii 234 2 232 Nonrelative: Alveteran Slaluses . . . ows vos es sms Es sma s ssn 215 42 173 VEIBIER . vv vs mr sm imemitmi dni Bia AEs Ab 0F 46 39 7 Nonveteran. . . ........ iii 169 3 166 Not stated: All Veteran StatUSES oz vis ws vu swe s Ere suis Uy Fuse h 4,085 654 3,431 VeIBIAN vv vos iis shims db amb Amemaw mans same s 667 603 64 Nonveteran. . . ........ iii 3,418 B 3,367 Death certificate informant and survey respondent were — Both spouse: All veteranstatuses . . . .......... ii 3,417 970 2,447 Veteran . oo... 1,002 914 88 NONVRIOIBN . 5 sv 6 vic 0 9 i 5 08 0 Ww wr 750 00 3 97% Ta 0 6 0 2,415 56 2,359 Not both spouse: Allveteranstatuses . . ............. 6,778 999 5,779 Veteran . ....... ii 1,014 915 99 INGIVBESIANS wv: ee wis iv wt v Tr ot 00s 1000 0 0 4 0 3 07 SO 50 0 5,764 84 5,680 Not stated: Allveteranstatuses . .................cc0i0vnnn 442 152 290 Veteran . . oo... 160 145 15 NONVBIBIAN. . » . wc: enews was ns wmsms wmv ms sms sre 282 7 275 NOTE: Oregon, Nebraska, Nevada, and New Mexico were excluded from the comparability analysis in this report because Oregon's confidentiality requirements precluded its participation in the 1986 National Mortality Followback Survey, and the primary matching criteria could not be applied for the other three States because they reissue death certificate numbers after the processing of the Current Mortality Sample. 46 Appendixes Contents IL U.S. Standard ‘Certificate Of Death «crv vv vss vassmersin vr sewers vnsms II. Instructions for filling death certificate items......................... III. Selected questionnaire items, 1986 National Mortality Followback Survey Se 0 0 0 4 se ss ss ss ee ee eee sees eee eee ee ee ss ss se ss ee eee ete ese ee eee eee 47 Appendix | U.S. Standard Certificate of Death TYPE OR PRINT IN PERMANENT INK FOR INSTRUCTIONS HANDBOOK IF DEATH OCCURRED IN INSTITUTION, SEE HANDBOOK REGARDING COMPLETION OF RESIDENCE ITEMS, 1978 REVISION EDUCATION, AND WELFARE—-PUBLIC HEALTH SERVICE—NATIONAL CENTER FOR HEALTH STATISTICS CONDITIONS IF ANY WHICH GAVE RISE TO IMMEDIATE CAUSE STATING THE UNDERLYING CAUSE LAST DEPARTMENT OF HEALTH, LIT DEATH HRA-162-1 Rev, 1/78 48 (PHYSICIAN, MEDICAL EXAMINER OR CORONER) Form Approved U.S. STANDARD OMB No. 68R 190] LOCAL FILE NUMBER CERTIFICATE OF DEATH STATE FILE NUMBER DECEDENT-NAME FIRST MIDDLE LAST SEX DATE OF DEATH (Mo., Day, Yr.) } 2 3 RACE~(e.g., White, Black, American AGE —Last Birthday UNDER 1 YEAR UNDER 1 DAY | DATE OF BIRTH (Mo., Dey, Yr.) [COUNTY OF DEATH Indian, etc.) (Specify) (Yrs) MOS, T DAYS HOURS 1 MINS. 4 Sa. 5b. Se. 1 6. Ta. CITY, TOWN OR LOCATION OF DEATH HOSPITAL OR OTHER INSTITUTION—Name (I/ not in either, give street and number) IF HOSP, OR INST, Indicate DOA, OP/Emer. Rm, Inpatient (Specify) 7b. Tc. 7d. STATE OF BIRTH (If notin US.A., [CITIZEN OF WHAT COUNTRY |MARRIED, NEVER MARRIED, |SURVIVING SPOUSE (If wife, give maiden name) WAS DECEDENT EVER IN U.S, name country) WIDOWED, DIVORCED (Specify) ARMED FORCES? (Specify Yes or No) 8. 9. 10. 11. 12. SOCIAL SECURITY NUMBER USUAL OCCUPATION (Give kind of work done during most of KIND OF BUSINESS OR INDUSTRY working life, even If retired) 13. 14s. 14b. RESIDENCE-STATE COUNTY CITY, TOWN OR LOCATION STREET AND NUMBER INSIDE CITY LIMITS (Specify Yes or No) 15a, 15b, 15¢. 15d. 15¢, FATHER-NAME FIRST MIDDLE LAST MOTHER-MAIDEN NAME FIRST MIDDLE LAST 16. 17. INFORMANT—NAME (Type or Print) MAILING ADDRESS STREET OR R.F.D. NO. CITY OR TOWN STATE z1P 18a. 18b. BURIAL, CREMATION, REMOVAL, OTHER (Specify) CEMETERY OR CREMATORY-NAME LOCATION CITY OR TOWN STATE 19a. 18b. 19c, FUNERAL SERVICE LICENSEE Or Person Acting As Such (Signature) 20a. » NAME OF FACILITY ADDRESS OF FACILITY 20b. 20c. F3 213. To the best of my knowledge, death occurred at the time, date and place and due to the 22. On the basis of examination and/or investigation, in my opinion death occurred at the time, ry causa(s) stated, « date and place and due 10 the cause (s) stated. 0 >W > 3 9 (Signature and Title) > 3 z & (Signature and Title) » i DATE SIGNED (Mo., Day, Yr.) HOUR OF DEATH is z DATE SIGNED (Mo., Dey, Yr.) HOUR OF DEATH 55 3 uw 3 82° 2m. 21c. M | 83g ab. 22. M ° SE NAME OF ATTENDING PHYSICIAN IF OTHER THAN CERTIFIER (Type or Print) 28% PRONOUNCED DEAD (Mo., Day, Yr.) PRONOUNCED DEAD (Hour) re Fu w 3 © 21d. 22d. ON 22¢. AT M NAME AND ADDRESS OF CERTIFIER (PHYSICIAN, MEDICAL EXAMINER OR CORONER) (Type or Print) 23. REGISTRAR DATE RECEIVED BY REGISTRAR (Mo,, Day, Yr.) 2a. signature) > 2. IMMEDIATE CAUSE [ENTER ONLY ONE CAUSE PER LINE FOR (a), (b), AND (c).] | Interval between onset and death PART | I (a) 1 DUE TO, OR AS A CONSEQUENCE OF: | Interval between onset and desth | (b) 1 DUE TO, OR AS A CONSEQUENCE OF: \ Interval between onset and desth | (c) | PART OTHER SIGNIFICANT CONDITIONS~Conditions contributing to death but not related to cause given in PART | (a) AUTOPSY (Specify Yes| WAS CASE REFERRED TO MEDICAL un or No) EXAMINER OR CORONER (Specify Yes or No) 26, 27. ACC., SUICIDE, HOM, UNDET., | DATE OF INJURY (Mo., Day, Yr.) HOUR OF INJURY DESCRIBE HOW INJURY OCCURRED OR PENDING INVEST, (Specify) 28a, 28b. 28c¢. M | 28d. INJURY AT WORK (Specify Yes |PLACE OF INJURY ~—At home, farm, street, factory, office building, LOCATION STREET OR R.F.D. No, CITY OR TOWN STATE or No) etc. (Specify) 28e. 281. 28g. ppendix II 1structions for filling death certificate items ba-c. Age Make an entry in either 5a, 5b, or 5c depending on the age of the decedent. Sa. LAST BIRTHDAY (YEARS) Enter the age of the decedent at last birthday. If the decedent was under 1 year of age, leave this item blank. 4. Race—White, Black, American Indian, Etc. (Specify) Enter the race of the decedent as stated by the informant. For groups other than white, black, or American Indian, obtain the national origin of the decedent, such as Chinese, Japanese, Korean, Filipino, Hawaiian, etc. If the informant indicates that the decedent is of “Mixed race,” enter both races or national origins. OTE: From Funeral Director’s Handbook on Death Registration and Fetal Death Reporting, U.S. Department of Health, Education and Welfare, Public Health Service, National Center for Health Statistics, DHEW Pub. No (PHS) 78-1109, July 1978. Origin or Descent A. ORIGIN OR DESCENT QUESTION Origin or descent (e.g., Italian, Mexican, Puerto Rican, English, Cuban, etc.). (Specify.) B. SPANISH-ORIGIN QUESTION Was decedent of Spanish origin? (Specify “Yes” or “No.”) If “Yes,” specify Mexican, Cuban, Puerto Rican, etc. These items are not on the U.S. Standard Certificate of Death, but are recommended by the National Center for Health Statistics and one or the other of the alternatives shown above has been adopted by a number of States. It may be in the form of a general origin or descent question (A) or a specific Spanish-origin question (B). For purposes of this item, origin or descent refers to the nationality group of the decedent or his ancestors before their arrival in the United States (except for American Indian and Alaskan native). There is no set rule as to how many generations are to be taken into account in determining ethnic origin. A person’s origin may be reported based on the origin of a parent, a grandparent, or some far removed ancestor. The response is to reflect what the person considered himself or herself to be, and is not based on percentages of ancestry. Some persons may not have identified with the foreign birthplace of their ancestors or with a nationality group and the informant may report “American.” If, after clarification of the intent of this item, the informant still feels that the decedent was “American,” enter “American” on the record. If the respondent reports that the decedent was of multiple origin, enter the origins as reported (e.g., English-German). If a religious group is reported (e.g., Jewish, Moslem, Prot- estant), ask for the country of origin or nationality group. This item is not a part of the race item. Both questions, race and origin or descent, should be asked independently. This will mean that for certain groups the entry will be the same in both items (e.g., Japanese, Chinese, Hawaiian). Even if they are the same the entry should be made in both items. 10. Married, Never Married, Widowed, Divorced (Specify) Enter the marital status of the decedent at time of death. Specify one of the following: married, never married, widowed, divorced. A person is legally married even if separated. If marital status cannot be determined, enter “Unknown.” Do not leave blank. 50 14a. USUAL OCCUPATION (GIVE KIND OF WORK DONE DURING MOST OF WORKING LIFE, EVEN IF RETIRED) Enter the usual occupation of the decedent. This is not necessarily the last occupation of the decedent. “Usual occupation” is the kind of work the decedent did during most of his or her working life, such as claim adjuster, farmhand, coal miner, housewife, janitor, store manager, college professor, civil engineer, etc. “Retired” is not an acceptable entry. Enter “Student” if the decedent was a student at the time of death and was never regularly employed. 14b. KIND OF BUSINESS OR INDUSTRY Enter the kind of business or industry to which the occupation listed in 14a was related, such as insurance, farming, coal mining, hardware store, retail clothing, university, government, etc. Do not enter firm or organization names. 12. Was Decedent Ever in U.S. Armed Forces? (Specify Yes or No) If the decedent was a veteran, enter “Yes.” If the decedent was not a veteran, enter “No.” If veteran status cannot be determined, enter “Unknown.” Do not leave blank. 7a-d. Place of Death 7a. COUNTY OF DEATH Enter the name of the county where death occurred. 7b. CITY, TOWN, OR LOCATION OF DEATH Enter the name of city, town, or location where death occurred. 7c. HOSPITAL OR OTHER INSTITUTION-NAME (IF NOT IN EITHER, GIVE STREET AND NUMBER.) Hospital Deaths If the death occurred in a hospital, enter the full name of the hospital. If death occurred en route to or on arrival at a hospital, enter the full name of the hospital. Nonhospital Deaths If the death occurred at home, enter the house number and street name of the place where death occurred. If the death occurred at some place other than those described above, enter the number and street name of the place. 7d. IF HOSPITAL OR INSTITUTION, INDICATE DOA, OP/EMERGENCY ROOM, INPATIENT (SPECIFY) If the decedent was pronounced dead in the hospital or other institution, specify whether dead on arrival, outpatient or emergency room patient, or inpatient. If death occurred on a moving conveyance other than en route to a hospital, enter as the place of death the address where the body was first removed from the conveyance. If death occurred in international waters or airspace, or in a foreign country, contact the State office of vital statistics for instructions. 51 Appendix li Selected questionnaire items, 1986 National Mortality Followback Survey Age 1. How old was the person at the time of death? 005 Age in years Race 15. Whi 147 5 Yiichomagory Be ST represaris 1 [J American Indian, Aleut, or Eskimo 2 [J Asian or Pacific Islander Mark (X) only one box. si Black 4 [J White Hispanic origin 16. Was this person of Spanish or Hispanic Jeo origin or descent? 10 Yes 20 No Marital status 19. At the time of his or her death, what was 187 : ; : the marital status of the person? 11] Married — Skip to question 21 2] Widowed — Go to next question 3] Divorced } Ski on 4] Separated pio-question 5s] Never married — Skip to question 26 on page 2 Occupation 134 had, what KIND OF WORK did he or she do the longest? (For example, electrical engineer, stock clerk, typist, farmer, in Armed Forces, etc.) 2. Of all the PAID jobs or businesses the person ever Industry INDUSTRY did he or she work in the longest? Describe the activity at the location where employed. (For example: TV and radio manufacturing, retail shoe store, State Labor Department, farm, Armed Forces, etc.) 4. In this occupation, what KIND OF BUSINESS OR 136 52 ‘eteran status 13. was the person ever on active duty in i. . the U.S. Armed Forces? 1 [J Yes — Go to next question 2 [J No — Skip to question 15 NOTE — Mark ’No”’ if all of the active duty service was related to training in the National Guard or military reserve. 'lace of death 29. Where did the person die? 093 1 [J In a hospital emergency room 2 [J In a hospital, not in emergency room 3 [J On the way to a hospital 4 [J In a nursing home or personal care home 5 [1 In his or her own home 6 [J In some other place — Specity 7 Mark (X) only one box. 1.8. Government Printing Office: 1983 — 301-019/80016 53 Reviews of New Re Ports From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics The National Mortality Followback Survey: 1986 Summary for the United States Series 20, No. 19 (PHS) 92-1856 Author: Seeman, I. For information contact: Kathi Brannan Scientific and Technical Information Branch 6525 Belcrest Road, Rm. 1064 Hyattsville, MD 20782 Tel: (301) 436-8500 The National Center for Health Statistics has just released a report that provides national estimates of the incidence of significant characteristics of adults who died in the United States in 1986. The report, “National Mortality Followback Survey: 1986 Summary, United States” presents data on the use of health services, disabilities, lifestyle practices that may affect health and mortality, and socioeconomic circumstances for the adults that were studied. Selective information from this survey has been weighted and is presented in 81 comprehensive tables. The data have been arranged into three categories: health care in the last year of life, lifestyle and health, and socioeconomic characteristics of decedents. According to the section on health care, some institutional care in the last year of life —in either a hospital or nursing home —was required for 81.1 percent of the decedents. Medicare covered 72.9 percent of all decedents and 92.3 percent of those age 65 and over. Even with coverage, about 12 percent of the elderly had somewhat serious to very serious problems with payments of medical bills. In the section on lifestyle, the incidence of heart attacks at some time during a lifetime was reported for 29.2 percent of the decedents; whereas, the incidence of Alzheimer’s disease and other memory impairments was 11.2 percent. According to the study, approximately 55.6 percent of the decedents smoked at least 100 cigarettes and 71.9 percent drank at least 12 alcoholic drinks in a lifetime. Approximately 16.9 percent of the decedents had exercised vigorously at least three times a week for at least 20 minutes each time, while 66.4 percent of the decedents hardly or never exercised. According to the section on socioeconomic characteristics, 27.2 percent of the decedents lived alone; for women it was 35.9 percent and for men 19.1 percent. Educational attainment was at the elementary school level for 32.5 percent of the decedents; the high school level for 42.1 percent; and the college level for 17.7 percent. Approximately 86.2 percent of the decedents worked at some time during their lifetime, and 13.1 percent of the decedents worked until the time of death. The data presented in this report are from the fifth survey in a series of National Mortality Followback Surveys (NMFS) conducted by NCHS. The 1986 NMFS data were collected from 16,598 informants; 81 percent of whom were the next of kin or another close relative of the decedent. Information about the decedent secured from informants was supplemented by data collected from hospitals, nursing homes, and other health care facilities in which the decedent spent at least one night during the last year of life. Copies of the report can be purchased from the U.S. Government Printing Office by completing the order form on the back of this release. SERVICE, oN a, § BEALTY 0 € %, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics CDC CENTERS FOR DISEASE CONTROL AND PREVENTION Publication Order Form Mail to: Superintendent of Documents Government Printing Office Washington, D.C. 20402 J YES, please send me__ copies of GPO Stock Number 017-022-01170-1 Price $15.00 The total cost of my orderis $________. Foreign orders please add an additional 25%. Prices include regular domestic postage and handling and are good through September 1993. After that date, please call Order and Information Desk at (202) 783-3238 to verify prices. Please Type or Print Please choose method of payment: [] Check payment to the Superintendent of Documents [] GPO Deposit Account [TTT ITTIT1] — [] (Additional address/attention line) [] VISA, MasterCard, or Choice Account CLIP PrP riry] (Company or personal name) (Street address) (City, State, ZIP Code) (Signature) ( ) (Daytime phone including area code) (Credit card expiration date) U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES BULK RATE Public Health Service POSTAGE & FEES PAID Centers for Disease Control and Prevention PHS/NCHS National Center for Health Statistics PERMIT No. G-281 6525 Belcrest Road Hyattsville, Maryland 20782 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 , he Library - UC Berkeley Received on: 05-05-94 IRS ital and health statistics. Series 2, Data evaluation and methods esearch une 1963-July 1971: Public Health Service pub Vital and Health Statistics From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics nvestigation of Nonresponse Bias: Hispanic Health ana Nutrition Examination survey December 1993 anita U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES 5 yy” Public Health Service i x Centers for Disease Control and Prevention Me National Center for Health Statistics CENTERS FOR DISEASE CONTROL AND PREVENTION Trade name disclaimer The use of trade names is for identification only and does not imply endorsement by the Public Health Service, U.S. Department of Health and Human Services. Copyright information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested citation Rowland ML, Forthofer RN. Investigation of nonresponse bias: Hispanic Health and Nutrition Examination Survey. National Center for Health Statistics. Vital Health Stat 2(119). 1993. Library of Congress Cataloging-in-Publication Data Rowland, Michael, 1945— Investigation of nonresponse bias : Hispanic Health and Nutrition Examination Survey / by Michael L. Rowland and Ronald N. Forthofer. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 119) (DHHS publication ; no. (PHS) 93-1393) “December 1993.” Includes bibliographical references. ISBN 0-8406-0485-8 1. Hispanic Health and Nutrition Examination Survey (U.S.) 2. Health surveys — United States — Statistical methods. 3. Hispanic Americans — Attitudes — Measurement. 4. Sampling (Statistics) I. Forthofer, Ron N., 1944— Il. National Center for Health Statistics (U.S.) Ill. Title. IV. Title: Nonresponse bias. V. Series. VI. Series: DHHS publication ; no. (PHS) 93-1393. RA409.U45 no. 119 [RA408.5] 362.1'0723 s—dc20 93-29686 [362.1'08968073) CIP For sale by the U.S. Government Printing Office Superintendent of Documents Mail Stop: SSOP Washington, DC 20402-9328 Vital and Health statistics Investigation of Nonresponse Bias: Hispanic Health and "ee musa Nutrition Examination MAY 17 1994 SU vey UNIVERSITY Ur 1p, wenacLeY Series 2: Data Evaluation and Methods Research " No. 119 This report presents an investigation of potential nonresponse bias in the Hispanic Health and Nutrition Examination Survey (HHANES) conducted during the period 1982-84. Data from a household and medical history interview were used to investigate factors related to examination status. The study includes a comparison of data for examinees in HHANES with data from interviewees in the National Health Interview Survey during 1982, 1983, and 1984. EE RE iii, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics Hyattsville, Maryland December 1993 U.6. DEPOSITORY DHHS Publication No. (PHS) 94-1393 FEB 08 1994 National Center for Health Statistics Manning Feinleib, M.D., Dr. P.H., Director Jack R. Anderson, Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D, 3ssociate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology David L. Larson, Assistant Director, Atlanta Division of Health Examination Statistics Robert S. Murphy! Director Kurt R. Maurer, Ph.D., Deputy Director Vicki L. Burt, Sc.M., R.N., Chief, Survey Planning and Development Branch Katherine M. Flegal, Ph.D., Chief, Medical Statistics Branch Anne C. Looker, Ph.D., R.D., Chief, Nutrition Statistics Branch Robert Krasowski, M.S., M.A., Chief, Computer Systems and Programming Branch Christopher T. Sempos, Ph.D., Chief, Longitudial Statistics Branch Jean S. Findlay, M.S., Chief, Survey Operations Branch Ronette Briefel, Dr. P.H., R.D., Coordinator for Nutrition Monitoring and Related Research Clifford Johnson, M.S.P.H., Special Assistant for Analysis and Information Management Contents ITU COAUCLION. ovis 8 ed Bt 0 Hi hn ec st RTH HR HI Rom sh 98S BRI RR Hohhot: 45 EE 005 00 1 Sources of dain and analytCaAl I8SIIEE: «ous 5s 5.¢ sims mais $0 aEm Em wn FR SAEs BOW 3g PE HER TERE SRE Sie 50S HE EE REE 4 SOUTCES uc. crvsins me0 us ar 00s as beds es insiss sews 9 50 1 § AFUE EIS Wmomsiasis mann iris HEI RTE RIFTS AFD Ab BLS mnacin s we a 600 ME hai 4 AATIBIVEICH] JEBUBS 1 «vv veo mie visio vis mie mi 50000 0 iw i ww 0 0 0 0 8 40 0 dR RR 4A WE Ry se 6 Statistical THCTHOOIOBY . «ovens mam ui wd ms wars msm sms ars 008 808% of 0s 9080 808 008 56 000 6 HRT 6 Oo 4 wi 0 6 4 0 8 08 7 SIRE TOSIING ors mm 553 5 aT Hn 0 i £m B18 Ei BEET B13. 94 0 30 4 8 40 3 BF 0 4 annie nian 8 BINGINGS : c5in vim womens 36 005 ER E6 FEA WERE 5 0 3 3 5 18 RRA TRH 6 58 350 20 ¥ 40 MEH 8 BEES Ea 90k 32 0 4 20 4 1 900% B38 BiH 90 6 9 MEXICAN AXTTICEECAIIN 5 ic 3:5 5: 0.010 0 wot i mom ms 50.18 500808 nh 03 0s rcs a 53 500 5 10 900 0 SER $05.00 0 no + 9 CUBBIE, or 00000008 55 0 00 tm I £0 IEE 131 0, 0 SoA SH 6 D5 0 HE 7 0 1000 0 To 5 0 5 0.00 4 11 PUCTIO BICANE. « vs oi viv wns wma G03 60 00 500 6 Bon ADE EE RRB FE A SE 4 HERRERA RRS a a Baa 12 DISCUSBION 2 nmanmmme v5: tor moras nevis wane mom iia ors on Bie gus aponon wen aig mt GR BEY 0% db Be a wo pa BE A EE TRA on way os 14 Areas Of POTENUAL DIAS «sce vv mse ne ma mne ss mine mem si oie #0 5% 08 REESE F 0% MRmE® 80 EE 56 BES HEBER E Wrd ns Hg WF BER SHE 14 BIAS CITT. 2 65: 2 8 12505 5 es om vm iow #9000 0 5 0 0, 2 305 0 mc wis i 0 0 0 00 9 onc se ai is tr 0 200 50 8 ee 0 15 Limitations and implications of the methodology. ........ c.count ei ee 16 RETSUCIIERS vv 10 27 505 303 i 99 0 91 v3 500 HE 008 TEAR 5 0 et 0 eo 0 2 TB 0 Ss 450 30. 0 1 0 1 8 0 00 J i som 18 Ti8E Of detailed CADIEE: i ory ics win suns sommaman mists on B00 or £180 0 + 20 rev oe 1 008 10 EH BIE 950 000 Br irw 3 or on 0 HE 420 50 02 0 90 ww 20 Appendixes L SEAISICAL TOES pvnrsins 310 prawns raves suite £708 Boh 0hp dase srnn puny org 90 HELE hie Wp os i po 45 8 ALERT? 08 BR iol is 48 IL. National OriZirCCO0E . . « vis vivvms ws sass wnis isn wis oi 06 205 015 0B 5S 6 85% 475 0 618 81 20 018 91 9 HEE 313 FRAA 24 54 ve RDB BEE 54 III. Definitions of demographic and SOCIOECONOMIC tEIMS . ....vvvutuuttettereeereeeeeeenneenn. £5 4 oe ee 56 IV. Items on the child sample person questionnaire used in nonresponse analyses. ..........coeeeveeeeeeenn.... 58 V. Items on the adult sample person questionnaire used in nonresponse analyses. ...........c.vveeuuneennn.nnn 61 : VI. HHANES-NHIS COMPATISON 5.5.50 0005 vn bon @ 800 0500 0005 50 Ahr0.0 050 00 SEH 4540141 400 0 050 0045040818550 985 8 di 68 Vil. CHAID PIOCEAUIR rw ennovmrn sents ad af ssn sess wi en me sin Sa Gossip ais 24 5s sins 5 8 AS BRan ATE Saas ny 70 VIII. Adjusting for pOsSiDIC NONTESPONSE DIRS 4: ov ws sw woes win wm 5 30810 90 0 ori ih #1200, 30 990 6 00030 950 B43 009-0 ie wi B19. meh 30 73 Text tables A. Health examination surveys conducted by the National Center for Health Statistics, by years of survey and ages of persons examined, 1960-80 ...........iiiiiitiii 1 B. Sample size and response rates for Hispanic persons 6 months-74 years of age, by survey area and specified Hispanic origin: Hispanic Health and Nutrition Examination Survey, 1982-84 ............. ccc. 5 C. Sample size and response rates for Hispanic persons 20-74 years of age in the fasting sample, by survey area and specified Hispanic origin: Hispanic Health and Nutrition Examination Survey, 1982-84............... 5 D: Study variable QlINIHONS i. «.. vvco ws isms wis oo 0005500 5 910 300 90 wow ame 0 gion ara of £9958 506 Boa are ae or iu avs 32 0p 98 30k 00 41 bw wie wine 6 E. Variables used in weighting adjustment for interview nonresponse, examination nonresponse, noncoverage, and poststratification, according to Hispanic subpopulation: Hispanic Health and Nutrition Examination SUIVEY, 1082-84 i vaivnuminmsinmn sso msmsio kes sis ase sis sams ba amass sans ssmssntnsainsssssietnsssiviesss 7 F. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination: Mexican American HHANES, 1082-84... oii i eee eee 15 G. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination: Cuban American HHANES, 1082-84. . . oii eee et etter 16 H. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination; Puerto Rican HHANES, 1082-84. .... 5551s 55 50% 5005 506 5 515 05. 0 dk 55 5k 20k #5 508 5 500 415 5 68 204 3: win mrs mace» 17 Figures 1. Examination response rates by age groups for surveys conducted by the National Center for Health Statistics FONT JOGO EhYOURN TOBA 2 2 5.000050 5/0 0.015 Sims riers x wien, S10 re ls TR ERB el 18 rm sm ert wo fs A Bal 9.0 i 2 2. Distribution of sample persons according to respondent status in the Mexican American subsample of te TI ANTES oo rn 0000 oie 200000 0 10 0k BURT $0040 0 45 Hn 0 56 F500 00 90 0 0 4 0 0 90 0 0 0 00 9 3. Distribution of sample persons according to respondent status in the Cuban subsample of the HHANES ....... 11 4. Distribution of sample persons according to respondent status in the Puerto Rican subsample of tHe HIE ANTS vs vrs vs vss 0300 503503 830003080 1 £07 2 eva se sf or i 04 0 900 50 00 0 0 2 6 are #0 EC R000 0 on 0 £0 12 ». ~- Symbols --- Data not available . Category not applicable - Quantity zero 0.0 Quantity more than zero but less than 0.05 2 Quantity more than zero but less than 500 where numbers are rounded to thousands * Figure does not meet standard of reliability or precision Investigation of Nonresponse Bias: Hispanic Health and Nutrition Examination Survey by Michael L. Rowland, Division of Health Examination Statistics, National Center for Health Statistics, and Ronald N. Forthofer, Ph.D., Biostatistician, Boulder, Colorado Introduction From 1960 through 1980 the National Center for Health Statistics (NCHS) conducted five population-based National Health Examination Surveys (table A). As with all surveys, the representativeness of the sample to the target population has been a primary concern. The first of a series of NCHS publications based on these surveys focused on an evaluation of “the similarity between the sample and universe it represents and the impact of nonresponse (1).” This concern has been echoed in a number of NCHS studies since then (2-4) and is the topic of the present paper —an evaluation of response status for Hispanics selected for examination in the Hispanic Health and Nutrition Examination Survey (HHANES), con- ducted from July 1982 through December 1984. As shown in figure 1 and table 1, the relatively high examination response rates of the first, second, and third National Health Examination Surveys of adults, children, and youths, respectively, conducted during the 1960’s have been followed by lower examination response rates in the first and second National Health and Nutrition Examina- tion Surveys (NHANES I and NHANES II) in the 1970’s and HHANES in the early 1980's. In a health examination survey, as well as any survey involving volunteer participation, the survey meets one of its severe problems after the sample is identified and the sample persons are requested to participate in the exami- nation. A sizable number of sample persons who initially are willing to complete the household information, and possibly some of the medical history questionnaires (which are done in the household), usually will not participate in the examination. Full participation by individuals is deter- mined by many factors, some of them uncontrollable by either the sample person or the survey personnel. For example, family health beliefs and practices, employment status, and access to transportation could affect participa- tion in the survey. Because nonresponse is a potential source of bias, intensive efforts were made in HHANES to develop and to implement procedures and inducements to reduce the number of nonrespondents and thereby reduce the poten- tial of bias due to nonresponse. Among these were remu- neration (that is, sample persons were given $20.00 after receiving the examination, as well as either taxi fare or milage costs of driving to and from the examination center), community outreach programs (a HHANES pub- lic affairs task force designed, developed, implemented, and coordinated a public affairs initiative, which was an integral part of the survey operations), Spanish-language translated questionnaires, and bilingual and/or bicultural household interviewers. These procedres are discussed in a Vital and Health Statistics series report (5). Despite response rates of 87, 79, and 89 percent at the household interview stage for Mexican American, Cuban, and Puerto Rican subsamples of the HHANES and inten- sive efforts of persuasion, only 76, 61, and 75 percent of sample persons for these groups, respectively, were exam- ined. Consequently, the potential for a sizable bias exists in the estimates from these subsamples. Using data from HHANES and from the National Health Interview Survey (NHIS), efforts have been made to examine possible demographic and health-related dif- ferences between examined and nonexamined persons. In addition to nonresponse to the examination (unit nonre- sponse), nonresponse to a particular examination compo- nent (component nonresponse) and nonresponse to particular items within an examination component (item nonresponse) are treated here in an estimation of poten- tial nonresponse bias. For the analyst who must evaluate nonresponse bias, both the results of the exploratory analysis presented here for the HHANES and the analytic approach used here involving questionnaire data internal to the HHANES survey and external to the HHANES data (comparable National Health Interview Survey questionnaire data) will be of interest. The methodologies used here find their Table A. Health examination surveys conducted by the National Center for Health Statistics, by years of survey and ages of persons examined, 1960-80 Survey Date Ages First National Health Examination Survey INHES I) sviv os wan ns nvmes wrasse 1960-62 18-79 years Second National Health Examination Survey EES HY. co sn nvaes punnis fh tas nam 1963-65 6-11 years Third National Health Examination Survey NHES M2 sss ss sv ems dn smems sme 1966-70 12-17 years First National Health and Nutrition Examination Survey (NHANES I) . ....... 1971-74 1-74 years Second National Health and Nutrition 6 months— Examination Survey (NHANES I)... ..... 1976-80 74 years 100 80 p= = NM | a a Percent of responses 8883 7 x IN N AN NR NN 77 7 30 p= 20 = 10 ie 0 NN NN i N NN NN NHANESI NHANES HHANES, HHANES, HHANES, NHES NHANES I NHANES HHANES, HHANES, HHANES, I Mexican Cuban Pureto I I Mexican Cuban Puerto American Rican American Rican Age group 6 months to 5 years Age group 6-11 years 100 p= - 90 p= lL 80 - A AO = NN i *E N N N\ re 55 NESE 60 %, - 5 40 = NN bee e L 10 p= — ° FN NAY RN N\ NN NN NN N NN NN I I Mexican Cuban Age group 12-17 years A NHES NHANESI NHANES HHANES, HHANES, HHANES, to American Rican NOTES: NHES ] is the first National Health Examination Survey. NHES II is the second National Health Examination Survey. NHESIII is the third National Health Examination Survey. NHANES I is the first National Health and Nutrition Examination Survey. NHANESTI is the second National Health and Nutrition Examination Survey. HHANES, Mexican American is the Hispanic Health and Nutrition Examination Survey, Mexican American portion. HHANES, Cuban is the Hispanic Health and Nutrition Examination Survey, Cuban portion. HHANES, Puerto Rican is the Hispanic Health and Nutrition Examination Survey, Puerto Rican portion. NHESI NHANESI NHANES HHANES, HHANES, HHANES, I Mexican Cuban Puerto American American Age group 18-74 years Figure 1. Examination response rates by age groups for surveys conducted by the National Center for Health Statistics from 1960 through 1984 antecedents in previous nonresponse studies, and the findings presented here can be placed in the context of those studies. More information on methodologies used in nonre- sponse adjustment have been published previously (6). As with any methodology, the assumptions made are critical to an evaluation of the results. All methods dealing with nonresponse adjustment, including statistical weight- ing, imputation, and probability approaches, find it neces- sary at some stage to make an assumption about the similarity of respondents and nonrespondents. For exam- ple, a common practice is to employ interview question- naire data in evaluating examination nonresponse (5). The assumption is made that any residual reporting bias of demographic and health history data is similar for respon- dents and nonrespondents. Where there is evidence to the contrary, differential reporting bias could confound a nonresponse evaluation such as the present one. The interpretation of nonresponse bias analyses must also be made in the context of other issues such as measurement error and other methodologic biases. For 2 example, in a survey like HHANES, the results from the physical examination measurements may not coincide with the results from the interview. This could be the result of bias in either of the components or both. Careful review and interpretation of both the statistical and methodolog- ical (physiologic or substantive) issues are as important in the analysis of nonresponse bias as they are in the basic descriptive or multivariate analysis. Reviews of nonresponse studies were conducted in Cycle I of the Health Examination Survey, 1960-62, and in the first National Health Nutrition Examination Survey, "1971-74, by Landis et al. (7). They summarize, During the early stages of NHANES I, when it became apparent that the response rate for the examinations was lower than in the preceding health examination surveys, a study of the effect of remuneration upon response in NHANES I was undertaken. The findings, published by NCHS (4), included remuneration as a routine procedure in NHANES I starting with the 21st and 22d examination locations. Using data from NHANES I and from an earlier survey, efforts have been made to examine possible health-related differences between examined and non- examined persons. An investigation of reasons for par- ticipation and nonparticipation in NHANES I was conducted by interviewing a sample of 406 people composed of 290 examined persons, 35 persons who had made appointments for the examination but who never came to the mobile examination center for the examination, and 81 persons who refused to participate in the survey (8)...They were asked to indicate why they did not choose to be examined in NHANES I. The primary reasons given were that they had no need for a physical examination (48 percent), or that the examina- tion times were inconvenient because of work schedules or other demands (15 percent). Only 6 percent of those persons who were not examined indicated that they refused the examination because of sickness, and 3 per- cent based their refusal on a fear of possible findings. Data on both examined and nonexamined (but interviewed) persons were analyzed by using infor- mation from the first 35 survey locations of NHANES I (9). For the health characteristics compared, the two groups were quite similar. For example, 20 per- cent of the examined people reported that a doctor had told them they had arthritis compared with 17 percent of the unexamined people. Similarly, 18 percent of both the examined and nonexamined persons had been told by a doctor that they had high blood pressure. Twelve percent of both groups reported that they were on a special diet and 6 percent of both groups said that they regularly used medication for nerves. In another study of factors relating to response in Cycle I of the Health Examination Survey, 36 percent of the nonexamined people viewed themselves as being in excellent health compared with 31 percent of the exam- ined people (2). A self-appraisal of poor health was made by 5 percent of the nonexamined persons and by 6 percent of those who were examined. In a different study of Cycle I findings, those who participated in the survey with no persuasion and those who participated only after a great deal of persuasion generally had few differences for numerous selected examination and questionnaire items (10). Forthofer evaluated nonresponse in NHANES II (11). This study used the Automatic Interaction Detection (AID) procedure (12) for identification of variables asso- ciated with nonresponse. Another analysis included a comparison with estimates from the NHIS. This study also included a review of previous health examination surveys and the factors associated with nonresponse in those surveys as well as in the NHANES II. Forthofer found that the factor most highly associated With examination status is whether people had problems that they wished to discuss with a physician. In his survey of the nonresponse literature, Forthofer found that response rates were high- est for those subjects reporting a health care need or condition (1,2,13-15). Another report provides an overview of nonresponse bias in NHANES II and, to a lesser extent, HHANES (16). This report included an evaluation of techniques for reducing item nonresponse bias. In addition, some prelim- inary investigations of nonresponse in HHANES have appeared in the literature (17-19). Sources of data and analytical issues Sources The Hispanic Health and Nutrition Examination Sur- vey (HHANES), conducted from July 1982 through De- cember 1984, is one of a series of health examination surveys conducted by NCHS. The major difference be- tween HHANES and other health examination surveys is that HHANES was a gurvey of three special subgroups of the population in selected areas of the United States rather than a national probability sample. The target population for HHANES ideally would have included all households with at least one member of Hispanic origin. However, the United States includes States and counties with very small numbers or proportions of Hispanic per- sons. Therefore, HHANES was restricted to those coun- ties in the three target areas of the country that had a sufficient number or proportion of Hispanic persons to permit the efficient operation of the survey. Thus, 97 per- cent of the 1980 Mexican-American population in the five Southwest States, 96 percent of the Cuban population in the Dade County, Florida, area, and 90 percent of the Puerto Rican population in the New York City area were eligible for inclusion in HHANES. Although HHANES was not designed to be representative of all Hispanics residing in the United States, the survey universe included approximately 76 percent of the 1980 Hispanic- origin population in the United States. The three Hispanic subgroups and the areas covered were: Mexican Ameri- cans residing in five southwestern States (Arizona, Califor- nia, Colorado, New Mexico, and Texas); Cuban Americans residing in Dade County, Florida; and Puerto Ricans residing in the New York City metropolitan area, includ- ing parts of New York, New Jersey, and Connecticut. Selected households were screened to identify eligible Hispanic families and to select sample persons from these families to be interviewed and examined. Eligibility for the survey was determined by the family unit. A family was considered eligible if at least one family member’s re- ported national origin or ancestry met the criteria for eligibility appropriate to the survey location. These crite- ria were as follows: Survey area National origin or ancestry Mexican or Mexicano, Mexican American, Chicano, Hispano, Spanish American or Spanish (when no other country of origin was mentioned) Southwest area . . .. Dade County, Fla, areca ....... Cuban or Cuban American New York City CH sans urmnsas Puerto Rican or Boricuan In cases where multiple origins were reported for the same individual on different questionnaires, the person was considered eligible if any one of the reported origins met these criteria. If a family were eligible for the survey, all members of that family were eligible to be selected for the interview and examination components. Therefore, some non- Hispanic persons residing in Hispanic households and some Hispanic persons not meeting the above criteria were selected and examined in each of the three geo- graphic areas. For this report, however, all findings are based on the examined persons within the households who were defined as being of Mexican origin or ancestry in the Southwest, of Cuban origin or ancestry in Dade County, Florida, and of Puerto Rican origin or ancestry in the New York City area. This report, therefore, excludes persons in the total sample who were non-Hispanic or of an origin that did not meet the eligibility criteria. Appendix II presents a more detailed description of how the Hispanic- origin recode used for this report was determined. Tables B and C show the sample sizes and re- sponse rates for each of the three survey areas in HHANES. In table B, the results are presented for both the total sample (including non-Hispanic per- sons) and for the specific-origin sample. Table C shows the sample sizes and response rates for Hispanic adults in the fasting sample. The fasting sample consisted of a randomly selected half sample of the examined adults ages 20-74 years. This “morning half sample” or “fasting half sample” was also designed to be representative of Hispan- ics in the designated areas. Persons in the half sample were asked to fast overnight for 10-16 hours and were examined in the morning session. No fasting instructions were given to those in the afternoon half sample or nonfasting half sample. The focus of this paper is on the full sample and the fasting half sample since these data sets provide the basis for the majority of analytic studies completed for the HHANES. HHANES, like previous examination surveys, con- sisted of two major components. Household interviews formed the first component; the second consisted of phys- ical examinations and additional interviews in examination centers. All interviews, examinations, tests, procedures, and laboratory determinations were performed following standardized protocols. Household interviews The household interview component involved collect- ing socioeconomic and demographic information from the family and sample persons within the family and complet- ing a medical history questionnaire for sample persons. Interviewers employed by the contract agency conducting the HHANES performed the initial household interviews and aided in the scheduling of appointments for examina- tion. This information was obtained prior to the examina- tion and was usually obtained from the sample person, or, when necessary, from a knowledgeable household member or a neighbor. Child and adult medical history interviews were also conducted in the household. Persons at least 18 years old responded for themselves unless they were physically or mentally unable to be interviewed. For sample persons 12-17 years of age, either self- or proxy-response was accepted. For sample persons under the age of 12, proxy respondents were required, except for a few questions addressed directly to children 6-11 years of age. An examination appointment was also made at the time of interview. In both the household interview and the examination, sample persons were given the choice of participating in either English or Spanish. Interviewers were bilingual and Spanish language questionnaires were available. Examination The examination component was performed in mobile examination centers specially designed for this study. The examination environment and equipment were standard- ized to minimize differences in findings among sample locations. The full-time examination teams were specifi- cally trained to follow the study protocols, which provided for standardization, quality control, and evaluation of team members’ performance. The examination consisted of a series of standardized tests and procedures that included the following: ® General medical examination and screening by a phy- sician, including additional medical history information ® Body measurements ® Dietary interview ® Selected diagnostic tests such as electrocardiograms, x rays, hearing, and diagnostic ultrasound for detec- tion of gallstones ® [Laboratory tests on whole blood, serum, and urine specimens Thus, HHANES provided the opportunity to assess key aspects of the Hispanic population’s health and nutri- tional status during a 2 1/2-year period and to collect baseline data that could be used to assess changes over time in selected Hispanic subgroups living in the United States. Table B. Sample size and response rates for Hispanic persons 6 months-74 years of age, by survey area and specified Hispanic origin: Hispanic Health and Nutrition Examination Survey, 1982-84 Interviewed Examined Survey area and Sample Se Hispanic origin size Number Percent Number Percent Southwest area Allpersons . . .......... 9,804 8,554 86.5 7,462 75.4 Mexican American. . . ..... 9,455 8,222 87.0 7,197 76.1 Dade County, Florida, area AIDErSonS . . + vnsws «ns 2,244 1,766 78.7 1,357 60.5 CUBAA «evi tw stra Ros 0 fk 2,125 1,877 78.9 1,291 60.8 New York City area AI PeISONS « « « + vv uns «vs 3,786 3,369 89.0 2,834 74.9 Puerto Rican. «=» + + «+ wi wie 3525 3,137 89.0 2,645 75.0 NOTE: See appendix Il for the definition of Hispanic origin. Table C. Sample size and response rates, for Hispanic persons 20-74 years of age in the fasting sample¥, by survey area and specified Hispanic origin: Hispanic Health and Nutrition Examination Survey, 1982-84 Interviewed Examined Survey area and Sample —mM@M™—MMM™M™M™M™ —— Hispanic origin size Number Percent Number Percent Southwest area Mexican American. . . ..... 2,360 1,969 83.4 1,655 70.1 Dade County, Florida, area COBB: 2 5 sc vrs em dw ns 741 565 76.2 426 57.5 New York City area Puerto RICAN. « +» «ov + x 5s 881 751 85.2 596 67.7 More detailed information on selected tests and pro- cedures referred to in this report are given as follows: Blood pressure—Two blood pressure measurements were taken on one occasion in the mobile examination center as part of a physician’s examination. Both measure- ments were taken with the patient seated, 5 minutes into the examination and 5 minutes apart. The average of the two readings was used for the estimates presented here. Systolic (first phase) and diastolic (fifth phase) blood pressure were measured to the nearest even digit using a standard mercury sphygmomanometer. Ultrasonography of the gallbladder — Real-time ultra- sonography of the gallbladder was performed by health technicians using an instrument with a 3-MHz rotary mechanical sector scanning transducer. Examinations were conducted with sample persons in both supine and left decubitus positions. A diagnosis of gallstones was made by commonly used criteria of echoes within the gallbladder with shadowing or movement of echoes. If a right upper quadrant or epigastric scar was observed and the gallblad- der was not seen, it was concluded that a cholecystectomy had been performed. Ultrasonography was done on the fasting half sample described previously in the text. Iron (Fe) status based on biochemical data — Impaired Fe status is used in combination with low hemoglobin as 5 an indicator of anemia. Impaired Fe status was calculated using the MCV model (20), which was developed by an expert panel for use with Health and Nutrition Examina- tion Survey (HANES) data. Measures are based on the results of venipuncture blood drawn from subjects at the time of the examination. Pregnant women were excluded from analyses because pregnancy affects the interpretation of Fe status indicators. Also, persons who lacked values for any of the Fe status indicators were excluded from the analytic sample. Total serum cholesterol — Serum total cholesterol was measured in venous blood specimens and corrected to the Abell-Kendall reference values (21). Height and weight — Technicians measured several an- thropometric dimensions, including standing height and weight. Body mass index (see definition in table D) was used as a measure of overweight. Definitions + The cutpoints and variables used to define the condi- tions referred to in this report were obtained from previ- ously published studies (20-24) based on the HHANES and are given in table D. Demographic and socioeconomic terms used in this paper are defined in appendix III. Items on the child and adult sample person questionnaires used in the nonresponse analysis are given in appendices IV and V, respectively. Analytical issues Survey design The Mexican American, Cuban, and Puerto Rican portions of the HHANES were each designed to be complex, multistage, stratified, probability cluster samples Table D. Study variable definitions of persons 6 months-74 years of age. There was oversam- pling of eligible Hispanics 6 months-19 years of age and 45-74 years of age. For more detail see appendix I. Statistical weighting To take into account oversampling and other sample design features, sample weights are provided with the HHANES survey data. Basic weights accounting for the probability of selection and oversampling of selected age groups were further adjusted for other factors related to nonresponse and noncoverage (table E). For Mexican Americans, weights were further ad- justed as follows: ® adjustments for interview nonresponse within catego- ries of age, income, household size, and geographical location ® adjustments for examination nonresponse within cate- gories of age, household size, and location ® adjustment for noncoverage within primary sampling units (PSU’s) according to family income group; and ® poststratification ratio adjustments by age and sex made to produce the final sample estimates of the population that correspond to the 1983 Bureau of the Census estimates of the civilian noninstitutionalized target population of Mexican Americans in the South- west (17) For Cubans, weights were further adjusted as follows: ® adjustments for interview nonresponse within catego- ries of age, gender, and income ® adjustments for examination nonresponse within cate- gories of age, gender, and household size ® adjustment for noncoverage within PSU’s (25) Variables Definitions Hypertension (based on examination). . . ........ medication. Hypertension (based on interview). . . .......... Defined as average of two blood pressure measurements = 140/90 mmHg or currently taking antihypertensive Defined as those subjects who reported in the medical interview that a doctor had told them that they had high blood pressure or hypertension. Gallstone disease (based on examination) . . . . .... Gallstones (based on interview). . . ............ gallstones. Impaired Fe status (based on examination) ....... Defined as subjects having gallstones or evidence of a previous cholecystectomy upon ultrasonography. Defined as those subjects who reported in the medical history interview a doctor had told them they had Defined as subjects with at least two of three Fe status indicators, namely, mean corpuscular volume <80fL, erythrocyte protoporphyrin >1245nmol/L RSC, transferrin saturation <16%. Pregnant women excluded from analysis. Anemia (based on interview) ................ had anemia. Defined as those subjects who reported in the medical history interview that a doctor had told them they Pregnant women excluded from analysis. Elevated cholesterol (based on examination) . . .... Defined as those subjects with a serum cholesterol level of 240 mg/dl or more. Defined as a body mass index (BMI) (weight in kilograms divided by height in meters squared) equal to or greater than that at the 85th percentile of men or women aged 20-29 years from the NHANES Il, 1976-80. Men are categorized as “overweight” when their BMI equals or exceeds 27.8. For women, the cutoff point is 27.3. Pregnant women excluded from analysis. Overweight (based on examination) . . . ......... Table E. Variables used in weighting adjustment for interview nonresponse, examination nonresponse, noncoverage, and poststratification, according to Hispanic subpopulation: Health and Nutrition Examination Survey, 1982-84 Interview (nonresponse) Examination (nonresponse) Noncoverage Poststratification Variable MA Cc PR MA PR MA Cc PR MA Cc PR BBE... sn vem 2% ms Hh BY X X BOB 56 555 00 3 5 wan ws X X X X X Household size. . ....... X X X X INCOMIBsrs mis vin vi mee X X X X X Primary sampling unit. . . . . X X X X X NOTES: MA =Mexican American, C= Cuban, PR= Puerto Rican. Poststratification was not done for the Cuban or Puerto Rican survey portions due to lack of adequate population estimates. For Puerto Ricans, weights were further adjusted as follows: ® adjustments for interview nonresponse within catego- ries of age, household size, and income ® adjustments for examination nonresponse within cate- gories of gender, age, and household ® adjustment for noncoverage within PSU’s according to family income (26) Statistical methodology The investigation of the potential nonresponse bias in the HHANES consisted of four parts. Part 1: Interview status — This investigation was limited to variables that the interviewer could obtain during the screener interview from the sample person, an adult household member, or a neighbor and to seasonal and geographic location information. Variables used in this part of the study were age, season of the year, gender, family size, language of the screener interview, and mobile examination center location (table 2). Because sample persons within a family tended to have the same interview status, the family was also used as the unit of analysis. The demographic data included in the family nonresponse analysis are those of the head of the family. The Chi-Square Automatic Interaction Detection (CHAID) (27) technique was used to summarize the data. CHAID is a descriptive procedure that provides the re- searcher with information about the relationships between the dependent variable (the interview status) and the predictor variables (other classification or descriptive vari- ables) by calculating the chi-square measure of association between the dependent and each independent variable. The predictor variable that has the most significant chi- square, after a Bonferroni adjustment (28) for the number of variable categories, is used to split the sample into groups. This process is repeated for each of the new groups until there are too few observations for further splitting. The result is a tree-like structure that suggests which predictor variables may be important and need future investigation (see appendix VII). Part 2: Examination status — In this stage of the analy- sis, the interview-weighted examination response rate was studied in relation to demographic and screener variables from the screener and family questionnaires and medical history variables from the child and adult medical history questionnaires (table 2). The CHAID technique was used to summarize the data in two steps. In the first step, the same set of variables used in the CHAID analysis of interview nonre- sponse was screened to identify relationships between examination status (the dependent variable) and a series of other variables commonly used in nonresponse weight- ing adjustments. In the second step, additional variables, selected from the family questionnaire and the child and adult medical history questionnaires (table 2), were in- cluded in the analysis as an aid to researchers in identify- ing potential sources of bias in analyses. Part 3: Comparison between HHANES and 1982-84 NHIS — This part of the study compared the HHANES with the combined 1982, 1983, and 1984 NHIS for the Mexican American, Cuban, and Puerto Rican subpopulations. The comparison consists of a display of the survey weighted proportion of various conditions or attributes for each sample. See appendix VI for more details on the HHANES-NHIS comparison. Part 4: Estimating possible bias in disease prevalence —In this, the final stage of the nonresponse bias analysis, an estimate of possible nonresponse bias is given for selected variables (diseases) that have appeared in published studies. For each variable, a bias-adjusted estimate is com- pared with a survey estimate based on the analytic sample. The conditional probability approach used to compute bias-adjusted estimates is described in detail in appendix VIII. Briefly, probabilities of the disease are computed conditional on the level of variables found to be associated with both the respondent status and the disease under study. Socioeconomic status was one of several variables from the medical history questionnaire found to be asso- ciated with respondent status in the HHANES and it has also been shown to be related to disease prevalence. Therefore, socioeconomic status, measured as below pov- erty or at or above the poverty level, was selected to be used in the creation of an adjusted prevalence estimate of the various diseases. Differences between the survey esti- mates and the bias-adjusted estimates measure the effect 7 of poverty status (Note, subjects with missing poverty status were excluded from the analysis). Since this method assumes that respondents and non- respondents have similar disease prevalence within each poverty status level, this assumption was examined empir- ically (see appendix VIII). Variable (disease) definitions, cutpoints, and age group- ings have been made comparable to those used in the studies. The dependent variables include hypertension (22), gallstone disease (23), impaired Fe status (20), elevated cholesterol (21), and overweight (24). In both the bias-adjustment analysis using examina- tion data and the empirical analysis using medical history data, estimates were computed using basic weights (the reciprocal of the probability of selection). Using the final nonresponse adjusted weights would have confounded any comparison between survey and adjusted estimates. Thus, because neither the survey weighting nonresponse adjust- ments nor the poststratification ratio adjustments (for Mexican Americans only) to the 1983 Bureau of the Census age-sex marginal distribution were included in the tabulations of survey estimates, and the survey estimates given in the tables will differ from the weighted published estimates. As was stated previously, the difference be- tween the adjusted estimate and the survey estimate is meant to measure the effect of nonresponse related to poverty status and independent of possible confounding of other nonresponse or poststratification weighting adjustments. Significance testing Testing for statistical significance was done at the 95-percent confidence level. The complex survey design used in the HHANES tended to increase the estimated variance of prevalence estimates over that which would have been obtained through simple random sampling (29). Average design effects (the ratio of the complex sample variance to the simple random sample variance) have been calculated for many analytic variables for the three His- panic subgroups. In general, the average design effects for the Mexican American and Puerto Rican subgroups tend to average 1.5, while those for the Cuban subgroup tend to be about 1.0. Thus, in these analyses, the weighted simple random sample standard errors were multiplied by the square root of the design effect. All data analyses were done using SAS procedures (30) or programs accessible through SAS (31). Criteria for presentation of data The following guidelines were used for the reporting of percents for the HHANES data. If the sample size in an analytic cell is less than 45, the percent was not reported. If the sample size is 45 or more, the percent is presented without caveat. Criteria for determination of bias Consistency and a measure of relative bias were the two criteria used in determining whether there was evi- dence of possible bias. Assuming that the adjusted esti- mate is the “true” prevalence, bias is defined as the difference between the survey estimate and the adjusted estimate. Variables that were identified as having the same directional bias across age groups within gender and that had a relative bias of at least 25 percent were consid- ered to have shown evidence of possible bias. Relative bias was defined as bias relative to the standard error of the estimate expressed as a percent bias relative bias = 100 s —mM8M8Mm™ standard error of the estimate Hansen, Hurwitz, and Madow (32) demonstrate that when biases can be shown to be relatively small, say, less than 25 percent of the standard error of an estimate, they can be neglected without any serious effect on the inter- pretation of the results. The estimated standard error (SE) of the survey estimate was computed as follows: SE = \/deff «+ /pelp n where deff is the design effect, p is the survey prevalence estimate, and » is the sample size. Findings Findings are presented separately for Mexican Amer- icans, Cubans, and Puerto Ricans, for each stage of the analysis identifying factors associated with nonresponse, comparison between HHANES and the 1982-84 NHIS, and bias estimation for selected disease conditions. Re- sults from a screening of potential predictors of interview response and examination response in the Mexican Amer- ican, Cuban, and Puerto Rican HHANES are summarized in text tables F, G, and H, respectively. Details of the Mexican American, Cuban, and Puerto Rican analyses are provided in the following paragraphs and in tables 3-17. Results from the bias estimation analysis for selected disease conditions are given in tables 18-41. Mexican Americans Factors associated with nonresponse Figure 2 shows the distribution of the Mexican Amer- ican sample according to respondent status. Seventy-six percent completed the examination as well as all interview components. Of the remaining 24 percent who did not receive an examination, medical history and demographic interview data were collected on 11 percent and demo- graphic information only was collected on 13 percent. Thus, in the following analysis, demographic information is used to evaluate interview response status, and demo- graphic together with medical history data are used to evaluate examination response status. Interview status — As shown in table 3, 87 percent of Mexican-American sample persons completed either the Child Sample Person Questionnaire or the Adult Sample Person Questionnaire. Interview rates differed the most by age, family size, location, and season. There was an inverse association between age and interview rate, rang- ing from 79 percent in the oldest group to 92 percent in the youngest group. For family size, there was a positive association between the number of family members and the interview rate. The response rates in the startup location (San Antonio, Texas) and the California locations were generally lower than response rates for the other locations. The response rate in summer was lower than in the other three seasons of the year. A CHAID analysis was also performed to examine the multiway relation between interview status and the predic- tor variables. The age variable had the largest association es XN \ 7 S 32 o, O Q & 0 QQ D N S ¥ S 0 A , Q * %. & 0 QQ $d Q Q 's, 2 1,025 10.84% Q <3 2 SB RY SB @ SB Ne IR x 5 x XK \¥ Nn -—- © 2 3 3 X 0 RX Q x2 > ™ ) 20 5% 2 QQ Za SO > > 5% o>, 55 QD K OP Y 3 SP o QQ KD >. (2 55 2 Yu y XR SSK CESK Dy RENN bo; 5 5% 0! > > > 52 2 ES 2525 ON OP Q Q 2 * ~ X02 Hi Screened, interviewed, Screened, not not examined interviewed, not examined RX Screened. interviewed. RA Son n ' Figure 2. Distribution of sample persons according to respondent status in the Mexican American subsample of the HHANES with interview status and was selected at the first level of variable selection. At the second level of variable selec- tion, family size, season, and stand location were the most important predictors. Response rates increased with fam- ily size for all age groups. The response rate was lowest for teenagers in the summer months, and response rates varied considerably by stand for the age groups 20-44 and 45-74 years. Subsequent levels of variable selec- tion included the gender and language of the screener interview and interviewer variables. The response rate for females was slightly higher than for males, and rates were higher for Spanish interviews compared with English interviews. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. In order of predictive ability, family size, season, geographic loca- 9 tion, interviewer, and language of screener interview were found to be predictors of nonresponse to the interview. Examination status —Of the interviewed Mexican American sample, 87 percent completed at least one com- ponent of the examination. The distribution of examina- tion response rates for the variables with the largest variation, as defined by the CHAID procedure, as a proportion of the interview sample is shown in tables 4-6 for children, adolescents, and adults, respectively. (Note, all examination response rates were interview weighted; that is, interview sampling weights were applied to all examination respondents and nonrespondents when calcu- lating examination response rates to account for interview nonresponse and noncoverage.) A two-stage CHAID analysis was also performed to examine the multiway relation between examination status and the predictor variables: e Stage 1, screener variables, only—The family size and age variables showed the largest differences in exami- nation response rates among categories. The CHAID analysis for the screener variables showed that family size was the most important predictor of examination status. Response rates ranged from 78 percent for small families (1 to 2 people) to 90 percent for large families (5 or more people). The variables age and gender were found to be important predictors at the second level of selection. The variables location, sea- son, and language of interview were selected at subse- quent levels of selection. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. The candidate demographic predictor variables were those of the family head. In order of predictive ability, family size, completed education of the head of household, poverty level (the poverty level variable was a dichot- omization of those at or above and those below the poverty index cutpoint (see appendix III)), age of the head of household, and location were found to be predictors of nonresponse to the examination. ® Stage 2, all variables — For children 6 months-11 years, the variables location, language parent usually uses at home, education of the head of household, and family size showed the largest differences in examination response rates. For adolescents 12-19 years of age, the family having received food stamps and family size showed the largest differences in response rates. For adults 20-74 years of age, gender, family size, and major activity during the previous 12 months showed the largest differences in response rates. The CHAID analysis showed that geographic location, having received food stamps, and gender were most impor- tant predictors for children, adolescents, and adults, respectively. Comparison between HHANES and 1982-84 NHIS The comparison among Mexican Americans in the HHANES interview and examination weighted samples 10 and the 1982-84 NHIS weighted data are shown in table 7. Distributions of the HHANES examination and interview samples are similar. In general, the distributions for age, sex, and body mass index for adults in all three samples were similar. Differences between the two HHANES (overlapping) subsamples and the NHIS occurred for the variables family income, education of head of household, health status, smoking status, and hypertension. Family income and education of the head of household were higher in the NHIS than in the HHANES. A higher percent of Mexican Americans in the NHIS population considered themselves in excellent or very good health than in the corresponding HHANES population. Fewer Mexican Americans reported being former or current smokers and fewer reported having hypertension in the NHIS compared with HHANES. Although the differences in prevalence for the health conditions shown here for the two surveys were not statistically significant, possibly due to the relatively small sample size in the NHIS, the observed estimates for the NHIS are generally lower than estimates for the HHANES. Possible bias The CHAID analysis and the HHANES-NHIS com- parison for Mexican Americans suggest that respondents and nonrespondents differ with respect to the distribution of socioeconomic status. Since disease prevalence may vary with socioeconomic status, it is important to evaluate potential bias that may have occurred due to differential response. In particular, 28.6 percent of the examined sam- ple were living below poverty compared with 26.6 percent for the nonexamined (but interviewed) sample. For the fasting half sample, 30.2 percent of the examined sample were living below poverty compared with 28.5 percent for the nonexamined (but interviewed) sample. The probabil- ity approach described in appendix VIII is used to esti- mate possible bias due to differential response by poverty level for a number of disease conditions reported in the research literature. As shown in tables 18-25, there is no evidence of nonresponse bias due to the greater response by subjects living below the poverty level for prevalence estimates of overweight, elevated cholesterol, self-reported anemia, self-reported hypertension, gallstone disease, and self- reported gallstone disease. However, prevalence estimates in females for the variables hypertension and impaired Fe status show differ- ences between the survey estimates and the bias-adjusted estimates. For hypertension in females, the survey esti- mates consistently underestimate by 1 percentage point or less. The bias expressed as a percent of the standard error is greater than 25 percent for the age groups 20-44 years and 45-55 years. For impaired Fe status in females, the survey estimates consistently overestimate by 1 percentage point or less. The bias expressed as a percent of the standard error is greater than 25 percent for the age groups 11-19 years, 45-64 years, and 65-74 years. Screened, interviewed, Screened, not not examined interviewed, not examined NX Screened, interviewed, Pe examined Figure 3. Distribution of sample persons according to respondent status in the Cuban subsample of the HHANES Cubans Factors associated with nonresponse Figure 3 shows the distribution of the Cuban sample according to respondent status. Sixty-one percent com- pleted the examination as well as all interview compo- nents. Of the remaining 39 percent who did not receive an examination, medical history and demographic interview data were collected on 18 percent and demographic infor- mation only was collected on 21 percent. Thus, in the following analysis, demographic information is used to evaluate interview response status, and demographic to- gether with medical history data are used to evaluate examination response status. Interview status — As shown in table 8, 79 percent of Cuban sample persons completed either the Child Sample Person Questionnaire or the Adult Sample Person Ques- tionnaire. Interview rates differed the most by interviewer, age, location, language of screener interview, and family size. Response rates among interviewers varied from a low of 67 percent to a high of 90 percent. There was an inverse association between age and interview rate, ranging from 76 percent in the oldest age group to 85 percent in the youngest age group. Rates were higher for Spanish inter- views compared with English interviews. For family size, there was a positive association between the number of family members and the interview rate. A CHALID analysis was also performed to examine the multiway relation between interview status and the predic- tor variables. The interviewer variable had the largest association with interview status and was selected at the first level of variable selection. At the second level of variable selection, age, location, and family size were the most important predictors. The relation of each of these variables with interview status was the same as shown in the univariate case discussed above. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. Family size was the only variable found to be a predictor of nonresponse to the interview. Examination status —Of the interviewed Cuban sam- ple, 77 percent completed at least one component of the examination. The distribution of examination response rates for the variables with the largest variation, as defined by the CHAID procedure, as a proportion of the interview sample is shown in tables 9-11 for children, adolescents, and adults, respectively. (Note, all examination response rates were interview weighted; that is, interview sampling weights were applied to all examination respondents and non- respondents when calculating examination response rates to account for interview nonresponse and noncoverage.) A two-stage CHAID analysis was also performed to examine the multiway relation between examination status and the predictor variables: ® Stage 1, screener variables, only—The total family income and interviewer variables showed the largest differences in examination response rates among cate- gories. The CHAID analysis for the screener variables showed that total family income was the most impor- tant predictor of examination status. The response rate was 80 percent for sample persons from families with an income less than $20,000 compared with 75 percent for those with an income of $20,000 or more. Education of head of household and interviewer were found to be important predictors at the second level of selection. Response was inversely associated with educational level. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. The candidate demographic predictor variables were those of the family head. In order of predictive ability, family size and season were found to be predictors of nonre- sponse to the examination. Family size was positively associated with response. Response was higher in the winter compared with the spring. ® Stage 2, all variables — For children 6 months-11 years, the variables age, geographic location, SMSA, central city/SMSA, not central city, and education of head of household showed the largest differences in examina- tion response rates. For adolescents 12-19 years of age, education of head of household, language of Adult Sample Person Questionnaire interview, self- perceived condition of teeth, and poverty status showed the largest differences in response rates. For adults 11 20-74 years of age, poverty status, self-perceived health status, self-report of ever having had anemia, self- perceived condition of teeth, geographic location, and having had trouble seeing showed the largest differ- ences in response rates. The CHAID analysis showed that age, education of head of household, and poverty status were the most important predictors for children, adolescents, and adults, respectively. Comparison between HHANES and 1982-84 NHIS The comparison among Cubans in the HHANES interview and examination weighted samples and the 1982-84 NHIS weighted data are shown in table 12. Dis- tributions of the HHANES examination and interview samples are similar. In general, the differences in the distributions for age, sex, and body mass index can be attributed to sampling variability. Differences between the two HHANES data sets and the NHIS occurred for the variables family income, education of head of household, and self-perceived health status. Family income was higher in the HHANES than in the NHIS. Conversely, education of the head of household was higher in the NHIS than in the HHANES. A higher percent of Cubans in the NHIS population considered themselves in excellent or very good health than in the corresponding HHANES population. Possible bias The CHAID analysis and the HHANES-NHIS com- parison for Cubans suggest that respondents and nonre- spondents differ with respect to the distribution of socioeconomic status. Since disease prevalence may vary with socioeconomic status, it is important to evaluate potential bias that may have occurred due to differential response. In particular, 20.4 percent of the examined sam- ple were living below poverty compared with 14.3 percent of the nonexamined (but interviewed) sample. For the fasting half sample, 22.3 percent of the examined sample were living below poverty compared with 16.2 percent of the nonexamined (but interviewed) sample. The probabil- ity approach described in appendix VIII is used to esti- mate possible bias due to differential response by poverty level for a number of conditions reported in the research literature. As shown in tables 26-33, there is no evidence of nonresponse bias due to the greater response by subjects living below poverty level for prevalence estimates of elevated cholesterol, impaired Fe status, hypertension, self-reported hypertension, gallstone disease, and self- reported gallstone disease. However, prevalence estimates for overweight and self-reported anemia (females only) show differences be- tween the survey estimates and the bias-adjusted esti- mates. For overweight in females, the survey estimates underestimate by no more than 5 percentage points. In males the survey estimates overestimate no more than 4 percentage points, although only the estimate for females 12 is statistically reliable. The bias expressed as a percent of the standard error is greater than 25 percent for the age groups 20-44 years, 55-64 years, and 65-74 years. For self-reported anemia in females, the survey estimates overestimate by no more than 7 percentage points. The bias expressed as a percent of the standard error is greater than 25 percent for the age groups 15-19 through 65-74 years. Puerto Ricans Factors associated with nonresponse Figure 4 shows the distribution of the Puerto Rican sample according to respondent status. Seventy-five per- cent completed the examination as well as all interview components. Of the remaining 25 percent who did not receive an examination, medical history and demographic interview data were collected on 14 percent and demo- graphic information only was collected on 11 percent. Thus, in the following analysis, demographic information is used to evaluate interview response status, and demo- graphic together with medical history data are used to evaluate examination response status. Interview status — As shown in table 13, 89 percent of Puerto Rican sample persons completed either the Child Sample Person Questionnaire or the Adult Sample Person Questionnaire. Interview rates differed the most by inter- viewer, geographic location, age, family size, and language 492 13.96% 388 1 01% 2,645 75.04% Screened, interviewed, Screened, not not examined interviewed, not examined BS BR Scruaned. interviewed, Figure 4. Distribution of sample persons according to respondent status in the Puerto Rican subsample of the HHANES of the screener interview. Response rates ranged from 67 percent to 96 percent among interviewers. There was an inverse association between age and interview rate, ranging from 87 percent in the oldest group to 92 percent in the youngest group. Response rates ranged from 79 per- cent to 96 percent among geographic locations. For family size, there was a positive association between the number of family members and the interview rate ranging from 83 percent for 1-2 member families to 91 percent for families of 5 or more. Rates were higher for Spanish interviews compared with English interviews, 92 percent versus 86 percent, respectively. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. In order of predictive ability, family size, interviewer, language of screener interview, and geographic location were found to be predictors of response to the interview. Examination status — Of the interviewed Puerto Rican sample, 84 percent completed at least one component of the examination. The distribution of examination response rates for each of the 10 variables with the largest variation, as defined by the CHAID procedure, as a proportion of the interview sample is shown in tables 14-16 for children, adolescents, and adults, respectively. (Note, all examina- tion response rates were interview-weighted; that is, inter- view sampling weights were applied to all examination respondents and nonrespondents when calculating exami- nation response rates to account for interview nonre- sponse and noncoverage.) A two-stage CHAID analysis was also performed to examine the multiway relation between examination status and the predictor variables: ® Stage 1, screener variables, only—The total family income, family size, poverty status, age, education of head of household, geographic location, and language of screener interview showed the largest differences in examination response rates among categories. The CHAID analysis for the screener variables showed that total family income was the most important pre- dictor of examination status. The variables geographic location and family size were found to be important predictors at the second level of selection. The CHAID analysis was repeated using the family rather than the individual as the unit of analysis. The candidate demo- graphic variables were those of the family head. In order of predictive ability, family size, age of head of household, sex of head of household, and geographic location were found to be predictors of response to the examination. Highest response rates were found for those families with the largest size, with a head of household less than 45 years of age, and with a female head of household. ® Stage 2, all variables — For children 6 months-11 years, the variables air-conditioning present, poverty status, language of screener interview, health insurance, received food stamps, size of place, family size, ever having had anemia, age, and SMSA showed the largest differences in examination response rates. For adoles- cents 12-19 years of age, the family having received food stamps and poverty status showed the largest differences in response rates. For adults 20-74 years of age, having received food stamps, generation in United States, poverty status, major activity during previous 12 months, family size, gender, self-perceived health sta- tus, education of head of household, ever had trouble hearing, and having health insurance showed the larg- est differences in response rates. The CHAID analysis showed having air-conditioning present, having re- ceived food stamps, and poverty status were the most important predictors for children, adolescents, and adults, respectively. Comparison between HHANES and 1982-84 NHIS The comparison among Puerto Ricans in the HHANES interview and examination weighted samples and the 1982-84 NHIS weighted sample are shown in table 17. Distributions of the HHANES examination and interview samples are similar. In general, differences in the observed distributions for age, sex, and income for the three sam: ples can be attributed to sampling variability. Differences between the two HHANES data sets and the NHIS occurred for the variables education of head of household, body mass index, and self-perceived health status. A higher percent had at least a high school education in the NHIS than in the HHANES. A higher percent of Puerto Ricans in the NHIS population considered themselves in excellent or very good health than in the corresponding HHANES population. Possible bias The CHAID analysis and the HHANES-NHIS com- parison for Puerto Ricans suggest that respondents and nonrespondents differ with respect to the distribution of socioeconomic status. Since disease prevalence may vary with socioeconomic status, it is important to evaluate potential bias that may have occurred due to differential response. In particular, 42.4 percent of the examined sam- ple were living below poverty compared with 36.0 percent of the nonexamined (but interviewed) sample. For the fasting half sample, 45.4 percent of the examined sample were living below poverty compared with 38.6 percent of the nonexamined (but interviewed) sample. A probability approach described in appendix VIII is used to estimate bias due to differential response in poverty groups for a number of conditions reported in research literature. As shown in tables 34-41, there is no evidence of nonresponse bias due to overresponse by subjects living below poverty level for prevalence estimates of the vari- ables considered in this study. 13 Discussion Areas of potential bias This study is meant to suggest to analysts possible sources of bias. An attempt has been made to identify the demographic, socioeconomic, and medical history vari- ables that are most strongly associated with interview and examination nonresponse. However, their relative impor- tance will vary from analysis to analysis depending on the strength of association with the analytic variable of interest. The analyses reported here suggest that there are a number of factors related to interview and examination status. A comparison of statistical weighting adjustment factors shown in table E with variables found to be signif- icant predictors of respondent status in tables F, G, and H suggest where potential for nonresponse bias exists. This comparison is done for each Hispanic subpopulation ac- cording to variable type—demographic, socioeconomic, and medical history. For Mexican Americans the combination of nonre- sponse and poststratification adjustments in the interview weight takes age, gender, household size, income, and geographic location into account. The combination of nonresponse and poststratification adjustments in the ex- amination weight takes location, age, household size, and gender into account. An additional noncoverage adjust- ment was made to compensate for the somewhat higher undercoverage of high-income Hispanic households. For these variables, the adjustments cause the weighted inter- view and examination samples to be distributed similarly to the civilian Mexican-American population residing in the southwestern region of the United States. Some predictor variables were not taken into account in weighting adjustments. Among demographic variables, language of interview was not accounted for in the weight- ing process. Among socioeconomic variables, education of the head of household, poverty index, and having received food stamps were not accounted for in adjustments for nonresponse to the examination. Some perceived medical problems were found more prevalent in respondents than in nonrespondents among children, adolescents, and adults. For Cubans nonresponse adjustments in the interview weight take gender, age, and income into account. Nonre- sponse adjustments in the examination weight take gen- der, age, and household size into account. An additional noncoverage adjustment was made to compensate for under- coverage of neighborhoods with few Hispanic residents. 14 Predictor variables that were not taken into account in the interview nonresponse weighting adjustments included location within Dade County, family size, and language of interview. Predictors that were not taken into account in the examination nonresponse weighting adjustments in- cluded geographic location, family income, poverty level, education of head of household, self-perceived condition of teeth in adolescents, and several health status variables for adults (self-perceived health status, self-report of hav- ing had anemia, self-perceived condition of teeth, and having had trouble seeing without glasses or contact lenses). Although location was statistically significant, its substan- tive importance is questionable. For Puerto Ricans, nonresponse adjustments in the interview weight take age, household, and income into account. Nonresponse adjustments in the examination weight take gender, age, and income into account. An additional noncoverage adjustment was made to compen- sate for somewhat higher undercoverage of high-income Hispanic households. A predictor variable that was not taken into account in the interview nonresponse weighting adjustments in- cluded language of interview. Predictors that were not taken into account in the examination nonresponse weight- ing adjustments included family size, gender of head of household, family income, poverty level, education of household, language of interview, self-report of having had anemia in children, self-report of health status, and self-report of trouble hearing. In summary, for the three Hispanic subpopulations, response rates were highest for those reporting a health care need or condition. This is consistent with evaluations of nonresponse in earlier health surveys (1,2,11,13-15). It should be noted that although the association of these variables with examination response was statistically signif- icant, the magnitude of differential nonresponse was prob- ably not large enough to cause a large bias in prevalence rates (6). The generally good agreement between the HHANES and the NHIS for the marginal distributions of age and sex contrasts with the differences in the distributions of socio- economic status (for Mexican Americans and Cubans) and self-perceived health status (for Mexican Americans, Cu- bans, and Puerto Ricans) as well as differences noted for the variables smoking status and self-reported hyperten- sion (for Mexican Americans). Table F. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination: Mexican American HHANES, 1982-84 Variable Survey component and age group Demographic-design Socioeconomic Medical history Interview: 6 months-74 years. . .......... age (Al) family size (D) geographic location season of year interviewer language of interview (S>E) Examination: 6 months-74 years. . . ......... family size (A,D) age (I) sex (F>M) geographic location season stand location (A) family size (D) SMSA/nonSMSA season BMONths-11Years. ... vv vs» language of interview (S >E) education of head of household (1) poverty index ratio (1) time since last physical (D) health status (1) language parent uses at home (S>E) education of head of household (1) poverty index ratio (1) language of medical history interview (S>E) language child uses at home (S>E) 12-19Y8aS. . cvs emma family size (D) received food stamps (A) ever had trouble seeing without education (1) acculturation glasses season generation in United States (I) condition of teeth population concentration ever did farm work 20-74 YBBIS. cvs swva a sy aWs un sex (F>M, A) received food stamps ever had anemia family size (D) language preference (S > E) body mass index (I) major activiity marital status geographic location SMSA/nonSMSA NOTES: A = most important predictor; | = inverse association with response rate; D = direct association with response rate; F>M = female response rate greater than male response rate; and S>E = Spanish response rate greater than English response rate. The differences noted here suggest that the NHIS represents a more well-to-do and healthier population of Hispanics than does the HHANES. Differences may be attributed in part to the tendency of those most likely to be medically underserved —those who are economically depressed, those who are without access to health care, and those with language barriers—to be most likely to respond to a health examination survey; and, conversely, those with adequate financial resources and with adequate health care are less likely to respond to a health examina- tion survey. These differences between the HHANES and the NHIS suggest that the HHANES interview nonresponse and noncoverage adjustments to the basic statistical weights may not have adequately compensated for the somewhat lower representation and undercoverage of high-income Hispanic households. This seems plausible since those adjustments were based in large part on imputed values for missing income (not obtained on the family question- naire) obtained as the 1980 Census median income of the neighborhood where the household was located (17). There are a number of variables that are related to response status at the interview and examination level. The adjustments have dealt successfully with a number of these variables and reduced the possibility of bias. Despite this, analysis of data internal to the surveys (CHAID) and comparison with data external to the surveys (NHIS) suggest that the weighted examined groups overrepresents the poorer less educated and those with lower health status and more health problems. The authors have elsewhere (6) estimated the effect that this overrepresentation of low socioeconomic status may have on two variables associated with income status, self-reported health status, and measured hypertension in Mexican Americans. The results indicate that bias was minimal. Bias estimation Estimates of the effect that this overrepresentation of low socioeconomic status individuals may have on vari- ables in the Mexican American, Cuban, and Puerto Rican samples have been made. For Mexican Americans there is evidence of slight nonresponse bias (less than 1 percentage point) due to the greater response of subjects living below poverty level for hypertension and impaired Fe status. For Cubans there was evidence of nonresponse bias for the variables overweight and self-reported anemia, due to overresponse of subjects living below poverty level. The 15 Table G. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination: Cuban HHANES, 1982-84 Survey component and Variable age group Demographic-design Socioeconomic Medical history Interview: 6 months-74 years. . .......... interviewer (A) age (I) geographic location language of interview (S>E) family size (D) Examination: 6 months-74 years. . . ......... interviewer geographic location family size (D) season 6 months-11years. . .......... age (A, D) geographic location SMSA central city/SMSA, not family income (I, A) poverty index ratio (I) education of head of household (I) education of head of household (I) central city 12199818. «cv i van mv min as education of head of self-perceived condition of teeth household (A, I) language of Adult Sample Person Questionnaire interview (S>E) poverty index ratio (I) 20-74y88. . sso issn ims geographic location poverty index ratio (A, I) self-perceived health status (1) received food stamps (D) self-report of having had anemia air-conditioning present (1) self-perceived condition of teeth have had trouble seeing without glasses or contacts NOTES: A = most important predictor; | = inverse association with response rate; D = direct association with response rate; and S>E = Spanish response rate greater than English response rate. survey prevalence estimates of overweight, shown in ta- ble 26, are similar to published estimates (24). A compar- ison of the bias-adjusted estimates with the survey estimates and with the published estimates suggest that the preva- lence of overweight was underestimated in females 65-74 years by 5 percentage points, and prevalence was overes- timated in males 65-74 years by 4 percentage points. However, only the results for females were statistically reliable. For Puerto Ricans there is no evidence of nonre- sponse bias for the variables considered in this study. Limitations and implications of the methodology The guiding philosophy in this paper has been an exploratory approach. Because of this, the model-free CHAID approach was used in this study instead of model fitting using logistic regression. The CHAID method does not produce an overall goodness-of-fit test statistic. Other limitations of this pro- 16 cedure are that it is not well suited for identifying interac- tions among variables; it requires a large sample size; and it does not accommodate low-prevalence predictors well. The above considerations influenced variable selection, but these factors were the tradeoff for allowing the data to speak for themselves. No preconceptions about how the data ought to behave were brought to the original screen- ing of variables to identify those related to response status. The final stage of the analysis concerned the estima- tion of nonresponse bias using a conditional probability approach. The limitations of this approach concern the assumption that must be made regarding the similarity of respondents and nonrespondents within adjustment cate- gories. This approach has been used before at NCHS (6,33-36); and the customary approach has been to as- sume that examined and nonexamined persons are similar with regard to the dependent variable within adjustment categories. This is, in fact, the same assumption underly- ing nonresponse adjustments to survey weights. Table H. Summary of significant predictors, as identified by the CHAID procedure, of response to the interview and examination: Puerto Rican HHANES, 1982-84 Survey component and age group Variable Demographic-design Socioeconomic Medical history Interview: 6 months-74 years. . . ......... Examination: 6 months-74 years. . .......... 6months-11years., . .....oovs «> interviewer (A) geographic location age (I) family size (D) language of interview (S > E) family size (D) age (I) geographic location sex of head of household (F>M) size of place (I) family size (D) age (D) SMSA/nonSMSA family income (A) poverty index ratio (I) education of head of household (I) language of interview (S>E) air-conditioning present (Al) poverty index ratio (I) language of screener interview (S>E) health insurance (I) received food stamps (D) self-report of having had anemia 12-19years. ...........ccu. received food stamps (A,D) poverty index ratio (1) 20-74years . . ..... 0... major activity received food stamps (A,D) self-perceived health family size (D) generation in United States (I) status (I) sex (F>M) poverty index ratio (1) self-report of trouble education of head of hearing (D) household (I) have health insurance (1) NOTES: A = most important predictor; | = inverse association with response rate; D = direct association with response rate; F>M = female response rate greater than male response rate; and S>E = Spanish response rate greater than English response rate. 17 References 10. 11. 12. 13. 14. 18 Gordon T, Miller HW. Cycle I of the Health Examination Survey; sample and response, United States, 1960-62. National Center for Health Statistics. Vital Health Stat 11(1). 1974. Miller H, Williams P. Factors related to response in a Health Examination Survey, United States, 1960-62. Na- tional Center for Health Statistics. Vital Health Stat 2(36). 1969. Schaible WL. Quality control in a National Health Exami- nation Survey. National Center for Health Statistics. Vital and Health Stat 2(44). 1972. Bryant EE, Kovar MG, Miller H. A study of the effect of remuneration upon response in the Health and Nutrition Examination Survey, United States. National Center for Health Statistics. Vital Health Stat 2(67). 1975. Maurer KR. Plan and operation of the Hispanic Health and Nutrition Examination Survey, 1982-84. National Center for Health Statistics. Vital Health Stat 1(19). 1985. Rowland ML, Forthofer RN. Adjusting for nonresponse bias in a health examination survey. Public Health Rep 108:380-6. 1993. Landis JR, Lepkowski JM, Eklund SA, Stehouwer SA. A statistical methodology for analyzing data from a complex survey: the first National Health and Nutrition Examination Survey. National Center for Health Statistics. Vital Health Stat 2(92). 1982. Institute for Survey Research. The HANES study final report. Contract No. HSM-110-73-376. Health Services and Mental Health Administration. Philadelphia: Temple Uni- versity. 1975. Chapman D. A comparison and analysis of examined and unexamined persons on medical history characteristics for the first round of the Health and Nutrition Examination Survey. Contract No. HSM-110-73-371. Health Services and Mental Health Administration. Rockville, Maryland: Westat, Inc. 1974. Schaible WL, Acting Chief, Methodological Research Branch, Division of Health Examination Statistics, National Center for Health Statistics. 1974. Memorandum to Arthur J. McDowell, Director, Division of Health Examination Statistics. Forthofer RN. Investigation of nonresponse bias in NHANES II. Am J Epidemiol 117(4):507-15. 1983. Morgan JN, Sonquist JA. Problems in the analysis of survey data. J Am Stat Assoc 58:415-34. 1963. Criqui MH, Austin M, Barrett-Conner E. The effect of nonresponse on risk ratios in a cardiovascular disease study. J Chronic Dis 32:633-8. 1979. Cobb S, King S, Chen E. Difference between respondents and nonrespondents in a morbidity survey involving clinical examination. J Chronic Dis 6:95-108. 1957. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 2s. 26. 27. 28. 29. Bryant E, Baird JT. Cooperation in health examination surveys. National Center for Health Statistics. Vital Health Stat 2(9). 1965. Kammerman LA, Nisselson H, Schreiber G. Evaluation of HHANES nonresponse bias. Contract No. 282-85-2104. Social & Scientific Systems, Inc. Rockville, Maryland. Westat, Inc. 1987. Gonzalez JF, Ezzati T, Lago J, Waksberg J. Estimation in the Southwest component of the Hispanic Health and Nu- trition Examination Survey. In: Proceedings of the Section on Survey Research Methods of the American Statistical Association. 1985. Gonzalez JF, McMillen M. Nonresponse and noncoverage analysis in the Southwest component of the Hispanic Health and Nutrition Examination Survey. In: Proceedings of the Section on Survey Research Methods of the American Statistical Association. 1986. Life Sciences Research Office, Federation of American Societies for Experimental Biology. Nutrition monitoring in the United States—an update report on nutrition monitor- ing. Washington: Public Health Service. 1989. Looker AC, Johnson CL, McDowell MA, Yetley EA. Iron status: prevalence of impairment in three Hispanic groups in the United States. Am J Clin Nutr 49:553-8. 1989. Carroll M, Sempos C, Fulwood R, et al. Serum lipids and lipoproteins of Hispanics, 1982-84. National Center for Health Statistics. Vital Health Stat 11(240). 1990. Pappas G, Gergen PJ, Carroll M. Hypertension prevalence and the status of awareness, treatment, and control in the Hispanic Health and Nutrition Examination Survey (HHANES), 1982-84. Am J Public Health 80(12):1431-6. 1990. Maurer KR, Everhart JE, Ezzati TM, et al. Prevalence of gallstone disease in Hispanic populations in the United States. Gastroenterology 96:487-492. 1989. Najjar MF, Kuczmarski RJ. Anthropometric data and prev- alence of overweight for Hispanics, 1982-84. National Cen- ter for Health Statistics. Vital Health Stat 11(239). 1989. Ezzati T, Leahy M. Sample weights for interviewed and examined SP’s in Dade County portion of HHANES. Mem- orandum. National Center for Health Statistics. 1986. Ezzati T, Leahy M. Sample weights for interviewed and examined SP’s in Puerto Rican portion of HHANES. Mem- orandum. National Center for Health Statistics. 1986. Kass GV. An exploratory technique for investigating large quantities of categorical data. Appl Stat 2:119-27. 1980. Miller RG. Simultaneous statistical inference. New York: McGraw-Hill Book Co. 1966. Kovar MG, Johnson C. Design effects from the Mexican American portion of the Hispanic Health and Nutrition 30. 31. 32. 33. 34. 35. 36. 37. 38. Examination Survey: A strategy for analysts. In: Proceedings of the Section on Survey Research Methods of the Ameri- can Statistical Association. 1986. SAS Institute Inc. SAS user’s guide. 5th ed. Cary, North Carolina. 1982. Research Triangle Institute. SUDAAN: Software for survey data analysis, version 5.30. Research Triangle Park, North Carolina. 1991. Hansen MH, Hurwitz WN, Madow WG. Sample survey methods and theory: Vol I methods and applications. New York: John Wiley and Sons, Inc. 1953. Hadden WC, Harris MI. Prevalence of diagnosed diabetes, undiagnosed diabetes, and impaired glucose tolerance in adults 20-74 years of age, United States, 1976-80. National Center for Health Statistics. Vital Health Stat 11(237). 1987. National Center for Health Statistics. Diabetes and OGTT Data, ages 20-74 years. Tape no 6506 (version 1). Hispanic Health and Nutrition Examination Survey, 1982-84. Public use data tape documentation. Public Health Service. 1988. Harris MI, Hadden WC, Knowler WC, Bennett PH. Preva- lence of diabetes and impaired glucose tolerance and plasma glucose levels in U.S. population aged 20-74 yr. Diabetes 36:523-34. 1987. Maurer KR. The epidemiology of gallstone disease in the Mexican American, Cuban American, and Puerto Rican populations of the United States, 1982-84. Ph.D. disserta- tion. The Johns Hopkins University. 1987. Goodman R, Kish L. Controlled selection—a technique in probability sampling. J Am Stat Assoc 45:350-72. 1950. Kish L. Survey sampling. New York: John Wiley and Sons, Inc. 1965. 39. 40. 41. 42. 43. 44. 45. 46. Bryant EE, Baird JT, Miller HW. Sample design and esti- mation procedures for a national health examination survey of children. National Center for Health Statistics. Vital Health Stat 2(43). 1971. McDowell A, Engel A, Massey JT, Maurer K. Plan and operation of the second National Health and Nutrition Examination Survey, 1976-80. National Center for Health Statistics. Vital Health Stat 1(15). 1981. Cuellar I, Harris LC, Jasso R. An acculturation scale for Mexican American normalized clinical populations. His- panic J Behav Sci 2(3):199-217. 1980. Dawson DA. Ethnic differences in female overweight: Data from the 1985 National Interview Survey. Am J Public Health 78:1326-29. 1989. Trevino FM, Moss AJ. Health indicators for Hispanic, black, and white Americans. National Center for Health Statistics. Vital Health Stat 10(148). 1984. Schoenborn CA, Cohen BH. Trends in smoking, alcohol consumption, and other health practices among U.S. adults, 1977 and 1983. Advance data from vital and health statistics; no 118. Hyattsville, Maryland: National Center for Health Statistics. 1986. National Center for Health Statistics. The National Health Interview Survey design, 1973-84, and procedures, 1975-83. National Center for Health Statistics. Vital Health Stat 1(18). 1985. Sonquist JA, Morgan JN. The detection of interaction effects. Survey Research Center, Institute for Social Research, The University of Michigan, monograph 35. 1964. 19 List of detailed tables 1. 2 Interview and examination response rates for health examination surveys of the National Center for Health SLASHES oo vvimis 5.000 50 SHEL HEE AER SERED HEE ERR A Variables from the screener, family, and medical his- tory interview included in the analysis of nonresponse to the examination. ... «co xsv ms mie 510 005 ow Eww 2020 Predictors of response Mexican Americans 3. Interview response rates by level of the predictor screener variables for Mexican Americans in the His- panic Health and Nutrition Examination Survey, 1082-83. svonm ow v0 50 05 3 5% 3x wR ww ee aE . Examination response rates for Mexican-American children 6 months-11 years—10 variables with the largest differences, as identified by the CHAID proce- dure: Hispanic Health and Nutrition Examination Sur- VEY 1082-83... cv cnn vn inan sw hin AR A EE gE Eee . Examination response rates for Mexican-American adolescents 12-19 years — 10 variables with the largest differences, as identified by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1092-83 iu ivueminsismrmmrmenwe mswmm maine ene . Examination response rates for Mexican-American adults 20-74 years— 10 variables with the largest dif- ferences, as identified by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1082-83 oc fener si RE EE RETR TE EE bi Eee . Weighted percent distribution of selected variables for Mexican Americans 20-74 years of age for 1982-84 NHIS and 1982-84 HHANES data. .............. Cubans 8. 10. 11. 20 Interview response rates by level of the predictor screener variables for Cubans in the Hispanic Health and Nutrition Examination Survey, 1982-84........ . Examination response rates for Cuban children 6 months-11 years —variables with the largest differ- ences, as identified by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1982-84 . . . .. Examination response rates for Cuban adolescents 12-19 years —variables with the largest differences, as identified by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1982-84 ........ Examination response rates for Cuban adults 20-74 years — variables with the largest differences, as identi- fied by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1982-84 ........... 22 3 23 24 25 26 27 28 28 29 12. Weighted percent distribution of selected variables for Cubans 20-74 years of age for 1982-84 NHIS and 1952-34 HHANES data ... +c concvvansnssnsnssss Puerto Ricans 13. 14. 15. 16. 17; Interview response rates by level of the predictor screener variables for Puerto Ricans in the Hispanic Health and Nutrition Examination Survey, 1982-84 . . . .. Examination response rates for Puerto Rican chil- dren 6 months—11 years — variables with the largest differences, as identified by the CHAID proce- dure: Hispanic Health and Nutrition Examination Survey, 1982-84... cour rsiuar isms nem mm Examination response rates for Puerto Rican ado- lescents 12-19 years—variables with the largest differences, as identified by the CHAID proce- dure: Hispanic Health and Nutrition Examination Survey, 1982-84... cou vanrnsbetarsssnsnenarnes Examination response rates for Puerto Rican adults 20-74 years—variables with the largest differences, as identified by the CHAID procedure: Hispanic Health and Nutrition Examination Survey, 1982-84... .. or Weighted percent distribution of selected variables for Puerto Ricans 20-74 years of age for 1982-84 NHIS and 1982-84 HHANES data. ................... Examination of potential bias Mexican Americans 18. 19. 20. 21. Potential bias in estimated percent prevalence of over- weight in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Mexican Americans: Hispanic Health and Nutrition Examination Survey, 1982-84 . . . .. Potential bias in estimated percent prevalence of ele- vated cholesterol in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Mexican Americans: Hispanic Health and Nutrition Examina- tion Survey, 1982-84... cvmsminmn ws menms mann eat Potential bias in estimated percent prevalence of im- paired Fe status using MCV model in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Mex- ican Americans: Hispanic Health and Nutrition Exam- ination Survey, 1982-84. ............. citi. Potential bias in estimated percent prevalence of self- reported anemia in the examined sample due to differ- ential reporting of anemia and poverty status in the interviewed-but-not-examined sample for Mexican 30 30 31 31 32 33 33 34 34 22. 23. 24. 25. Americans: Hispanic Health and Nutrition Examina- tion Survey, 1082-84. . ...cn vunrmmmpmny cwvmy x ery Potential bias in estimated percent prevalence of hy- pertension in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Mexican Americans: Hispanic Health and Nutrition Examination Survey, 1982-84 . . . . . Potential bias in estimated percent prevalence of self- reported hypertension in the examined sample due to differential reporting of hypertension and poverty sta- tus in the interviewed-but-not-examined sample for Mexican Americans: Hispanic Health and Nutrition Examination Survey, 1982-84 ................... Potential bias in estimated percent prevalence of gall- stone disease in the examined sample due to differen- tial reporting of poverty status in the interviewed-but- not-examined sample for Mexican Americans: Hispanic Health and Nutrition Examination Survey, 1982-84 . . Potential bias in estimated percent prevalence of self- reported gallstones in the examined sample due to differential reporting of gallstone disease and poverty status in the interviewed-but-not-examined sample for Mexican Americans: Hispanic Health and Nutrition Examination Survey, 1982-84 ...usans ms ave nsness Cubans 26. 27. 28. 29. 30. 31, Potential bias in estimated percent prevalence of over- weight in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84 ........... Potential bias in estimated percent prevalence of ele- vated cholesterol in the examined sample due to differential reporting of poverty status in the interviewed- but-not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84... ..... Potential bias in estimated percent prevalence of im- paired Fe status using MCV model in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, TOB2Ba ive mia 5 sre ns mimi wk wis ms 00 4 SR RE A Potential bias in estimated percent prevalence of self- reported anemia in the examined sample due to differ- ential reporting of anemia and poverty status in the interviewed-but-not-examined sample for Cubans: His- panic Health and Nutrition Examination Survey, IOB2Ba sin mai ans RE EERE Sa bons hm mS Potential bias in estimated percent prevalence of hy- pertension in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84 ........... Potential bias in estimated percent prevalence of self- reported hypertension in the examined sample due to differential reporting of hypertension and poverty sta- tus in the interviewed-but-not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-88. . «ian i 50m is mos mmisim somain mimi nin oe 35 35 36 36 37 37 38 38 39 39 40 32. 33. Potential bias in estimated percent prevalence of gall- stone disease in the examined sample due to differen- tial reporting of poverty status in the interviewed-but- not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84 ........ Potential bias in estimated percent prevalence of self- reported gallstones in the examined sample due to differential reporting of gallstone disease and poverty status in the interviewed-but-not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84. sus unswrwsinimsmasinv ins wean Puerto Ricans 34. 35. 36. 37. 38. 39. 40. 41. Potential bias in estimated percent prevalence of over- weight in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Puerto Ricans: Hispanic Health and Nutrition Examination Survey, 1982-84 ........ Potential bias in estimated percent prevalence of ele- vated cholesterol in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Puerto Ricans: Hispanic Health and Nutrition Examination Survey, 1982-84. susws sus nse sism sess be swiss ws Potential bias in estimated percent prevalence of im- paired Fe status using MCV model in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Pu- erto Ricans: Hispanic Health and Nutrition Examina- tion Survey, 1982-84 ........... iit, Potential bias in estimated percent prevalence of self- reported anemia in the examined sample due to differ- ential reporting of anemia and poverty status in the interviewed-but-not-examined sample for Puerto Ricans: Hispanic Health and Nutrition Examination Survey, 1982-84... .. oe Potential bias in estimated percent prevalence of hy- pertension in the examined sample due to differential reporting of poverty status in the interviewed-but-not- examined sample for Puerto Ricans: Hispanic Health and Nutrition Examination Survey, 1982-84 ........ Potential bias in estimated percent prevalence of self- reported hypertension in the examined sample due to differential reporting of hypertension and poverty sta- tus in the interviewed-but-not-examined sample for Puerto Ricans: Hispanic Health and Nutrition Exami- nation Survey, 1982-84 . . viv vninnrnrensmres Potential bias in estimated percent prevalence of gall- stone disease in the examined sample due to differen- tial reporting of poverty status in the interviewed-but- not-examined sample for Puerto Ricans: Hispanic Health and Nutrition Examination Survey, 1982-84 . . . .. Potential bias in estimated percent prevalence of self- reported gallstones in the examined sample due to differential reporting of gallstone disease and poverty status in the interviewed-but-not-examined sample for Puerto Ricans: Hispanic Health and Nutrition Exami- nation Survey, 1982-84 cco vvvurnsinninans vas 40 41 41 42 42 43 43 44 44 45 21 Table 1. Interview and examination response rates for health examination surveys of the National Center for Health Statistics Survey date Response rate Sample size Survey and age group. Beginning Ending Midsurvey Interview Exam Sample Interview Exam NHES | (18-79 years). . . .......... Oct 1959 Dec 1962 --- --- 87 7,710 --- 6,672 NHES Il (6-11 years) . ............ Jul 1963 Dec 1965 Aug 1, 1964 --- 96 7,417 --- 7,119 NHES lll (12-17 years) . . . ......... Mar 1966 Mar 1970 Mar 9, 1968 --- 90 7,514 --- 6,768 NHANES «cic ioc msvavmamus wen Apr 1971 June 1974 Nov 1, 1972 =TAYCAIE swiss wi sma Aula mas 99 74 28,043 27,753 20,749 VB YAS. » viv is wis win www 100 82 3,530 3,516 2,895 C-11Y6S. . c.nrnvamsnsimsn 99 85 2,415 2,401 2,057 V2-NT YRS «ic viv coi 5 i © wows in 99 84 2,526 2,505 2,126 UB-TAYBAUS + ui sins nis ip Hw umn 99 70 19,572 19,331 13,671 1B-BAYOAIS. . vv ax wi mins aise 99 73 12,289 12,131 8,925 55-74years. .............. 99 65 7,283 7,200 4,746 NHANEB MH. © .ov ae params wwe aa ain Feb 1976 Feb 1980 Mar 1, 1978 6months-74years. . ........... 91 73 27,801 25,286 20,322 6 months-5years ............ 96 81 5,069 4,876 4,118 B11 BAIS: « 5 5 win 5 bows re BE 94 83 2,085 1,963 1,725 ET 95 81 2,438 2,304 1,975 18-74 VBS. : =: vv 5 0 vw Fins OAS 89 69 18,209 16,143 12,504 18-58 YBAIS. + vv vv vv vs vais 91 72 10,129 9,181 7,333 BETA YBAB. . « vn v vr sw ow ven 86 64 8,080 6,962 5,171 HHANES Mexican American. . ....... Jul 1982 Nov 1983 Mar 1, 1983 6 months-74 years. . ........... 87 76 9,455 8,222 7,197 Bmonths-5years ....... ++» 92 84 1,492 1,377 1,250 6-11years. ................ 92 85 1,508 1,384 1,287 12-917 Y8RISv ; +s sms wm sms mu un 90 79 1,325 1,188 1,053 VE-TRYBAIS « « ov ins vy 0 oi 04 1 mim 83 70 5,130 4,273 3,607 18-BAYBAIS. . + vii ives wain 84 71 4,183 3,520 2,983 BE-TAYBAIS. . . . 00 nu swans ii 80 66 947 753 624 HHANES Cuban . . . ............. Jan 1984 Apr 1984 Feb 1984 6 months-74years. ............ 79 61 2,125 1,677 1,291 6months-5years ............ 84 58 165 139 95 6-11years. ................ 85 71 178 152 126 V2-V7YORIS os + 5 » won oi 3 6 0 9 2 86 72 222 191 159 1B-TA YEAS. : «crn vimmr mannan 77 58 1,560 1,195 911 18-54 years. . ............. 76 59 1,070 816 630 BETA YOarS. + 5 + wie vues 4 bi 81% 77 57 490 379 281 HHANES Puerto Rican . . .......... May 1984 Dec 1984 Sept 1984 6 months-74years. ............ 89 75 3,625 3,137 2,645 Bmonths-5Y8aIS . . uu wewiws vss 91 78 496 451 389 GTN YBAIS. «wie vi viv me won om 92 84 501 463 420 12-17 YBAIB co «tnd vw mw in ww 94 83 586 550 484 PB~TAYORIS . vo nv wis 9a Wn ® 4 0% 86 70 1,942 1,673 1,352 18-BA YAS. « + + «cn veri vm 86 71 1,544 1,332 1,094 B5-74Y0IS. + «cvs rcv nsws 86 65 398 341 258 NOTES: NHES | is the first National Health Examination Survey. NHES Il is the second National Health Examination Survey. NHES lll is the third National Health Examination Survey. NHANES | is the first National Health and Nutrition Examination Survey. NHANES II is the second National Health and Nutrition Examination Survey. HHANES, Mexican American is the Hispanic Health and Nutrition Examination Survey, Mexican American portion. HHANES, Cuban is the Hispanic Health and Nutrition Examination Survey, Cuban portion. HHANES, Puerto Rican is the Hispanic Health and Nutrition Examination Survey, Puerto Rican portion. 22 Table 2. Variables from the screener, family, and medical history interview included in the analysis of nonresponse to the examination Screener 6 months-74 years Family 6 months—74 years Child medical history 6 months—11 years Adult medical history 12-74 years Sex Family size Age Language of interview SMSA Size of place Season of year Telephone present Interviewer Health insurance WIC program’ Education Air-conditioning Food stamps Marital status Place of birth Language of interview Health status Dental care Anemia Weight status Vision problems Hearing problems Breastfed Asthma Language child usually speaks Language parents usually speak Language of interview Health status Dental care Anemia Weight status Vision problems Hearing problems Farm work Last physical exam Acculturation — Southwest only Generation in United States Language usually spoken Language preferred Work status Recreation and exercise Activity level Diabetes High blood pressure Kidney stones Chest pain Smoker Gallstones Kidney problems Coughing WIC = Women, infants, and children. NOTE: Variables shown are a topical summary of questions. Table 3. Interview response rates by level of the predictor screener variables for Mexican Americans in the Hispanic Health and Nutrition Examination Survey, 1982-83 Variable n Rate? Variable n! Rate? TOE cv snsne amram nrwnrn 9,455 87 Geographic location:3 1 San Antonio, TX. . . ..... 492 80 Age:? 3 Houston, TX .......... 611 87 6 months-11years........... 3,000 92 BGreslay, CO ... ;cvvnvns 626 89 12-19years . .............. 1,720 89 7 Midland, TX. . . ........ 648 91 20-48 YBAIS + ox vo sins vs van ws 2,828 86 9 Tucumcari, NM ........ 542 93 A5-TEYBHIS vv «vrs svn sia 1,907 79 11 Brownsville, TX ........ 605 91 13 Beeville, TX. . ov vow vv vs 513 88 1B EY PaO, TX. «sows iw nim ois 571 88 FZ TUCSON, AZo v5 5 on 2v 24 576 88 Season:? 19 San Diego, CA. . ....... 602 87 WIBIBE «vv min 0% 4 00m i ms ie We 1,660 88 21 Los Angeles, CA. . . ..... 640 91 SPAN sv wv = De 9a a wv wim in wa in 1,829 89 23 Los Angeles, CA. . . ..... 587 88 SUMMNBL.. @ oo > - Kidney problems? An eye injury? ® 64 PESTICIDE EXPOSURE L1. Have you ever done farm work, either paid or unpaid? Some examples of farm work are working with crops or animals and supervising other workers on farms or orchards. 10 Y 2 0 N(L27) ACCULTURATION M1. Do you speak any Spanish? 8 10Y 2 OO N(M4) M2. Would you say that you speak mostly Spanish, or mostly English, or do you speak Spanish and English about the same? mostly Spanish mostly English both about the same ® N ogo M3. What language do you prefer: Spanish only, mostly Spanish, mostly English, English only, or Spanish and English about equally? Spanish only mostly Spanish mostly English English only both equally Or WwN = 0ooaoo M4. Can you read Spanish? 10Y 20 N M5. Can you read English? 10Y 20 N IF “YES” TO BOTH M4 AND M5, ASK: M6. Which do you read better? Spanish English both the same N ooo M7. Can you write in Spanish? M8. Can you write in English? IF “YES” TO BOTH M7 AND M8, ASK: M9. In which language do you write better? —— —— — fe —_——d —_—— ee Spanish English both the same WN = i 65 HAND CARD ASP 4 specify country 11 OO other Spanish or other Hispanic 12 specify country 13 J American 14 OJ Anglo-American 15 OJ other group 16 | M10. Which of those groups best describes your ethnic &) 01 [J Boricuan identification? 02 OO Puerto Rican | 03 [J Cuban 04 [J Cuban-American os [0 Mexican/Mexicano 06 [J Chicano | 07 [J Mexican-American | 08 [J Hispano o9 [J Latin American 10 OO Other Spanish or other Hispanic | 11 OO American 12 OO Anglo-American | 13 O other group 14 | specify IF ANY BOX BELOW THE LINE IN M10 IS CHECKED, ASK: M11. What is your country of origin? | 1 specify | M12. Which of those groups best describes your mother’s | o1 OU Boricuan ethnic identification? | 02 [J Puerto Rican 03 [J Cuban 04 [J Cuban-American | 05 [J Mexican/Mexicano 06 [J Chicano 07 [O Mexican-American | o8 [J Hispano o9 [J Latin American 10 specify country 11 0 other Spanish or other Hispanic 12 specify country 13 [J American | 14 [J Anglo-American 15 [J other group 16 specify | M13. Which of those groups best describes your father’s 600) 01 [J Boricuan ethnic identification? | 02 [0 Puerto Rican 03 OJ Cuban 04 [J Cuban-American | 05 [J Mexican/Mexicano | 06 [J Chicano 07 [J Mexican-American | o8 [J Hispano 09 [J Latin American 10 1 specify 66 M14. In what country or State was your father born? 1 O U.S., except Puerto Rico 2 [J Puerto Rico 3 [J Cuba 4 OJ Mexico 5 [J other 6 specify M15. In what country or State was your mother born? ® 1 O U.S., except Puerto Rico 2 (J Puerto Rico 3 [0 Cuba 4 [J Mexico 5 (J other & specify 67 Appendix VI HHANES-NHIS comparison The National Health Interview Survey (NHIS) is an important source of data on the reported health status of Hispanics(42-44), and it can provide a point of compari- son with similarly collected data from the HHANES examined group. Similarities and dissimilarities between the two surveys that should be considered when interpret- ing these results are discussed here. Detailed information on the plan and operation of the NHIS has been docu- mented (45). The two surveys have important design and opera- tional features in common including the following: 1. Both are large-scale surveys utilizing stratified, multistage probability designs involving the selection of geographically defined areas (primary sampling units). 2. The two surveys share some of the same primary sampling units in the areas of the country in which the HHANES was conducted. 3. Similar demographic and medical history data were collected during household interviews, although relative positioning of specific questionnaire items in the inter- views were different. Limitations on the comparability of the two surveys include the following: 1. SMSA non-SMSA — Self-representing standard met- ropolitan statistical areas from Texas, California, Miami, and New York included in the NHIS were chosen for the comparison study. This may have led to an “urban” bias in the NHIS data. 2. Proxy status — For the NHIS, all persons 19 years or over or any age if ever married were eligible to respond for himself or herself and for any other related household member not present. For the HHANES, proxy response was allowed only for demographic and family information (including age, sex, income, and education); but medical history data were required to be self-reported. 3. Language —Both the NHIS and HHANES were based on interviews. The NHIS interviews were conducted by Bureau of the Census employees. For those interview- ers who were not Spanish-speaking, household members, neighbors, or friends of the sample person were allowed to interpret. While there was no Spanish translation of the NHIS core questionnaire, Spanish-language flashcards were used. For the HHANES, bilingual interviewers were em- ployed and a Spanish-language interpretation of the ques- tionnaire was available. 68 4. Differing primary sampling units (PSU’s) — The two surveys shared PSU’s in Los Angeles and San Diego, California; Houston, Texas; Miami, Florida; and New York, New York. In the Mexican American sample, approximately 40 percent of the HHANES sample and 70 percent of the NHIS comparison sample were drawn from these areas. 5. Nonresponse —In past surveys, nonresponse to the NHIS has generally been smaller than nonresponse to the medical history interview component of the National Health Examination Surveys (NHES). This is due primarily to the NHIS practice of allowing proxy response to medical history questions while the NHES has required self- response among adults. If there is close agreement between the two surveys, it adds to the sense of comparability and credibility of these two large-scale surveys. The comparison consists of the display of the weighted proportion of various conditions or attributes for each sample. The composition of the HHANES weights has already been described. The NHIS weights were the reciprocal of the probability of selection with adjustments for nonresponse and with poststratifica- tion to the population distribution as estimated by the Bureau of the Census. One of the strengths of the NHIS is the ability to combine data over multiple years (43). To increase the stability of the estimates, years of data were combined. To maximize comparability with the HHANES, this compari- son was limited to the combined 1982, 1983, and 1984 NHIS weighted samples. Reanalysis limited to just those SMSA’s included in both surveys did not alter the conclu- sions of the study. The comparison of the NHIS and the HHANES should be interpreted in light of the limitations mentioned — the use of proxy respondents in the NHIS and the avail- ability of a Spanish-language-translated questionnaire in the HHANES but not in the NHIS. First, the use of proxy response allowed the NHIS to collect information on those who would have been nonrespondents in the HHANES. The HHANES approach was to assume that nonrespondents were similar to respondents within nonre- sponse weighting adjustment categories. When this assump- tion was not true, estimates from the two surveys would diverge. Second, it is not certain what effect the lack of a Spanish-language questionnaire in the NHIS may have had on NHIS estimates. Although the opportunity for Lacking more direct and complete information on the conducting the interview in Spanish as well as English was socioeconomic and health status of the HHANES nonin- available in both surveys, the uniformity of translations terviewed group, the HHANES versus NHIS comparison was less exact in the NHIS than in the HHANES. This is suggests the nature and direction of possible nonresponse clearly a subject for further research. bias. Appendix VII CHAID procedure The Chi-Square Automatic Interaction Detection (CHAID) technique was used to summarize the data. CHAID is a descriptive procedure that provides the re- searcher with information about the relationships between the dependent variable (the interview status) and the predictor variables (other classification or descriptive vari- ables) by calculating the chi-square measure of association between the dependent and each independent variable. (Note, “unknown” or “missing” category data were treated as “floating” response categories and were allowed to combine with other response categories.) The predictor variable that has the most significant chi-square, after a Bonferroni adjustment for the number of variable catego- ries, is used to split the sample into groups. This process is repeated for each of the new groups until there are too few observations for further splitting. The result is a tree-like structure that suggests which predictor variables may be important and need future investigation. The computer software SI-CHAID (SI-CHAID® is a regis- tered trademark of Statistical Innovations Inc., Belmont, Massachusetts) was used to perform the analysis. Background The CHAID technique was originally developed by Kass (27) as a procedure for predicting the outcome of a categorical dependent variable on the basis of a set of independent categorical variables. But this type of “tree analysis” has its origin in the Sonquist and Morgan Auto- matic Interaction Detection (AID) program developed at the University of Michigan’s Institute for Social Research in 1964 (12,44). Advantages and limitations CHAID was found to be particularly suitable to the present analyses for two reasons: First, a large number of variables were to be screened as potential predictors of response status. CHAID is a multivariable procedure but not a multivariate one. All of the variables are not considered simultaneously, but rather are considered sequentially. Thus, the sample size prob- lems endemic to multivariate approaches (multiple regres- sion or logistic regression) are avoided. 70 Second, there was no reason to assume that the relationships between response and the dependent vari- ables were linear. CHAID is a model-free approach de- pending on the structure of the data rather than the a priori structure assumed by a model. A limitation of the CHAID approach is that it re- quires a large sample size. However, the sample sizes in the HHANES were sufficient for this type of analysis. Thus, in an exploratory-data-analysis approach, such as the present one, CHAID was the method of choice. An empirical example The illustrative analysis chosen here is the interview response analysis for Cubans 6 months-74 years in the HHANES. In this analysis, the goal was to identify screener variables that were predictors of interview response. Data for the analysis were drawn from the responses of a sample of 2,125 persons who were chosen into the Cuban portion of the HHANES. The criterion variable for the analysis is completion of at least one examination component. The criterion is scored on a 1-2 basis. The independent variables and their associated re- sponse categories for the analysis are summarized in figure I. In addition, figure I indicates whether the variable was treated as nominal (free) or monotonic (mono) in the analysis and gives the frequency distribution for each variable. The results of the CHAID analysis are summarized in figures II and III. Figure II shows an analysis summary for the total sample providing for each variable the signifi- cance level, a measure of correlation analogous to the usual r-squared, and a summary describing how the cate- gories were merged. Figure III shows the results of the sequential analysis. Initially, the interviewer variable (INV) was chosen as the “best” predictor (or independent) variable. The predictor with the smallest chi-square signif- icance is considered “best.” Having selected a best predic- tor, SI-CHAID carried out the same analysis for each population or “segment” (group of interviewers) de- scribed by the categories of the selected predictor. The completed analysis is depicted with a tree diagram in figure III. As shown, age, stand, and family size are identified as significant predictors at the second level of analysis. HHANES Interview Response Rates for SPs: Cuban Americans SI-CHAID (R), Copyright (C) 1984-1987 Statistical Innovations Inc. 375 Concord Avenue, Belmont, MA 02178 Technical Parameters... Run Mode Analysis Depth Limit: 30 Automatic Significance Levels... Predictor: 0.050 Detailed Tables Requested ... Category: 0.050 *No detailed tables= Mininum Segment Sizes... Before Split: 200 After Split: 100 Default Summary Table: Row % Bonferroni Adjustment? yes Frequency Variable: =*none= Weight Variable: *nonex Missing values are included. Dependent Category Frequency Variable Levels # Label Counts RESPONSE 2 1 yes 1677 2 no 448 Combine. .... Category Frequency Predictor Levels Type Sig # Sym Label Counts AGE 4 Mono 0.050 1 1 1t 12 yrs 343 2° 1 12-19 yrs 301 3 2 20-44 yrs 610 4 4 45-74 yrs 871 SEX 2 Free 0.050 i m male 999 2 f female 1126 SEASON 2 Free 0.0S0 1 W Winter 1610 S Spring 215 SIZE 3 Mono 0.050 1 1 1-2 576 2 3 3-4 1002 3 5 5S or more 547 LANGUAGE 2 Free 0.050 1 1 1 788 2 2 2 1337 STAND 4 Free 0.050 1 1 35 S41 2 2 37 530 3 3 39 539 4 4 41 518 Free 0.050 19 241 133 INV 15 2 2 242 128 3 3 243 122 4 4 247 115 5S S5 248 113 6 6 249 108 7 7 251 106 8 8 253 103 9 9 254 102 10 A 255 97 iy 8 2957 93 12" C 289 81 13 D 260 81 14 E 885 70 15 F other 673 Figure I. SI-CHAID program output: summary frequency distributions for dependent and independent variables The CHAID procedure culminating in the tree shown (combined) response levels for a set of predictor variables. in figure III relies on a sequential, semihierarchical search The search procedure is directed by Bonferroni adjusted procedure to partition response groups on the basis of the chi-square values. 71 HHANES Interview Response Rates for SPs: Cuban Americans Analysis of total group. Predictor INV AGE STAND LANGUAGE SIZE SEX SEASON WNP = p-value 0.15e-8 0.00017 0.0011 0.0031 0.0091 1.00 1.00 r-sq 0.033 0.008 0.007 0.004 0.004 0.000 0.000 groups 3 158D 237EF 469ABC Figure Il. SI-CHAID program output: analysis summary HHANES Interview Response Rates for SPs: Cuban Americans INV 64.29 140 Total 78.92 2.125 INV INV INV 158D 237EF 469ABC 89.77 72.07 83.72 430 1.099 596 AGE AGE STAND STAND SIZE SIZE SIZE 1 24 12 34 1 3 5 95.14 87.06 68.32 77.00 71.34 85.35 92.77 144 286 625 474 157 273 166 INV JEP 82.34 334 AGE 112 89.74 156 AGE 4 75.84 178 Figure Ill. SI-CHAID program output: tree diagram 72 Appendix Vill Adjusting for possible nonresponse bias The approach used in this report to adjust for nonre- sponse bias has been used previously at NCHS (6,23,33- 35). The development of the following approach makes clear the potential effect that the magnitude of nonre- sponse to the examination may have on prevalence esti- mates. A model is also developed to estimate the “true” prevalence when there is evidence that respondents may differ from nonrespondents. This model incorporates a variable to modify the sample estimates of the population parameters. This variable is related to both the response status and the variable for which the parameter is being estimated. An analysis of the sensitivity of estimates based on this model to differing assumptions also follows. Magnitude of nonresponse The validity of prevalence estimates based on the HHANES examination sample rests on an assumption that the prevalence of sample persons participating in the examination did not differ from that of sample persons not participating. The importance of this assumption is illus- trated in table XI. This table models the dependence of the results of the survey on the response rate and the prevalence of the attribute being estimated in respondents and nonrespondents. This model is based on the following equation: P(C) where P(C) true prevalence for a condition C P(R) = proportion of sample responding P(NR) = proportion of sample not responding P(Cg) = prevalence rate estimated based on respondents P(Cyr) = prevalence rate in nonrespondents P(Cg) e P(R) + P(Cygr) ¢ P(NR) 1) This equation shows that true prevalence is the sum of prevalences in respondents and nonrespondents weighted by the proportions of respondents and nonrespondents, respectively. If B is the ratio of prevalence in nonrespon- dents to prevalence in respondents (P(Cyr)/P(CRr)), then P(C) = P(R)+P(Cg) + [1-P(R)] « Be P(Cy) = P(Cg) * [P(R) + B-B « P(R)] ©) and the percent bias is 100[P(Cg)-P(C))/P(C) = [100(1-P(R)-B+B « P(R))] /[[P(R)+B-B + P(R)] 3) The numbers in table XI were obtained by substitut- ing values for P(R) and B in the above equation. The table shows that bias is related to both response rate and a difference in prevalence rates. There is no bias when the prevalences are equal for respondents and nonrespon- dents. When the prevalences differ, the percent bias is higher at lower response rates. With only 60 percent response, if the prevalence in nonrespondents were 25 per- cent lower or higher, than in respondents, the survey estimate would be 11 percent overestimated, or 9 percent underestimated. Estimating the “true” prevalence Looking again at equation (1), we know P(R) (the proportion of the sample responding to the examination), P(Cr) (for example, the prevalence of overweight esti- mated based on respondents), and 1-P(R). Assumptions can be made about the nature of P(Cyg) given what is known about the relationship between the condition of interest (overweight) and a variable that has been found Table XI. Percent bias for selected respondent-nonrespondent prevalence ratios and selected response rates: Hispanic Health and Nutrition Examination Survey, 1982-84 Ratio of prevalence rate for Percent of population responding nonrespondents to prevalence rate for respondents 30 40 50 55 60 65 70 75 80 85 O80 i ns mm oh ina oF adleas nade 54 43 33 29 25 21 18 14 11 8 078. so vss tonvvrs nase ves 21 18 14 13 11 10 8 7 5 4 B90: va srr ren sree AEE E 8 6 5 5 4 4 3 3 2 2 1000: 35 vost envi e viens s0sd 0 0 0 0 0 0 0 0 0 0 130 vo vv nnin van srw poms ve -7 -6 -5 -4 —4 -3 -3 -2 -2 -1 1.28 css vs nni ni une svar ns on -15 -13 -11 -10 -9 -8 -7 -6 -5 —4 180s d+ vive ® & San pry wae we -26 -23 -20 -18 -17 -15 -13 -11 -9 -7 73 Table XII. Potential bias in estimated percent prevalence of overweight in the examined sample due to differential reporting of poverty status in the interviewed-but-not-examined sample for Cubans: Hispanic Health and Nutrition Examination Survey, 1982-84 Examination Relative Survey Adjusted bias? Sex and age Sample size Response rate estimate! estimate! Difference (percent) TOMB viv 10 pe dia i aA We Re Ws EEF 860 58.2 31.95 32.59 -0.64 -40 Female DOA YEAS; i «iv x i 5 0s 2 65 ai wk intend 3 Wd 204 60.0 26.22 27.02 -0.80 -26 AB-BAYBRIB. «oi Sais iis © ov sie ok frm ms 0 tn iw 119 61.7 37.31 37.71 -0.40 -9 BE-BAYBAIS 0-1 x: 3 5: Wis bial #8 18 its & 008 I 3k 97 59.5 51.45 52.90 -1.45 -29 BE-FAYBAIR .. st s.0:5 5 #0 5 3% 4 gv W008 ar 4 08 nn ie 64 53.8 40.05 45.20 -5.15 -84 Male 20-A4YBBIB., . , vs vars wi cas BE wd? Canes 143 53.6 25.01 24.93 0.08 2 AB-BAYBAIS . voi 5 ov bis vd dip i widen Bi Bae 4 114 60.6 34.75 34.19 0.56 13 BB-BAYOBIB ce sv 4 400 vi 4 HL WH SERED aH n 78 56.9 32.00 30.80 1.20 23 BE~TFAYOBRIS . + vv + inv v 02 ov % 48 5056 & % kis 41 * * - * * Computed using basic weights (reciprocal of probability of selection). 2Relative bias =100 (survey estimate-adjusted estimate)/(standard error of survey estimate). to be related both to this condition and to response status (for example, poverty status). Thus, the “true” prevalence of overweight can be reexpressed in terms of these vari- ables and the relationship with poverty status. The model is shown as follows: P(C) = [P(Cr)]¢ P(R) + [P(Cnr)]* P(NR) = [P(CIV,r) . P(Vir) + P(CIVyg) e P(V,yr)] « P(R) + [P(CIV NR) * P(Vinr) + P(ClVynr) ® P(Vanr)] . P(NR) 4) The terms in the brackets are P(Cg) and P(Cyr), respectively; and P(C) is the true prevalence of overweight P(R) is the proportion responding to the examination P(Vir) is the proportion of the respondents living below poverty P(Vyr) is the proportion of the respondents living at or above poverty is the conditional probability of being overweight given that the person lives below poverty and was examined is the conditional probability of being overweight given that the sample person lives at or above poverty and was examined is the proportion not responding to the examination is the proportion of the nonrespondents living below poverty is the proportion of the nonrespondents living at or above poverty is the conditional probability of being overweight given that the sample person lives below poverty and was not examined is the conditional probability of being overweight given that the sample person lives at or above poverty and was not examined P(CIV,g) P (CIVaR) P(NR) P(Ving) P(Van rR) P(CIVinr) P(CIV nr) The components of P (Cg) are known (that is, can be computed from what is known for the examined sample). The components of P(Cygr) must be estimated based on two assumptions. First assumption: It is assumed that 74 the relation between the prevalence of the condition C (overweight) and the variable V (poverty status) is the same for respondents and nonrespondents (P(CIVg) = P(CIVnr)). The examination data provides an estimate of this relation. Second assumption: The distribution of the adjustment variable (P(Vyg), the distribution of poverty status among nonrespondents) is known for the nonexam- ined sampled persons who were interviewed, but is not known for the noninterviewed-nonexamined sample. Thus, it is assumed that poverty status is distributed the same among all nonexamined persons as it is among the interviewed-nonexamined group. Given these assumptions, all the pieces of this equa- tion are known, and it is possible to obtain an adjusted estimate of the prevalence of C. Values for the terms in equation (4) were estimated from the Cuban HHANES data using the basic weights (reciprocal of the probability of selection before adjust- ment for nonresponse). The adjusted estimates are com- pared with the unadjusted estimates in table XII. The adjusted estimate for total prevalence was 1.97 percent higher than the survey estimate. But the difference for age-sex specific cells varied from a 14.31 percent (5.2 per- centage points) underestimate in males 65-74 years to an 11.40 percent (3.8 percentage points) overestimate in fe- males 65-74 years. Sensitivity of bias-adjusted estimate to assumptions of analysis As previously mentioned, it is necessary in adjusting for nonresponse bias to make two assumptions about the similarity of respondents and nonrespondents. Although these assumptions cannot be verified, if they were incor- rect, an error in either direction could have been intro- duced in the final bias-adjusted estimate. To evaluate the potential impact of differential response, a sensitivity analysis was done. The first assumption is evaluated in table XIII. Values for the ratio of P(CIV,g) to P(CIV ng) and P(CIV,R) to Table XIII. Sensitivity of estimated prevalences of overweight to response selection bias in Cubans 20-74 years living above or below poverty level: Hispanic Health and Nutrition Examination Survey, 1982-84 Ratio of prevalence for respondents to prevalence for nonrespondents Ratio of prevalence for respondents to living at or above poverty level prevalence for nonrespondents living below poverty level 0.75 0.90 1.00 1.10 1.25 Percent who are overweight O75 ior an et 3 1 BE BE oR Bk Sl Th Si 27.87 30.18 31.72 33.26 35.57 O00: 4 ie: 5.0 wih i: 2 Wl EE A 28.39 30.70 32.24 33.78 36.10 00s 5 0 ow ie + rn 8 4 BH BO EAH E Sle 28.74 31.05 32.59 34.13 36.45 ToV0 01% 00 5 0000 ® Sine rt EE WA AE 29.09 31.40 32.94 34.48 36.80 BaD, Stee: nest oi eK 8 HE BIR 1 RE 29.61 31.92 33.47 35.01 37.32 P(CIV,nr) Were assigned as shown in the row and column labels, respectively. The cells show the resulting values for P(C). The stronger effect on the estimates would be caused by error in estimating the prevalence of overweight in the group living at or above poverty level because this is the larger group (about 81 percent of the population). Overestimating or underestimating the prevalence of over- weight in these nonrespondents by 25 percent would cause about a 14 percent error in the survey estimates. The second assumption is evaluated in table XIV. The ratios of the prevalence of poverty in the nonrespondents to the prevalence of poverty in the respondents are shown in the row labels. The cells show the resulting values for P(C). Based on this table, a deviation from a ratio of 1.00 by plus or minus 50 percent would result in less than 1 percentage point change in the adjusted estimate. It is clear that the first assumption about the nonre- spondents was the more critical assumption. Summary The answer to the question, “To what extent does nonresponse affect the estimated prevalence of overweight in Cuban adults 20-74 years of age?” has been made using varying assumptions about the nonrespondent group. As- suming a deviation of no more than 25 percent in these assumption parameters, the population prevalence would *U.S. Government Printing Office: 1983 — 301-019/80024 Table XIV. Sensitivity of estimated prevalences of overweight to assumptions about distribution of poverty status in nonrespondents for Cubans 20-74 years: Hispanic Health and Nutrition Examination Survey, 1982-84 Ratio of prevalence of poverty in Bias-adjusted nonrespondents to prevalence of prevalence estimate poverty in respondents of overweight Percent who are overweight OBO. ws vmsmnimsimditnsmss ss 32.66 078. mms smile pr imetnmamele oe 32.58 000: i «= ov ie ow wa ww 32.53 B00 ims +1 wm ot ssw 4 Bt WE 0 8 EB 32.49 Lid0e i vs wvin im v6 os ® ia wis own wie 32.46 2B i iva hm a wR Ei EE 32.41 HB 0 2 teh fume Bi vied 3A 4 ts 32.33 differ no more than 14 percent (or less than 5 percentage points) from the estimate based on the examined sample alone. This analysis has also shown that potential bias for individual age-sex groups could be considerably greater than for the total group. The magnitude and direction of bias differed for males and females and was greatest in the oldest age group within each gender, the groups with the smallest sample size. 75 Series 2 No. 120 409 tal and Health Statistics From the CENTERS FOR DISEASE CONTROL AND PREVENTION/National Center for Health Statistics Evaluation of National Health Interview Survey bo pi Reporting he Library - UC Berkeley Received on: 06-15-94 ital and health statistics. Series 2, Data evaluation and methods research une 1963-July 1971: Public He a ervice publication February 1994 aa U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES 3 7 Ve [foR alte iA CU loLe) 2 a Centers for Disease Control and Prevention National Center for Health Statistics Pim el PETS SN To) Copyright Information All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. Suggested Citation Edwards WS, Winn DM, Kurlantzick V, et al. Evaluation of National Health Interview Survey Diagnostic Reporting. National Center for Health Statistics. Vital Health Stat 2(120). 1994. Library of Congress Cataloging-in-Publication Data Evaluation of National Health Interview Survey diagnostic reporting by W. Sherman Edwards ... [et al.]. p. cm. — (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 120) (DHHS publication ; no. (PHS) 93-1394) “December 1993.” Includes bibliographical references ISBN 0-8406-0486-6 1. Chronic diseases — Reporting — United States — Evaluation. 2. National Health Interview Survey (U.S.) — Evaluation. I. Edwards, W. Sherman. II. National Center for Health Statistics (U.S.) Ill. Series. IV. Series: DHHS publication ; no. (PHS) 93-1394. RA409.U45 no. 120 [RAB44.6) 362.1'0723 s—dc20 [362.1'0723) 93-29687 CIP For sale by the U.S. Government Printing Office Superintendent of Documents Mail Stop: SSOP Washington, DC 20402-9328 Vital and Health Statistics Evaluation of National Health Interview Survey Diagnostic Reporting Series 2: Data Evaluation and Methods Research No. 120 This report presents the results of a study of the reporting of chronic conditions in the National Health Interview Survey. The analysis compares the reporting of certain chronic conditions by household interview respondents against the presence of these conditions in medical records, examining the differences in agreement across conditions and across respondent characteristics. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Centers for Disease Control and Prevention National Center for Health Statistics Hyattsville, Maryland February 1994 DHHS Publication No. (PHS) 94-1394 National Center for Health Statistics Manning Feinleib, M.D., Dr.P.H., Director Jack R. Anderson, Deputy Director Jacob J. Feldman, Ph.D., Associate Director for Analysis and Epidemiology Gail F. Fisher, Ph.D., Associate Director for Planning and Extramural Programs Peter L. Hurley, Associate Director for Vital and Health Statistics Systems Robert A. Israel, Associate Director for International Statistics Stephen E. Nieberding, Associate Director for Management Charles J. Rothwell, Associate Director for Data Processing and Services Monroe G. Sirken, Ph.D., Associate Director for Research and Methodology Division of Health Interview Statistics Owen T. Thornberry, Jr., Ph.D., Director John E. Mounts, Deputy Director for Operations Gerry E. Hendershot, Ph.D. Special Assistant for Science Ann M. Hardy, Dr.P.H., Chief, Illness and Disability Statistics Branch Nelma B. Keen, Chief, Systems and Programming Branch Stewart C. Rice, Jr., Chief, Survey Planning and Development Branch Robert A. Wright, Chief, Utilization and Expenditure Statistics Branch Contents TO SOUUICTRON + vie 2 eve 8 00 008 1 ew 02 0: 540 9 TE 08 4 SL 0 O10 6 a0 ns FR BR wa es ev sgh sos i i 9 56 9 at FAT 8 1 BEIT on cv no sion em imi 000 5 1 0 i 0 SE FT 0 FEU © ELE 5000 0 i 6. 26 nd I 8 90 0 BR 3 Previous research on interview reliability and VAUBIY . «oo. vvw vs wri oo wan vs wns se mou ve 0080s ain 00 30s 0a ae 840 aie 5 Interviews compared With PrOVIAET TOPOTES « i. uv vw iw sisi sims viv siems@ nee vie os oie ae 50080 50% 80 £00 04 2 08 3k win 5 Results of studies of reporting medical cONIHONS. . «.cvvvrrvrcrnviniasimrmriasarisrsnsmresenmsnnsnssnemen 7 Studies Of Interview AESIZIT IVTIICTICE! iv ov ovis mein mss ivf 08 09 585 65 5550 90k 008 58h on 00 30100 04 500000 1 00 68 0) 20 0 6 00 9 IOISCUSSION 5 5: #7. 5.2555 50300 ew 50 5k 09 00 5 550 Fk 04 0 400 0 30 VE 400 E00 8 309 0 0 36 0 0 004 0 3 00 306 6 0.0 hm 0 i a 500 0 2 0 10 MIBHIOUE 5: nav 5.7 0 in 4 oi ro 99 S000 03 0 0 0 0 B00 B02 8 8 AL 08 3 0 08 66 Bk 4 3 HF i 08 8 0 ek 4 930 11 SANIDIE ARBEIT. in 5550 04 0500 5000 00 in 00 0h 6 0 8 BR 00 $503 50 0 0 ERE 08 6 0 0 HI 6 #8 Ft 48 11 Data CONCCHON TOTING wo msn 27 00m 09 20% 2 5205 50h 05.8 10 21) 42s 11 mim: 11 wm mm 0g 4 54 mk 34 SETH 18 10 90304 8 OR 00 STE. 16 Dat COUCEEUON. ino 150 vrs 2 50 50000 2.10 500 00 548 1053 (0 BEET ntl 551 nt (0) nd 2 1 oad somos cs ss hs wn mals it. 4.9300 WOR ine rr 4, 0 0G a 8 17 Abstracting medical 1E00THS... + osm swans 100 500 3388080 WERTH ERIG BE S000 AE 20d 05 HEE RR Re 0 9 i 5 Wi 8 ke 18 Data preparation and PrOCESEING. « cs sis sev swiss » ois 0185 008 00 508 4703 $08 $308 418 #180 E55 4 $3255 67 $30 Bin 31978 $85.59, 6000 9 18 DISCUSSION vow ree tir mm kins 30 20 FR ans 34% we + ims slsty ot £20 2m ot Strato or EE on A REE RTL SR BA Se A 19 ANALYSIS ACTOR cone rn 10 0 le Be 000 00 Gh 158 F000 00.0 0181000 cs en ri bh 9 i 0 i 0 Hr 20 RESUNE ..... . 0 scsi sisnsisis om maniminmsa sy or SiR BAS 58 SED ERIE EG EIRRE 5 I BEF RR SIRE Sth R08 54 8 hid Fie 50 0 Eh 8 2A Survey statistical AiIEreNEES. .. .c vovsvi vv sions win pe mre $18 E ESET EEE SER EEE EE a 21 Interview reporting compared with medical record, by condition ..................o iii 21 Effect of person chiaraeteristios On TePOrIING. 5 + c+ 455 vis samme rams smainss sss smmsm ages 340s £5.08 304 2E02 32 IDISCUSSIOIY. +r rsrnmecai a words Sos 318) RHE TERT TUE ER 00m HRS 191 ot rt 8 1 3 snr 5 0 FEAR A 0 ER EP EE 34 BR INCI. « Fosrsnscsuarosns iri soos fn si 50 LD 0 ST, 5 8 ST. dB. ht cr cs Bs co 1 6 36 TASCOL Belo LADIES. vores orn ae 5150000 00 510.05 1 45 10 3, 80 0 L315. 0 Ss i 0 4 0 38 Appendixes 1. Health Interview Evaluation Survey QUESHONNBIIE oo 1 wos 15 0 5 5.0506 55.4 50% #06000 4000 20 in #00 00 5 0w008 w ww 0 0 $50: 00 55 JI. Health Interview Bvaluation Survey AbSIracting ProCeatl®8. uv vise em asm oi vam ah 55a 0s 0 08 506 5000s 9000 5 6005000 5 101 TIL. Loose match 1ecOmMMENBAIONS . «oo cov waua iis mi we ina ws in 010 00806 0991950 03 850340008 824 50400 400 0 038 814 9 4% 214 in 8 114 IV. Definitions of terms USS IN EMS TOPOTE cams m vss sioimsmmsn sn sa mamas sana samme anf gsansms a uame ns wees 116 List of text tables Marquis’ (1984) basic record-check matrix for binary variable no missing data, by survey response and notation Of Condition In TSHTCAL YOCOM. 5.« 5 0 vais v0 50505 3001s 50 0 05895 060.0 55590450 008 308.095 910 608 0% 0.0 5 9 nck 4 0 900 0 5 Condition prevalence per 1,000 persons, by source of information and type of condition .................... 6 Determinants of higher agreement between interview and record data. ............ooooiiiiiiiiiie... 7 Planned allocation of persons cooperating in Health Interview Evaluation Survey, by sex and age group ...... 13 Actual allocation of persons cooperating in Health Interview Evaluation Survey, by sex and age group........ 13 Planned allocation of list-sample persons cooperating in Health Interview Evaluation Survey, by event history, BEC, AN SOX (vse 500 WIRE SOB 8 700 50 BIE 0 0 Eo: 414108 0 ER 0 0 0 0 SA RO RE WE EDR 0 YB A I RP 14 Planned allocation of list-sample persons cooperating in Health Interview Evaluation survey, by event history revised to analyze telescoping, age, aNd SEX .....i vi viiiuiisssiiiissts tats sas tats easasaniansuentaane 14 Actual number of list-sample persons available for analysis, by event history from the medical record, age, and m0 mmuoE p K. L. Actual number of supplementary-sample persons available for analysis, by event history according to medical TECOTA, AEE, AMA SEX. «ot titi t tite t ttt ttte teeta eee eeeeeseeeneseeeeeeeeeneeeeseeenneeenenennn 15 Number and percent of initial draw and response rates for Health Interview Evaluation Survey, by utilization DITOUIDY 0 ov 00 m0 07 0 0570 9% 10 470 0 0 30 0 350 9 0 00 50 070 0 00 9 010 0 0 0 0 0 0 0 A I RL 15 Matrix for matching interview with medical record reports of chronic conditions ........................... 20 List of figures 1 2. 3. Comparison of National Health Interview Survey procedures and Health Interview Evaluation Survey design CIEMBNIE firs rsn min my Te RHE IRR MSRM ES RIA REM LET LE SEDER AE INIA RAR EAN SAE EERIE DI HARES 0s wi HHA 11 Summary of Health Interview Evaluation Survey sample deSIBN ....cvrciuvssrenisnsinsnsansovensorsssnsins 12 Percent overreport by interview compared with medical record in the Health Interview Evaluation Survey... .. 2 Evaluation of National Health Interview Survey Diagnostic Reporting by W. Sherman Edwards, M.B.A., Westat, Inc.; Deborah M. Winn, Ph.D, National Center for Health Statistics; Vera Kurlantzick, M.A., Project HOPE; Samantha Sheridan, M.A., Westat, Inc.; Marc L. Berk, Ph.D., and Sheldon Retchin, M.D., Project HOPE; and John Gary Collins, M.B.A., National Center for Health Statistics. Introduction The National Health Interview Survey (NHIS) is a continuous cross-sectional survey of the civilian noninsti- tutionalized population of the United States, conducted by the National Center for Health Statistics (NCHS) and the U.S. Bureau of the Census. The NHIS core interview provides national estimates of, for example, the use of physician and hospital services, and of functional limita- tions and restrictions of everyday activities for health reasons. Annual supplements provide timely information on other topics of health policy interest. The NHIS has also become a major source of esti- mates for the prevalence of certain chronic conditions and physical impairments in the United States. Since its incep- tion in 1957, the NHIS has included checklists of chronic conditions and impairments. The procedures for collecting this information have been revised several times, with the current approach (since 1978) of asking questions about more than 100 chronic conditions and impairments in six separate checklists (one checklist per household inter- viewed). There are several advantages to collecting infor- mation on the presence of chronic conditions in a national survey, as opposed to other sources. For example, a survey can capture conditions or health characteristics that may not result in medical care or may not be entered in medical records. A survey also includes persons, such as the poor and minority groups, who may be less likely to get into the health care system than others. However, a survey has drawbacks as well. Survey respondents are not medi- cally trained, so they may not know of the presence of a condition or may misdiagnose symptoms, and they may be unwilling to report the presence of certain embarrassing or stigmatizing conditions. The Health Interview Evaluation Survey (HIES) was designed to evaluate the reporting of service utilization and chronic conditions in the NHIS (the first such evalu- ation in nearly 20 years) by comparing interview responses with medical records for the same individuals. In addition, it is the first evaluation since the 1978 introduction of the current NHIS questionnaire and procedures for making prevalence estimates of chronic conditions. Comparing household interview reports with medical records can improve our understanding of data from both sources and may also shed light on people’s understanding of their own health and how well the health care system meets their needs for information. Evaluations using record-check designs are difficult; if one simply interviews persons and checks the sources they mention, it is likely that sources will be missed. Similarly, a design starting with medical records and following up with interviews will miss persons who have not sought profes- sional medical care. Like the previous studies of the reporting of chronic conditions in the NHIS (1-3), HIES drew its subjects from the membership of a health main- tenance organization (HMO) to allow as complete a verification of reports of chronic conditions as possible. However, the selection of an HMO as a source of the sample has drawbacks. The evaluation cannot examine differences by provider because there is in essence only one provider, nor can it examine the effects of variations in access to care. Also, persons belonging to an HMO may exhibit different care-seeking behavior from the general population, and they may differ in other ways. Because of interest in possible reporting differences by race, the study design of the HIES included an over- sample of black persons. The sample was also stratified by age and sex, with oversamples of older persons. Because chronic conditions generally are far less prevalent among children than adults, the selection of list-sample persons was limited to persons 18 years of age and over. To accommodate the examination of doctor visits within 2 weeks of the interview and hospitalizations within 13 months, persons identified in the medical record as having recent utilization were oversampled. HIES methods and procedures followed those of the NHIS as closely as possible. The questionnaire included a slightly modified core NHIS, with a composite condition list that included the most prevalent chronic conditions and impairments. To avoid confounding the examination of data on the list-sample persons by whether a self- or proxy report was obtained, all list-sample persons re- sponded for themselves. Other household members, in- cluding children, were included in the interview, as in the NHIS. To the extent that these persons were members of the HMO and permitted access to their medical records, they are included in some analyses. In analyzing differences between the interview report and medical record, the medical record was viewed as “the truth” for the presence and timing of doctor visits. For the presence of chronic conditions, however, the medical record may not represent a “gold standard,” and this issue is examined in this report. An analysis of the reporting of 2-week doctor visits will be published separately. This report includes a review of previous research on the reporting of chronic conditions by household respon- dents, describes the methods used in the HIES, and presents results relating to the reporting of chronic condi- tions. Appendixes present the HIES questionnaire, proce- dures used for abstracting from medical records, and specifications for a detailed analysis. The HIES was conceived and mandated by NCHS. It was conducted by Westat, Inc.; the Project HOPE Center for Health Affairs shared the design and analysis respon- sibilities. The study sample was drawn from the member- ship of the Group Health Association, whose staff provided essential assistance in identifying the sample and in mak- ing available participants’ medical records. Highlights The findings from the Health Interview Evaluation Survey (HIES) support the observations from previous research on the National Health Interview Survey (NHIS) and from other studies that survey interviews and medical records often provide very different pictures of the preva- lence of chronic conditions in a population. The HIES design and analysis did not assume the medical record to be a “gold standard” with regard to the presence of chronic conditions but rather focused on interpreting the differences between the two data sources. Some of these differences are artifacts of the procedural differences in acquiring and interpreting reports from the two sources, but others are inherent in the definitions, manifestations, and need for professional medical care of the conditions studied. Regardless of the reason for the differences, their existence sheds some light on the accuracy of survey-based prevalence estimates of chronic conditions. Chronic conditions may be classified in several ways. Conditions that require a physician’s diagnosis to identify and are very likely to require ongoing medical care showed the highest levels of agreement between the interview and medical record. The conditions studied that fall into this category include diabetes, most heart conditions, high blood pressure, and asthma. The presence of these condi- tions, once diagnosed, is likely to be noted in the medical record. Each of these conditions (with the exception of heart murmurs, a special case among heart conditions) was underreported in the HIES interview, and most other conditions were apparently overreported. Thus, interview reports of these conditions are likely to be accurate, but their prevalence may be underestimated by survey data. The problem of underestimation may be particularly se- vere for heart disease, where individuals with more than one heart condition according to the medical record often reported fewer in the interview. The other conditions apparently underreported by HIES respondents were cataracts and dermatitis. Al- though the medical record may have overstated the prev- alence of cataracts (counting some that were surgically removed before the “past year”), it is likely that cataracts were underreported by survey respondents. Many nota- tions of “beginning cataracts” or “early cataracts” were in the records, but these conditions may not have been serious enough for respondents to remember or may not have even been mentioned by the provider who discovered them. Dermatitis is a condition for which chronicity is difficult to determine from the medical record — the appar- ent HIES underreport is unlikely to indicate a correspond- ing underreport from the NHIS. At the other end of the spectrum from the first group of conditions are those that can only be diagnosed by patient report. In the HIES, constipation and tinnitus meet this criterion. Both were significantly overreported by HIES list-sample persons, and both had very low rates of agreement with the medical record. For these condi- tions, the medical records shed almost no light on the accuracy of interview-based prevalence estimates. How- ever, the records do suggest that many people do not report these conditions to their physicians, so medical records would almost certainly underestimate prevalence. Another group of conditions is those that may be quite salient to the persons suffering from them but that may not require ongoing treatment and thus may not be in the medical record. These include orthopedic impair- ments, visual and hearing impairments, migraine head- aches, varicose veins, allergic rhinitis, and chronic sinusitis. These conditions were all substantially overreported in general in HIES interviews, but with the exception of visual and hearing impairments, all had a number of underreports in the medical record as well. The presence of impairment is a somewhat subjective determination, whether by a provider or an individual; other conditions in this group, such as constipation, may tend to be self- diagnosed. Overall, it is clear that medical records alone would provide a very different picture of prevalence for this group of conditions than do interviews and that the rates from medical record data would likely be consider- ably lower. Some conditions studied are less well defined than others, from both the household respondents’ perspective and the clinical perspective as well. Some interview re- ports of arthritis, for example, although technically “false positives,” appear to match clinically equivalent condi- tions in the medical record. The extent to which other reports of arthritis may reflect more generalized joint pain could not be determined. Circulatory conditions provide particular definitional problems for respondents. Persons with several heart or other circulatory conditions seem to tend to group them under one heading. The HIES found evidence of this tendency for heart conditions; it may be true for the larger family of circulatory conditions as well. That is, persons with heart disease may report “high blood 3 pressure” as the overall condition that encompasses all their circulatory problems. The HIES design, using an HMO membership, stud- ied only persons with good access to health care, including preventive care. Some evidence from the HIES analysis and previous research indicates that people who receive medical care are better able to report the presence of chronic conditions. This is clearly true for the first group of conditions described earlier, because a physician’s diag- nosis is necessary for patients to know that they have a condition. Among the general population, many of whom have less access to medical care than the study sample, it may be that the conditions underreported in the HIES are even more underreported for the general population. This is because some people may not have received a diagnosis and some may not have sought medical care after receiv- ing a diagnosis. Conversely, it may be that self-diagnosed conditions might be more overreported among the general population than in the HMO study sample because people with limited access to care might have less chance to have their own diagnoses refuted. Finally, some proxy effects seem to be present in the reporting of chronic conditions. Although the HIES did not include a formal study of proxy reporting, a compari- son of probable self-responders and persons with proxy reports among household members indicated that proxy reports included considerably less overreporting, but agree- ment with the medical record was about the same as for self-responders. The net effect of proxy reports on NHIS prevalence estimates is difficult to determine from the analysis possible in the HIES. Previous research on interview reliability and validity For more than 50 years, the accuracy of data reported on household health surveys has been studied by examin- ing medical provider reports and through review of medi- cal records, provider surveys, or physical examinations of study subjects. These methodologies have been applied to large national surveys, including the 1935-36 National Health Survey (4), the Hunterdon County Health Study (5), the National Health Interview Study (NHIS) (1,3,6), the Center for Health Administration Studies 1970 Health Survey (7), and the 1977 National Medical Care Expendi- ture Survey (NMCES) (8-10), as well as to smaller, more focused studies. Although each study has a different design, all of the studies attempt to describe the error in survey results. Most of the studies sponsored by NCHS referred to in this report are described in greater depth by Jabine (11), who reviews findings from methodological research on health interview surveys as they relate to chronic condition reporting. He discusses sources of information used, the size of various components of nonsampling error, and relationships of these errors to data requirements, respon- dent and interviewer characteristics, and survey design features. In addition, Jabine describes current NHIS ob- jectives and historical changes in the survey, as well as NHIS operating procedures. He evaluates the quality of chronic conditions data, discussing which chronic condi- tions should be reported, as well as alternative evaluation methods. Interviews compared with provider reports The NHIS is the principal source of prevalence esti- mates for many chronic conditions in the noninstitutional- ized U.S. population. Prevalence estimates from household reports are subject to various kinds of reporting error. Underreporting may occur for several reasons. Interview respondents may not be aware of the presence of a condition, particularly if they are reporting for others. They may not know the proper name for a condition or they may forget that it was present. They may also choose not to report a condition. Overreporting may also occur as respondents misdiagnose medical problems or confuse or not remember names of conditions. Because of these limitations of household respondents as sources of clinical information, one might consider another possible method for producing such prevalence estimates, through the review of a nationally representa- tive sample of medical records. However, as Marquis (12) and others have described, medical records have shortcom- ings as sources of prevalence data. Perhaps the most significant limitation for prevalence estimates is that only people seeking medical care are included. Limitations of record checks Marquis identified limitations of particular record- check methodologies. He was concerned with response bias, the systematic overreporting or underreporting of a medical condition or health service use. He described a basic record-check typology in terms of the values ob- tained for a binary variable (i.e., a variable with two possible values) from two different sources, specifically a household interview and medical records. This typology is reproduced as table A. Cell A may be referred to as “positive match” and cell D as “negative match.” Cells B and C represent disagreement between the two sources; if the record is taken as truth, cell B would be considered a false positive or overreport, and cell C would be a false negative or underreport. Marquis extended this model to describe the design of record checks. A design in which a sample of persons with a particular characteristic (such as the presence of a certain chronic condition) is drawn from records and the characteristic is then tested for in a survey he labeled “AC,” noting that such a design would not capture over- reports, i.e., responses in cell B. On the other hand, a design in which a survey is conducted first and record checks performed on persons reporting a characteristic of interest (“AB” design) would fail to capture underreports, i.e., responses in cell C. Record checks of either AB or AC Table A. Marquis’ basic record-check matrix for binary variable with no missing data, by survey response and notation of condition in medical records Survey response Condition noted in medical record Yes No Both responses 1 A A Cc A+C NO: sum mwa B D All conditions . . A+B A+B+C+D NOTES: A is positive match, B is false positive, C is false negative, and D is negative match. design would thus not measure response bias accurately; estimates of bias would be skewed by the limitations of the design. Fully designed record checks identify a population and sample from it independently of records, obtain survey and record information for each sampled element, and compare the two data sources. Thus, Marquis believes that cognitive research on health surveys should contain external validation features such as fully designed record checks or other careful strategies to measure the correlation of survey responses with true values. Furthermore, because of the problems inherent in certain types of record checking, it cannot be assumed that respondent forgetfulness is the dominant response problem in health surveys. In addition, record checking has inherent limitations; for example, it does not explain why respondents give incorrect answers. Looking more generally at the use of records in survey research, Edwards and Cantor (13) expanded Tou- rangeau’s (14) cognitive model of survey response pro- cesses to include responses based on a review of records. They pointed out idiosyncratic sources of error in using records, including error that may arise during the creation of records and error resulting from using records devel- oped for a purpose other than research. Records thus have a different “error structure” than do interviews, where one is concerned, for example, with how well respondents understand questions, how well they recall relevant information, and how willing they are to report potentially embarrassing facts. Thus, even for a population for which the operational difficulties of selecting a repre- sentative sample of medical records are overcome (such as the membership of an HMO), one would expect that the inherent differences in the data would almost certainly result in different prevalence estimates. Physical examinations and other data sources Physical examinations appear to yield yet a different set of prevalence estimates from interviews or medical records. A comparison of clinical examination and medical history in the National Health Survey (15) found that only about half of adults 25-74 years of age classified as “definite hypertensive” in the examination reported being told by a doctor that they had high blood pressure. Gordon (16) described a three-way record check compar- ing self-administered medical history reports, physical examinations, and private physicians’ reports, for heart conditions and hypertension. In the full sample, the prev- alence of heart disease was slightly higher using physical examination as the source than using medical history and slightly lower for hypertension. Table B compares preva- lence estimates from all three sources for the subsample subjected to the medical records verification. For both heart disease and hypertension, the medical record showed the lowest prevalence among the three sources. Heart disease had the highest rate using examination as the source, and hypertension had the highest rate using self- administered medical history. 6 Table B. Condition prevalence per 1,000 persons, by source of information and type of condition Source of information Medical Condition history Physical Medical examination records Rate per 1,000 persons Heartdisease. ............... 161.9 192.6 118.9 Hypertension .... .« . wv wav sv svss 204.9 172.1 133.2 SOURCE: Gordon (16). Methodology of previous studies The two previous studies of diagnostic data in the NHIS, one reported by Balamuth (1) with a sample drawn from the Health Insurance Plan (HIP) of New York City, and the other reported by Madow (2,3) with Kaiser Permanente (KP) members in California, both used re- verse record-check designs, or “AC” designs, using Mar- quis’ typology from table A. Interview responses were compared with diagnostic information from medical records; the HIP study used an existing form routinely completed for medical encounters, and the KP study used a specially designed form completed by KP physicians for 1 year. Both studies acknowledged the limitations of medical records as validation. Balamuth largely limited analysis of the HIP study to conditions reported in the medical record, merely pointing out differences between household- reported conditions that were and were not in the record. Madow’s study was limited to conditions entered in the medical records or about which respondents said they had spoken with a physician during the year. Thus, a condition a respondent reported, but for which no physician had been seen during that year, would appear as an overre- port, even though a physician may have been seen in the previous year. Harlow and Linet (17) reviewed studies comparing questionnaire responses of chronic conditions to medical records. Accuracy of recall was measured by agreement between the two data sources, although not all studies quantified the agreement. The authors noted that accu- racy of recall includes correct reporting of medical condi- tions and absence of medical conditions in both data sources. Thus, if medical records are reviewed only for subjects reporting disease, the measure of agreement does not assess false negatives. The converse is also true; if interviews are conducted only for subjects whose medical records contain notation of disease, false positives cannot be calculated. Harlow and Linet noted that in these studies, because the condition data were derived from two different sources rather than being a repeat measurement, the term ‘“‘reli- ability” is not appropriate to describe the accuracy of reporting. They concluded that the Kappa statistic (see “Kappa statistic as a measure of agreement”) and overall proportion of agreement remain the most useful summary measures. Subsequently, Harlow and Linet (18) and Hertz- Picciotto (19) refined these views, stating that the use of medical records for assessment of accuracy is inappropri- ate for conditions in which medical service use depends upon self-identification of medical problems and subse- quent care-seeking behavior. Harlow and Linet believe that medical records are appropriate for the assessment of conditions that have clear and unambiguous diagnostic criteria, are relatively severe, and require frequent physi- cian contact. They conclude that agreement between med- ical records and self-reports cannot be generalized across conditions or across severities of conditions. Kappa statistic as a measure of agreement The Kappa statistic is widely used as a measure of interrater agreement, a method for analyzing the variation in different observer responses to the same phenomenon (Landis and Koch (20)). The Kappa statistic is a weighted proportion that summarizes the extent of agreement, adjusted for the rate of agreement expected by chance. Landis and Koch suggested value labels corresponding to the range of possible values for the statistic, with the labels providing benchn /rks for interpreting the statistic. In a critique of ‘fic Kappa statistic, Maclure and Willett (21) noted tk it it was originally conceived as a measure of agreement between two observers who sought to classify subjects into two nominal categories. The Kappa statistic has also been interpreted as a measure of validity. According to Maclure and Willett, this is not an appropri- ate use of the Kappa statistic. The authors cited as the Kappa statistic’s major weakness the fact that it is a measure of the frequency of exact agreement, not a measure of the degree of agreement. The same weakness applies to simpler measures of agreement such as percent agreement and percent over- or underreporting. Several studies of the reporting of medical conditions (1,3,10) have addressed this problem by examining “loose matches” of interview and medical record data in which categorical definitions were expanded. Another criticism of the Kappa statistic is that it measures agreement, which may or may not be equivalent to accuracy. Thus, if two raters agree on an incorrect judgment, resulting statistics may be biased (22). It is difficult to imagine a resolution of this weakness for the current application because a third source of information may itself be subject to idiosyncratic error, as noted earlier in an examination of interview, medical records, and physical examinations (16). Results of studies of reporting medical conditions This section describes some results from previous studies examining the reporting of medical conditions by survey respondents. Table C lists sources that describe characteristics of persons and correspondence with higher levels of agreement between interview and medical record reports. Details are provided in the sections that follow. Early studies The first use of a physician report to verify household- reported data occurred more than 50 years ago, in the Table C. Determinants of higher agreement between interview and record data Factor Reference Male sex Male sex, nonelderly Female sex Female sex, elderly Nonelderly age Age over 44 years White race Self-report versus proxy Proxy for spouse versus child Proxy reporting stigmatizing condition Less threatening condition? More salient condition® More numerous conditions More numerous physician visits Recent physician visit Higher levels of expenditure Medication for condition Fewer household members Urban location Daugherty! (7); Linet et al. (23) Madow (3) Balamuth (1) Madow (3) Daugherty! (7) Balamuth (1) Daugherty! (7); Linet et al. (23) Balamuth (1); Linet et al. (23) Balamuth (1) Berk et al.2 (8) Daugherty! (7); Cox and lachan* (10); Trussell and Elinson (5) Daugherty! (7) Daugherty1 (7) Daugherty! (7); Madow (3); Balamuth (1) Balamuth (1) Daugherty! (7) Madow (3) Balamuth (1) Daugherty? (7) Daugherty analyzed “physician visit conditions,” that is, any condition for which a respondent had a physician visit during the survey year. 2Respondents reported “physician visit conditions." 3A classification of threatening versus nonthreatening diseases was developed by Cannell and Fowler (6) to identify conditions most likely to be misreported because they are threatening or embarrassing. These conditions are called stigmatizing by other authors (8). 4The condition for this study was “physician visit condition.” Ssalience is used to mean severity of the condition or importance of the condition to the patient. 1935-36 National Health Survey (4). Trussell and Elinson (5) also verified each major condition classification in their Hunterdon County study. Trussell and Elinson found that 30 percent of medically-attended conditions mentioned by the attending physicians were not reported in the family interview. For some conditions such as obesity, Trussell and Elinson noted that about 80 percent of the time a condition was listed in the medical record but not re- ported in the household interview. Studies using health maintenance organization members The HIP and KP studies described earlier examined the quality of NHIS diagnostic data in HMO settings, where it is relatively easy to collect medical provider record data from all providers (without missing those not reported by the respondent). The HIP study surveyed members who sought care during a specific 12-month period (1). The diagnoses from their medical records were summarized, and the summary records were compared with interview reports of chronic illnesses taken at the end of the study period. Conditions noted during physician visits that were not diagnoses were not included on the summary records. Families in which at least one person had received a medical service related to a selected list of conditions were sampled three times as intensively as other families. Two recode classifications were used for matching health conditions in the HIP study: recode number 1, which had 278 detailed titles, and recode number 3, which had 43 more general categories. Three types of matches were recorded: two that matched according to each recode type, and one that did not fit a recode, but had character- istics recorded in the interview that allowed a match to be made to the summary record. Fewer than half of the conditions gleaned from med- ical record summaries were reported in interviews, with underreporting ranging from 4 to 76 percent. The authors suggested the following factors not related to accuracy that may have contributed to low match rates: Conditions from the summary records may have been errors, some conditions judged from the medical record to be chronic may have been acute and thus not appropriate for men- tion in the interview, and lack of training or experience of the interviewers. Self-reports were more often matched to summary records than were proxy reports in 21 of 32 class or diagnosis categories. The study also concluded that the proportion of all conditions inferred from the summary records that are correspondingly reported in interviews remains constant no matter how many summary record diagnoses are sustained by the given individual. In addi- tion, the HIP survey found some underreporting of physi- cian contacts in both the 2-week preinterview period and the previous year. This study did not examine whether medical care reported as occurring in a given time interval did, in fact, occur within that interval. 8 In the second HMO study, Stanford Research Insti- tute compared interviews of a sample of members of the Kaiser Foundation Health Plan’s Southern California Per- manente Medical Group with medical encounter forms developed specifically for the study. The study was limited to conditions that were entered in the medical records or about which respondents said they had spoken with a physician during the year. The results can be applied only to conditions for which a physician had been seen in the past year. The study sample was designed to be able to measure the effect the number of visits had on accuracy of recall. The survey found 15,417 chronic conditions reported in interviews or records, but only 7,182 after exclusions noted in the previous paragraph were made. Many condi- tions that were under- or overreported were those for which only a single physician visit was made during the study year. The matching of respondent and record re- ports of conditions improved markedly as the number of physician visits increased and also when medication for the condition was taken on a regular basis. The most accurate reporting was found for diabetes, vascular lesions of the central nervous system, heart conditions, diseases of the gallbladder, and absence of fingers and toes. Many more medical records than interviews noted benign and unspecified neoplasms, mental illness, menstrual disor- ders, and skin diseases, and there were few household reports that were unconfirmed by medical records. The opposite pattern was observed for allergic rhinitis, asthma, tuberculosis, headache and migraine, hypertension, hem- orrhoids, rheumatic fever, sinusitis, bronchitis, visual and hearing impairments, and speech defects. These had many unconfirmed household reports and few medical record notations not matched by interview responses. The au- thors suggested that overreporting of these conditions could be the result of their long duration, as they may have begun well before the period covered by reviewed medical records. Center for Health Administration Studies research In the 1970 Center for Health Administration Studies of the University of Chicago survey, Daugherty (7) noted a strong inverse relationship between underreporting and the condition’s effect on the individual. The greater the effect, the less often it was underreported. Daugherty explained that the type of illness affects patient reporting. She also described error estimates for reporting of three types of illnesses. In her study, 99 physician visit condi- tions were matched on a per person rather than a per visit basis. Daugherty found that 40-50 percent of conditions reported by one source were not reported by the other. Overall, 35 percent of physician-mentioned conditions matched patient reporting. The author noted that report- ing accuracy did not vary widely by age. She found that males were slightly better reporters than females and that the largest differences in reporting were by race and by urban versus rural location. White people had some- what higher agreement with medical records than people of other races, and urban more than rural residents. The study found that self-reporting was not more accurate than proxy reporting. People with more conditions had a higher accuracy score, probably because of the greater effect on their lives, called the “salience effect.” Those reporting six or more conditions had an overall accuracy score of 73 percent, which was well above the overall mean. The most serious conditions have an overall accu- racy score of 62 percent, also above the overall mean. Underreporting by respondents was found to be greater than overreporting. This study differs from others in that it found a greater effect of demographic factors on reporting error. Record checks of special populations Linet et al. (23) studied people diagnosed with chronic lymphocytic leukemia, comparing medical records with questionnaire responses about a wide range of health conditions. The overall proportion of agreement for each condition, Kappa statistic, and confidence interval for the Kappa statistic were calculated. The Kappa statistic was calculated because it incorporates an adjustment for chance agreement. Agreement for self-respondents and proxies was compared. Based on their results, the authors concluded that some specific diseases are more accurately identified in medical records, but other conditions (such as allergic rhinitis) are more accurately ascertained from interviews. They suggest that other conditions such as asthma, may be best determined from a combination of medical records and interviews. The authors also concluded that recall is consistently better for self-respondents than for surrogates. Studies using the National Medical Care Expenditure Survey Cox and Cohen (9) explored whether a household interview survey could be used to predict provider re- sponse to a survey. The basic research question reflected the authors’ supposition that the medical record is the most reliable data source for conditions of the population. The authors compared reason-for-visit reports from the Medical Provider Survey component of the NMCES with reports from the household component, which obtained data on the use of and costs for health services from a national probability sample of the civilian, noninstitution- alized population. The Medical Provider Survey oversam- pled providers of survey respondents believed to be poor reporters, based on social demographics. The authors found that only 30 percent of conditions reported by physicians were reported by households, and only 40 per- cent of the conditions reported by households were re- ported by providers. Subsequently, the authors collapsed conditions into 16 categories to determine whether this poor match was caused by inability to match at a greater level of detail. This change substantially improved agree- ment between the two sources. Thus, the authors sug- gested that the coding system should be modified to reflect the less precise nomenclature more familiar to nonprofes- sionals. True agreement could be detected at a less detailed level. They also recommended relying on medical provider surveys rather than household interviews for resolution of differences between sources. Cox and Cohen concluded that the relationship between these two data sources is too weak to allow prediction of the provider’s report of reason for visit from a household survey report. Cox and Iachan (10) investigated the effectiveness of household reports of conditions in describing providers’ corresponding diagnoses. The authors compared re- sponses to the NMCES with those included in the Medical Provider Survey component. Respondents were asked for conditions and providers were asked for diagnoses. The principal measures used were the percent of household reports matching provider reports and the percent of provider reports matching respondent reports. In addi- tion, to examine overall agreement by demographic cate- gory, the (overall) probability of agreement was calculated. This was the sum across 63 conditions of the weighted percent of visits for which the two reports matched. Because of the large number of conditions, no correction for chance agreement was made (Kappa statistics were not used). Agreement for specific conditions was generally weak, but it improved when conditions were grouped together to be less specific. Questionnaire revisions also would have improved agreement because different ques- tions were used for household respondents and providers. Berk, Horgan, and Meyers (8) challenged the notion that self-respondents were better reporters than were proxies on health interview surveys. Using a survey of all the health providers for respondents to the NMCES for comparison, the authors were able to evaluate whether proxies or self-respondents had more accurate reports. This study focused specifically on ‘“‘stigmatizing condi- tions” that can be very serious or that may be embarrass- ing to the patient. It was found that for these conditions, although the number of conditions reported was higher for self-respondents than for persons with proxy re- sponses, the proxies reported as well as or better than actual patients when compared with medical records. Previous examinations of proxies versus self-reporters in the NHIS (24,25,26) had found self-reports generally superior to proxy reports for a variety of health indicators. Berk et al. (8) suggested that fewer conditions were reported by proxy respondents because persons not present during a household interview have fewer medickl condi- tions than those who are present. This factor had Rot been controlled for in other studies. This report concluded that the use of a household proxy does not result in indgeased reporting bias for the conditions examined. Studies of interview design influence Questionnaire types The Michigan Survey Research Center Study (27) was conducted in 1968 to compare differences in reporting of 9 health conditions (both chronic and acute) obtained through three different questionnaire types and data collection procedures. In a survey of nonelderly white adult females of low-to-middle socioeconomic status in Detroit, Michi- gan, three questionnaire versions were tested: (1) an extensive interview using multiple cognitive frames of reference, multiple cues, additional probes, and recogni- tion of items through numerous questions; (2) a respon- dent diary completed daily for 1 week, followed by an interview visit; (3) a control procedure with one interview and a shorter questionnaire with the same major items and questions as the 1968 NHIS. Although there was little difference in reporting levels from the diary and control procedures, the extensive procedure resulted in a 58- percent higher mean number of conditions reported per person than the control group; most of the increase came from conditions not on the checklist. Furthermore, the majority of newly reported conditions were shown to have significant public health implications. Comparison of condition and person approaches From July 1967 through June 1968, a multistage national probability sample of 43,600 households was interviewed to compare the results of a “condition” ap- proach and a “person” approach in the NHIS (28,29). The condition-approach survey, in the same format as the previous NHIS, included a series of direct questions on accidents, injuries, and illnesses of short and long dura- tion, followed by a checklist of selected chronic conditions and impairments. The person approach also collected reports of health conditions from questions at the begin- ning of the interview on bed days, activity limitations, and physician contacts. It differed from the condition ap- proach by starting the collection of health condition data through the effects of conditions. In addition, the person approach limited collection of prevalence data on chronic conditions to one of six specific body systems during a given interview. During the test period, information was collected on chronic conditions affecting the digestive system. The person-approach survey resulted in significantly higher overall prevalence for the conditions studied (those affecting the digestive system) than the condition- approach survey. The increase was considered an improve- ment and was attributed to the greater detail in the checklist used in the person approach. This finding led to the present system of using more detailed checklists cov- ering only a single body system in each interview. Further- more, beginning in January 1969, the NHIS replaced its condition approach with a person approach. Discussion Using provider reports to assess the accuracy of diag- nostic data reported in household interviews has provided useful information in a number of studies. However, this methodology cannot be said to provide precise measures 10 of reporting error; it is limited by flaws in medical records as a data source, by the statistics available to interpret comparative data, and by lack of knowledge of the true prevalence of medical conditions. The basic assumption of most investigations using record-check designs is that the provider record is “the truth.” However, this may hold only for those conditions that have clear diagnostic criteria, are perceived as rela- tively severe by the patient, and require frequent physician contact (17-19). Chronic conditions that do not require ongoing physician contact may be correctly reported by household interview but may not be in provider records. Some record-check study designs do not include full household interviews and medical record review. Rather, only those households for which certain conditions were found on medical records might be interviewed. Alterna- tively, only those provider records for households mention- ing certain conditions might be examined. These incomplete survey designs can result in erroneous estimates of the bias attributable to the use of a survey report (1-3). The record-check design is also limited by the differ- ence between the questions asked in the household survey and the medical record. The sources can be expected to differ in content (10). Even when the patient and provider intend to describe the same condition, they may use different nomenclature, resulting in mismatches (1,3,9,10,30). “Agreement” between household and provider sources has been inconsistently defined in previous studies. How closely do the two sources have to match to be considered in agreement? Some investigators have explored matching in broader categories with generally improved match rates. The Kappa statistic controls for marginal variations in measuring agreement, but because it only measures exact agreement, it may be considered to measure only the frequency of use of the same medical terminology (20) in reporting of health conditions. The Kappa statistic cannot, of course, measure the intent or true meaning of what was actually reported from either source being compared and whether that intent is in agreement between sources. A record check to examine the reporting of chronic conditions in a household survey would be most likely to be successful if the following conditions are met: ® It uses a full design, allowing evaluation of both over- and underreports. ® It measures agreement in some standard ways but goes further to examine the nature of disagreements to determine whether two sources are really talking about the same phenomenon in different terms. ® It does not necessarily assume the medical record to be the truth, but considers the possible reasons why some chronic conditions might legitimately be absent from the medical record. The design and analysis of the HIES, described in the following sections, were planned with awareness of these lessons from previous research in evaluating survey re- ports of chronic conditions. Methods This section presents the methodology used to con- duct the HIES. The evaluation was designed to mimic the content and procedures of the NHIS as closely as possible within certain design and analytic constraints. The differ- ences in design and conduct between the HIES and the NHIS are presented in figure 1. Sample design The HIES was conceived as a full-design records- check study. That is, following the typology used by Marquis (12) presented in table A, the intent was to examine the reporting of chronic medical conditions by interview respondents in such a way that both apparent interview overreports (cell B in table A) and underreports (cell C in table A) could be detected. Further, the design was to allow interpretation of the absence of reports of a condition from both sources as agreement that the condi- tion was not present. To this end, the study universe was members of Group Health Association (GHA), a staff- model HMO in the greater Washington, D.C., area. The use of a staff-model HMO with centralized records was the surest and most efficient way to implement a full design because the HMO’s records provide a nearly com- plete inventory of members’ use of health care services. The sample was restricted to individuals who had been GHA members for at least 3 years before selection to maximize the completeness of participants’ medical records and to further strengthen the record-check design. The study design was further guided by the desire to evaluate the reporting of chronic conditions by age, race, and sex, as well as to evaluate the reporting of medical events (doctor visits and hospital stays), which will be the subject of a separate report. However, this secondary objective strongly affected the overall sample design by leading to oversampling of persons with recent medical utilization. Because of cost considerations early in the planning of the HIES, the target sample size was 1,000 self-responding adults selected from the GHA membership rolls. Children were omitted from this list sample because of their rela- tively low prevalence of chronic conditions. (The most Data collection period Contact procedures Respondent selection Questionnaire content Data preparation Continuous survey; cases targeted for 2—week field period In person; sought household informant Knowledgeable audut in household Core and supplement(s) U.S. Bureau of the Census and National Center for Health Statistics staff rules for resolving discrepancies Area NHIS' practice HIES? procedures Sample frame Area probability; nationally representative List of members of Washington, D.C., area health maintenance organization Sample design Multistage selection, oversampling of areas with higher Disproportionate sampling by age and whether recent doctor proportion of black residents visit or hospital stay Interviewer selection U.S. Bureau of the Census staff; mostly experienced Westat staff, many new hires Interviewer training Verbatim training by U.S. Bureau of the Census staff Verbatim training by U.S. Bureau of the Census staff Field work lasted 6 months; cases targeted for 2—week field period Telephone appointment allowed; sample person only Sample person only Modified core only Westat staff; same procedures except: refer to questionnnaire for resolving discrepancies TNHIS is National Health Interview Survey. 2HIES is Health Interview Evaluation Survey. Figure 1. Comparison of National Health Interview Survey procedures and Health Interview Evaluation Survey design elements 11 common chronic conditions among children occur at a rate of about 50 cases per 1,000 persons.) Because the NHIS is a household interview, HIES interview data were collected for all household members as well as the list- sample persons. Many of these household members were also GHA members. The total sample available for analy- sis included, in addition to the list sample, all such household members who signed permission forms allowing access to their GHA medical records and for whom records were located. This group was called the “supple- mentary” sample or “household members,” as distin- guished from the “primary” sample or “list-sample persons.” Because GHA contracts with the Federal Govern- ment to provide health coverage to employees and be- cause Federal employees may be atypical in their reporting of chronic conditions in a Government-sponsored survey, the number of Federal employees in the list sample was limited in the study design. Employees of GHA, Westat, NCHS, and the U.S. Bureau of the Census were excluded from the list sample. The sample design is summarized in figure 2. Selection of medical centers GHA was serving approximately 160,000 people at nine medical centers in the greater Washington, D.C, area at the time of the study. To reduce the burden on GHA staff and to increase the clustering of the sample for more efficient field work, two medical centers were se- lected to provide the sample of study subjects. The criteria for selection included the desire to have one urban and one suburban center, a need to limit the number of Federal employees selected, and the requirement that the sample include an oversample (compared with the na- tional population) of black persons. GHA records in- cluded other person-level information required for sample stratification (see next section), but did not include system- atic notation of members’ race. Thus, the oversample of black persons was affected by selecting medical centers in communities with a high proportion of black residents. The analytic sample turned out to be predominantly black persons: The list sample was two-thirds black persons, and the supplementary sample almost 70 percent black persons. Explicit stratification of list sample A sample intended to produce 1,000 completed inter- views with medical record data was selected from three list frames: persons having a recent ambulatory care visit (60 percent); persons having a recent hospital stay (20 per- cent); and persons from the general membership rolls (20 percent). Individuals from the three lists were strati- fied by demographic characteristics (age and sex), by recency of ambulatory visit (ambulatory care frame only), and by employer group (Federal Government or not). Sampling rates within each stratum followed the guide- lines described later. Although some separate analyses were planned for each of these subgroups, the intent was not to create a fully crossed design for analysis but to ensure that list-sample persons included appropriate rep- resentation by these key characteristics. For stratification by age, two major groups (those 18-64 years of age and those 65 years and over) were broken into four age categories typically used in NHIS reports of chronic condition prevalence: 18-44 years, 45-64 Selection of list-sample Sample limited to adults (18 years and over) persons Stratification of sample by age Oversampling of persons 65 years and over those aged 45-64 years Limitation of number of Federal employees Design element Description Purpose or reason Study population Members of a Washington, D.C. area health To allow a full study design maintenance organization Selection of medical Two of nine centers selected from which to draw study ~~ To reduce burden on health maintenance organization staff; to centers sample cluster sample and reduce field costs; to ensure representation Among those 18-64 years of age, oversampling of Oversample of persons with recent doctor visits Oversample of persons with recent hospital stays black persons Chronic conditions to rare among children To ensure representation of all age groups To allow separate analysis by age group To ensure sufficient reports of chronic condition To guard against response bias possibly associated with Federal employment To allow analysis of reporting of doctor visits To allow analysis of reporting of hospital stays Figure 2. Summary of Health Interview Evaluation Survey sample design 12 Table D. Planned allocation of persons cooperating in Health Interview Evaluation Survey, by sex and age group Supplementary List sample sample Age Males Females Males Females Number of persons ANBOBS wus vmmis wns amas 500 500 374 374 O-17YeAIS. « wu swiss memsns si - — 125 125 18-64 years 18-44Y08I8. . oc .v svn amin 146 146 72 72 ABBE YORB, vo vv 0 vn 3 hw nv 187 188 93 94 65 years and over BB=PAYBBIS. «. iv: + 5 i 00 v0 x 1p 7 1m 100 100 50 50 75Years BNC OVEY « « «vs vs «4 + 67 66 34 33 years, 65-74 years, and 75 years and over. The sample was divided between the two major age groups so that each would be expected to yield at least 40 reports of the 10 most prevalent chronic conditions for that age group. Thus, persons 65 years and over were selected at a higher rate than those persons aged 18-64 years. Within the younger group, persons 45-64 years of age were selected at a higher rate than those in the general population to increase the number of chronic condition reports expected for the overall group. Within the older group, persons 65-74 years of age and those 75 years and over were selected so as to be represented at the same rates as in the general population. Equal numbers of males and females were selected in each age group. Table D presents the planned sample allocation by age and sex. The first two columns represent the planned allocation for the list sample; the next two columns represent estimated yields from the supplementary sample of other household mem- bers. A “household” is defined as one or more families sharing common cooking facilities. The distribution of the analytic sample by age and sex is shown in table E. Again, the first two columns represent list-sample persons and the second two represent other Table E. Actual allocation of persons cooperating in Health Interview Evaluation Survey, by sex and age group Supplementary List sample sample Age Males Females Males Females Number of persons Alages .:rnsavsuracsnsnses 460 545 310 393 O-17Y88IS. iin tn rah dh 00m 3 - - 147 138 18-64 years 18-34 Y88IS., . cc corns mrinns 145 164 69 104 ASBL YOBISI. «cx vio #10 2 nw 4 wie 4 171 202 50 88 65 years and over BE-TAYBBIB. + «vv xs vv vieves 85 108 30 41 75yearsandover. .........: 59 71 14 22 household members. The list sample produced more fe- males than males, partly because of higher nonresponse among males and partly because of the difficulty of iden- tifying sufficient numbers of older men from GHA visit logs. The larger number of females in the supplementary sample may be the result of the household composition patterns of the Washington, D.C., area, a greater likeli- hood that female household members would be available and willing to sign permission forms, and perhaps a greater likelihood that female household members would also be GHA members. Although the primary objective of the study was to evaluate the validity of patient reports of medical condi- tions, secondary objectives related to the validity of re- ports of the number and timing of doctor visits and hospitalizations. A random sample of the GHA member- ship would be unlikely to yield sufficient reports of doctor visits within the 2-week NHIS reference period or hospi- talizations within the 13-month reference period for mean- ingful analysis. Therefore, the study design oversampled persons known to have had visits or stays within the appropriate periods. To identify persons with recent doctor visits, a sample was drawn weekly from primary care encounter forms filled out for each patient visit. For persons with hospital- izations within the past 13 months and for a general sample of persons with neither recent doctor visits nor hospitalizations, GHA’s central records system provided the sampling frame. The sample of persons with recent doctor visits and recent hospital stays was further stratified so that approximately equal numbers of persons would recall visits or stays over given time intervals. Age and sex strata were also imposed within the utilization groups, as was an upper limit of one-third of the sample representing Federal employees on each list frame. This sampling strategy did not guarantee mutually exclusive groups as some persons selected from the gen- eral rolls could have visited a medical center between the time they were selected and the time of the interview. In addition, persons selected because of recent doctor visits may also have had hospital stays within the reference period and vice versa. Sampling procedures did not allow the selection of a given person more than once either within or across categories. However, probabilities of selection were not calculated, and the sample was not weighted for analysis. Table F presents the planned allocation of the list sample by event history (that is, by whether the person had a recent doctor visit or hospital stay). The supplemen- tary sample was expected to fall largely in the “persons with neither” category. The reporting of the number and timing of medical events is subject to recall error of various kinds. Two complementary kinds of recall error are forgetting and “telescoping,” or drawing in events from outside a refer- ence period. The study design, as described, would allow analysis of forgetting or of misplacing an event within the reference period. It would not allow any meaningful 13 Table F. Planned allocation of list-sample persons cooperating in Health Interview Evaluation Survey, by event history, age, and sex Persons Persons with with Persons All recent hospital- with Characteristic persons visits izations neither Number of persons All age groups, both sexes . . 1,001 601 200 200 Age 18-44years. ........... 292 176 58 58 45-64years. . .......... 375 225 75 75 BETA YORE. , «vv» vk wn 200 120 40 40 75yearsandover........ 134 80 27 27 Sex MAB cova usasamyemyas 500 300 100 100 FEMald. cu ov swims wumass 500 300 100 100 Table G. Planned allocation of list-sample persons cooperating in Health Interview Evaluation Survey, by event history revised to analyze telescoping, age, and sex Persons with hospitalizations Persons with recent visit Sm————| —————— . EE (}])S All 0-2 2-4 0-13 14-19 with Characteristic persons weeks weeks months months neither Number of persons All age groups, both BOXES . vv vw ve 1,001 400 201 134 66 200 Age 18-44 years. ..... 292 117 59 39 19 58 45-64 years. . . . .. 375 150 75 50 25 75 65-74 years. . . . .. 200 80 40 27 13 40 75 years and over . . 134 53 27 18 9 27 Sex Male comme wae 500 200 100 67 33 100 FOmMalay, «vv vw wwe 500 200 100 67 33 100 analysis of the extent to which telescoping affects NHIS reporting of medical visits and hospital stays. To analyze forward telescoping, the sample of recent doctor visits was extended to include patients who had visits just outside of the reference period, in the preceding 2 weeks. Similarly, the sample of persons with recent hospitalizations was extended to include the 6 months before the 13-month reference period. This strategy resulted in the allocation presented in table G. Again, the categories are not mutu- ally exclusive because persons may visit the doctor in both 2-week periods. The actual analytic sample was affected by slippage in the field period for many cases. That is, persons selected because of a doctor visit within the 2-week reference period were often not interviewed in the designated week, and the reference period shifted. Tables H and J present, respectively, the actual list and supplementary samples available for analysis by event history as noted in the medical record. 14 Sample selection and response rates The sample design described in the previous section is summarized in table K. Implementing the sample selec- tion and obtaining cooperation from selected persons was a multistep process. GHA required an initial passive informed-consent process before releasing members’ names for contact. Thus, the initial step in obtaining cooperation was to mail letters to all qualifying members of the two GHA medical centers selected for the study. The letter included a return postcard that members were to send to GHA if they did not want GHA to release their names to the study. Thirteen percent of notified members returned these postcards. The sample cases were selected and fielded over 26 weeks beginning in June 1990. Each week, a sample of recent doctor visits and recent hospitalizations and a general (“no utilization”) sample were fielded. The recent- utilization cases were stratified so that equal numbers were from the previous week and from the week before, and equal numbers were from each of the preceding 2 weeks. Thus, each interview wave included members from all sampling cells, with the timing of recent-visit and recent-stay groups spread across the reference periods and the extended reference periods for analysis of tele- scoping. Interviewers were expected to complete their assignments in each wave within 1 week; however, a number of cases in each wave slid into the second week or later. NHIS rules indicate that such “holdover” cases have the reference period updated; the HIES followed this procedure. Because the HIES sample was an unclustered list, as opposed to the NHIS clustered-area sample, HIES interviewers fared worse than their NHIS counterparts in completing interviews during the assigned weeks. Interviews were conducted with list-sample persons and any household members who happened to be present. Following NHIS procedures, proxy responses were ob- tained for household members not present during the interview. At the conclusion of the interview, list-sample persons and any household members also belonging to GHA were asked for written permission to abstract infor- mation from medical records. Second permission forms were later required for certain patients with medical problems of a sensitive nature; these were requested by mail. A total of 1,846 household members were identified in 1,077 interviews. Of these, 1,312 were reported as GHA members. Only limited followup for permission forms was attempted for household members not available when the list-sample person was asked to sign a permission form. Of the 733 household members who did sign permission forms, medical records data were not obtained for a total of 70 persons; of these, 54 required further followup beyond what the schedule or resources would permit, 11 refused second permission forms, and 5 turned out not to be GHA members. Table H. Actual number of list-sample persons available for analysis, by event history from the medical record, age, and sex Persons with recent visits Persons with hospitalizations 2-4 weeks but 14-19 months but Persons Characteristic All persons 0-2 weeks not 0-2 weeks 0-13 months not 0-13 months with neither All age groups, both sexes . ...... 1,005 433 Age 18-44vyears. ................ 309 116 ABBA YBBIB. «vv vs xn vv s vu www» 373 164 BB-TAYBAIS . ov vos mivits bods gm os 193 86 75yearsandover............. 130 67 Sex Male . .cnnmmvvmv mmm wanes 460 187 FOINBIB : vows caus THANE S500 4 545 246 Number of persons 233 145 51 287 7 27 16 105 73 58 16 109 50 33 9 51 33 27 10 22 114 73 21 133 119 72 30 154 NOTE: Columns add to more than total because of overlap between persons with visits and persons with stays. Table J. Actual number of supplementary-sample persons available for analysis, by event history according to medical record, age, and sex Persons with recent visits Persons with hospitalizations 0-2 2-4 weeks but 0-13 14-19 months but Persons Characteristic All persons weeks not 0-2 weeks months not 0-13 months with neither Number of persons All age groups, both sexes . .... 703 103 79 18 10 512 Age 0-17vyears. . .............. 285 29 35 5 2 217 18-44.¥8I8, 1» 5 ssw i sus wa se 173 21 12 2 1 139 45-64years. .............. 138 27 24 5 6 85 BE=74Y0BIS. « «+ vrs crn bi 71 14 6 3 0 51 75yearsandover........... 36 12 2 3 1 20 Sex Male ................... 310 40 38 8 6 227 Femal, ; vo smvamsmemmowsns 393 63 41 10 4 285 NOTE: Columns add to more than total because of overlap between persons with visits and persons with stays. Table K. Number and percent of initial draw and response rates for Health Interview Evaluation Survey, by utilization group Utilization group Recent doctor visit Recent hospital stay No recent utilization Item Total Number Percent Number Percent Number Percent Initial draw. . . «o.oo 1,615 1,132 ad 277 ce 206 oc. Locatingrate . .......... ... rn bk 0.96 ak 0.93 ces 0.93 Number located . ........... 1,540 1,090 258 en 192 Number ineligible . . ......... 130 70 34 cen 26 Interview requested . . . ....... 1,410 1,020 Gi 224 sue 166 wine Interview response rate. . . ..... Bt 2s 0.76 NE 0.78 EE 0.77 Permission form requested . . . . . 1.077 775 174 wis 128 Cooperation rate for permission JOINS 5 va nto mmm cn wm os ie Pha 0.94 she 0.95 a 0.96 Usablecases . . ............ 1,017 728 166 PR 123 NOTE: Twelve additional cases were dropped because the respondents refused to sign a second permission form required by Group Health Association for certain patients. Table K presents the number and percent of list- sample persons at each stage of the locating, interviewing, and permission form process. The refusal rate was higher than anticipated (all interviews were conducted in metro- politan Washington, D.C., a traditionally difficult area in which to interview), but the locating and permission-form rates were somewhat higher than expected. The selection rates were adjusted during the field period in response to the slippage in reference weeks described earlier so that more persons than originally anticipated were selected in 15 the “recent doctor visit” group. (See tables G and H to compare the effective sample against the original alloca- tion by event history.) Data collection forms Questionnaire The selected GHA members were administered the NHIS core questionnaire with several modifications. Al- though the sampled GHA members were selected as individuals, the NHIS questionnaire is a household inter- view. Thus, the interview included the households of the sampled individuals. The NHIS core interview includes the following sections: ® Household composition: names of all household mem- bers, relationships, ages, full-time active duty, hospital probe. e® Limitations of activities: current limitations and under- lying conditions. All conditions mentioned are re- corded for later review in condition sections. ® Other: ongoing list of conditions, other information required for administering interview. ® Restricted activities: restrictions of activities (days missed work, school, or work around the house, days in bed, cut-down days) and underlying conditions for the previous 2 weeks. Conditions recorded as previ- ously described. ® Doctor visits: number of doctor visits or phone calls to doctor in previous 2 weeks. ® Doctor visit details: details of doctor visits reported in previous item, including condition necessitating visit. Conditions recorded under “Other.” ® Health indicators: other accident or injury in previous 2 weeks, total bed days and doctor visits in last 12 months, perceived health status, height, weight. ® Condition lists (one per interview): (1) skeletal, mus- cular, skin disorders; (2) hearing, vision, or speech impairments; (3) digestive conditions; (4) glandular, anemia, nervous system, genitourinary system disor- ders; (5) heart and circulatory system problems; (6) respiratory system disorders. ® Hospital page: details of each hospital stay reported in previous 12 months (since “13-month hospital date”), including entering condition, operations, and name of hospital. eo Condition page: details of each condition reported in “Limitations of activities,” “Restricted activities,” “Doc- tor visit details,” “Condition lists,” and “Hospital page.” ® Demographic background: information including mili- tary experience, education, race, ethnicity, employ- ment status, marital status, income, father’s last name, and social security number. Three kinds of changes were made to this core inter- view for the HIES: 16 ® The six categories under “Condition lists” were abridged and condensed into one list asked of every respondent. ® To assist in matching visits reported by household respondents with visits in the medical records, ques- tions on the location of each visit were added to the “Doctor visits details” section. e® The HIES household composition put the list-sample person in the first column and collected relationships to this person. The selection of chronic conditions from the six NHIS lists was based on the expected prevalence of the condi- tions in the sample. The 10 most prevalent conditions in each adult age group (according to NHIS estimates) were included, with the goal of having at least 40 reports of each condition from the list sample for analysis. Groups of conditions, such as heart conditions and impairments, were all included, even if some did not meet the preva- lence rules, because it was felt that some conditions might be reported in response to a probe for another condition in the group. For example, a hearing impairment (which met the quantitative criteria for inclusion) might be re- ported in response to the probe for deafness (which did not meet the quantitative criteria), leading to the need to have both conditions included in the HIES list. The HIES questionnaire is included as appendix I. Medical records abstraction form Most abstraction of medical records was done from photocopies of the past 3 years’ records (before the interview date) from an individuals file. A direct data entry form was tested, but the abstractors preferred a paper form, largely because of the time required to type condition names. The purposes of the abstracting were to identify all medical conditions and impairments men- tioned in the record and to identify doctor visits and hospital stays within the relevant reference periods. The medical records abstracting form and instructions for abstractors are presented as appendix II. To limit the amount of resources required for abstract- ing, encounter-specific information was abstracted for a limited number of medical encounters —only those felt to have direct relevance to the planned analysis. Specifically, encounters abstracted included the following: ® Any encounter within 2 months ‘before the interview date. ® The most recent encounter, if there were none within the 2 months before the interview. ® Any health assessment (comprehensive medical checkup) within 3 years before the interview date. ® Any hospital stay within 19 months before the inter- view date. In addition, abstractors recorded and coded all medical conditions from the entire 3-year medical record, regard- less of the type or date of encounter for which it was noted. Supporting materials Materials were developed or adapted from NHIS materials for a variety of data collection support purposes, including: ® Advance letters— GHA required an initial postcard mailing to all members of the selected medical centers for informed consent. An advance letter from NCHS was sent to sampled persons. ® Labels, logs, and assignment materials — For each week’s wave of sample members, computer-generated inter- viewer assignment materials were prepared, including a face sheet, reporting log, and a receipt-control log, as well as mailing labels to attach to the advance letter. ® Interviewer manual —The HIES used a modified ver- sion of the NHIS interviewer manual, a comprehensive guide to the conduct of the core interview, with addi- tional sections on idiosyncratic administrative proce- dures, the purpose of the evaluation study, and how to describe the study to respondents, as well as question- by-question specifications for new questions. ® Abstractor manual — Abstracting procedures, defini- tions, and code categories were detailed in an abstrac- tor’s manual (appendix II). Data collection Selection and training of field interviewers Twenty-four interviewers were recruited for the HIES, 20 of whom had previous interviewing experience. Five of the interviewers were based outside the Washington, D.C, area and traveled there as needed to supplement the local staff. All HIES interviewers were trained as if they were new interviewers for the NHIS. An experienced U.S. Bureau of the Census trainer conducted the session, using NHIS materials that included a verbatim guide with par- ticipative lectures, practice, and exercises. Supervisory staff observed the NHIS session and conducted additional training in specific procedures, including receiving assign- ments, and contacting, locating, reporting, and submitting of completed work. The 3% —day project-specific training included orien- tation to U.S. Bureau of the Census format and question- naire conventions, training on the NHIS core interview, and training on additional questions. Some interviewers attended an additional day of training in general inter- viewer training, adapted from U.S. Bureau of the Census general training procedures. Field data collection Advance contact by mail—As noted in the section entitled “Review,” the HIES included two advance con- tacts by mail. The first was a letter from GHA mailed to all members at the two selected medical centers. It gave a very brief description of the research and included a postpaid return postcard for members to return if they did not want their name released. The second letter, from the Director of NCHS, was sent to persons selected for interview. Much of the content and language of this letter was specified by the Privacy Act of 1974 and NCHS enabling legislation. Contacting and interviewing— Unlike NHIS proce- dures, in which interviewers approach addresses from an area probability sampling frame, interviewers contacted HIES sample members directly, knowing their names. The initial contact was made by telephone (when a number was available). HIES required the sample person to be present for the interview. Other family members present could respond for themselves; the sample person an- swered for family members not present. Following the interview, the interviewer asked all GHA members in the family for written permission to review their medical records. For adults, this permission could only be given by the persons themselves; for chil- dren under the age of 14 years, the interviewer requested the signature of a parent or guardian. For children aged 14 to 17 years, the interviewer asked for the signatures of both the child and a parent or guardian. Interviewers attempted return visits to obtain permission forms for household members (supplementary sample) not available at the time the interview was conducted. Data collection schedule and staffing — Interviewing was conducted over 26 weeks, from June to November 1990. Interviewing went on longer than expected for a variety of reasons. Because the sample was a list of individuals, contact was required with the particular person selected, which is more time-consuming than simply interviewing a household informant. Second, the sample was not geo- graphically clustered, resulting in further inefficiency in interviewer time. Third, GHA addresses were not always up to date, which sometimes required locating. These factors affected the average time per completed case as well. The estimate of 4.75 interviewer hours per completed interview used in planning the study was quite low; the actual overall average was just under 7 hours per case. As interviewers tend to work part time regardless of how much work is available, the level of staffing was less than optimal, because staffing estimates were based on the lower number of hours per complete interviews. The interviewers reported to a field supervisor, who in turn reported to the field director. The field supervisor discussed each interviewer’s workload at least weekly and oversaw quality assurance measures. Field quality assurance measures — Following the NHIS model, interviewer performance was measured in three ways: review of hard-copy interviews and feedback; obser- vation of interviews; and verification reinterviews. Each interviewer’s first two completed cases, and 10 percent of his or her cases thereafter, were thoroughly reviewed by inhouse staff. The reviewer completed a feedback form detailing both good performance and per- formance requiring improvement. The field supervisor reviewed these forms with the interviewer within 1 week of 17 receipt of the cases. Interviewers not meeting minimum performance standards were retrained or dismissed. Each interviewer was observed in person twice during the field period by an experienced observer. The observer followed each interview carefully, noting examples of both good and bad performance, and reviewed the results with the interviewer after they left the household. The field supervisor conducted a short (3 minute) reinterview with 10 percent of respondents, using the standard NHIS verification interview. Verification inter- views for households without telephones were conducted in person. Comparison of verification interviews with the completed work revealed no evidence of falsification. Abstracting medical records All signed permission forms were sent to GHA, which in turn copied the corresponding medical records, going back 3 years. Medical records abstractors then reviewed the records and identified and recorded all medical condi- tions and relevant encounters noted in the records. Receipt of records The copying process took much longer than antici- pated. GHA staff were very busy, and the records were often very long. A variety of problems were encountered that prevented or delayed the copying of records. These problems included: @ Some household members signing permission forms were not GHA members or had lapsed memberships. ® Some records were temporarily unavailable when sought. ® Some records included referrals to mental health ser- vices, drug or alcohol treatment, or acquired immuno- deficiency syndrome (AIDS); for such persons, GHA requested an additional permission form specifically acknowledging the sensitive information. When this additional permission was refused, no information was obtained. Abstracting process Abstracting procedures are described in appendix II. The abstractors were experienced at abstracting from medical records and condition coding. The staff partici- pated in development of the abstracting form and wrote the manual. Because of delays in receiving the records from GHA, some of the abstracting work was subcon- tracted. The subcontracted work was subject to the same quality control and strict confidentiality procedures as that performed by Westat Inc., and they reviewed all subcon- tracted work. Abstractors recorded information on abstract sheets for each case. For 10 percent of the cases, a second abstractor then reabstracted the record, noting discrepan- cies as they were discovered. A third member of the 18 abstracting staff acted as arbitrator, working with the first two abstractors to resolve discrepancies. Data preparation and processing Data preparation activities for the HIES included medical condition coding, other coding of hard-copy ques- tionnaires from the field interviews, retrieval of missing or ambiguous critical items, key entry of the questionnaires and abstract forms, and machine edits of all study data. Condition coding Conditions from both the household interview and the medical record were coded according to the International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM), as modified by the NHIS. All condition coding was subject to 10-percent recoding by a second coder. A third coder then worked with the first two coders to resolve discrepancies. Coding data into machine-readable format NHIS coding specifications were applied to the house- hold interview data. Coders reviewed all hard-copy ques- tionnaires for legibility, missing or incorrectly entered data, and assignment of numeric codes to any non- numeric values (such as “2” for “Feb” or “other — specify” fields). Coders received a full day of training, and 100 per- cent of their first batches (about 25 cases) were reviewed. Ten percent of subsequent batches were reviewed. Retrievai of missing or ambiguous data Project staff identified critical items for analysis as part of the coding specifications. When coders found missing or ambiguous responses in these critical fields, the cases were flagged for retrieval. The field supervisor conducted data retrieval, including recontacting of inter- view respondents. Keying and verifying data File layouts were prepared and keying procedures developed for the survey questionnaires to ensure compa- rability with the NHIS. All key entry was 100-percent verified and adjudicated by a keying supervisor. Abstract forms were also key entered, with 100-percent verification and resolution. Editing and correcting computer files After data were keyed and verified, they were com- puter edited against an exhaustive set of machine specifi- cations. The specifications were adapted from those used by the NHIS to run on Westat software. The cleaning specifications included skip, range, and logic checks. Cod- ers reviewed all fail-edit cases and made appropriate corrections to the data file. The edit specifications were rerun until all discrepancies were resolved. Coders also reviewed frequency distributions for all data items as a final step in machine editing. Machine editing was also performed on the medical records data. A series of skip and logical checks paralleled those done on the household data. Discrepancies were resolved when possible by the abstracting or analysis staff. Discussion The current NHIS method for obtaining reports of chronic conditions includes collecting the names of condi- tions associated with restricted activity, limitations of activity, doctor visits in the 2 weeks before the interview, or hospital stays in the 13 months before the interview. Following these indirect probes, the interview presents the respondent with one of the six condition checklists. The checklist asks directly: “Does anyone in the family NOW have . ...” Depending upon the condition, the reference period is “now,” “ever,” or “the last 12 months.” Inter- viewers record “yes” responses by writing the condition name on the interview booklet, alongside any conditions reported earlier through the indirect probes. Later in the interview, the interviewer asks more detailed questions about each of the conditions he or she has recorded. In data preparation, medical coders review the re- ported condition names and the other information about that condition, and assign a code from the NHIS adapta- tion of the ICD-9-CM. One condition report may lead to multiple codes, and duplicate reports of the same coded condition are collapsed. Coders also determine whether any given condition is chronic or acute, adding an indica- tor to the modified ICD-9-CM code. Many conditions, such as diabetes and hypertension, are “chronic by defini- tion,” that is, they are always coded as chronic. Conditions not defined as chronic are considered chronic if they last 3 months or longer, from date of onset to the date of interview or cure. These dates are part of the detailed questionnaire condition section. A series of computer edits (11) ensure that reported conditions identified as chronic in the data file meet the NHIS definition for chronicity. The NHIS produces prevalence estimates for chronic conditions defined by groups of NHIS-modified ICD- 9-CM codes; these groups are referred to as “recode C” codes. For this report, recode C is referred to as the “NHIS recode.” Note that there is no one-to-one match between the condition checklists in the questionnaire and the NHIS recode conditions for prevalence estimates and that responses to the checklist do not define whether a person has a particular NHIS recode condition. For exam- ple, a person may say “yes” to the checklist probe for dermatitis but give information in the condition section that leads to an ICD-9-CM code outside the NHIS recode group for dermatitis, such as athlete’s foot. On the other hand, a person may say “no” to the checklist probe but report a condition coded into the NHIS recode group as the reason for a 2-week doctor visit. To compare reports of chronic conditions between interview and medical record data, it would be desirable to adapt the NHIS procedures to the use of medical records. Procedures for abstracting medical records are detailed in appendix II. Abstractors recorded all condi- tions mentioned in 3 years’ worth of medical records, using the NHIS adaptation of the ICD-9-CM, except for the chronicity indicator, which was not used. The ab- stracted conditions were then collapsed into NHIS recode groups. The interview and medical record procedures differ in potentially significant ways. First, the medical record re- view includes no stimulus comparable to the interview’s condition checklist. A more comparable procedure would be to ask a physician who had examined the person, “Does this person have . . . ?”” This approach was used by the Baltimore and Hunterdon County studies (31,5). The absence of such a checklist may result in failure of the medical record to confirm accurate interview reports. For example, conditions may have been noticed by an attend- ing physician or other medical professional but not en- tered in the record. A second problem of comparability occurs in the timing of condition reports. As previously noted, the household interview asks about “now” for some condi- tions, or “ever” or “in the past 12 months” for others. Medical records are dependent upon when a person seeks care and often do not include information on duration. Thus, medical record reports are subject to error for “now” conditions because the person may or may not have been seen at a time near the interview date. “In the past 12 months” may also be difficult: Positive reports within the 12-month period are fairly clear, but a positive report occurring only outside the 12 months is likely not to have information on whether the condition continued to be present into the 12-month period. The only systematic review of this problem instituted for the HIES was for cataracts, where mention of cataract surgery more than 12 months before the interview date was not considered “cataracts in the past 12 months.” A third problem lies in the NHIS definition of “chronic.” As already noted, some conditions are “chronic by definition.” This rule is straightforward in the medical record context. However, the “chronic by duration” rule is not easily transferrable. One could derive rules for coding conditions as chronic by duration (if, for example, the medical record documented two encounters about the same condition separated 3 months or more), but these rules would not be comparable to the NHIS interview, in which duration is asked about specifically. The HIES procedures counted all conditions mentioned in the med- ical record as chronic, except those that most obviously were not (i.e., sunburn and poison ivy under “dermatitis’). A fourth problem in comparing prevalence from med- ical records with that from household interviews is refer- ences to a “history of” a condition in the record. This problem ties in with the second one previously mentioned — the timing of the condition report. For those conditions 19 about which the interview asks “ever had,” “history of” is a comparable indication of the presence of the condition. For the “now” or “past 12 months” conditions, however, the correct treatment is less clear. HIES procedures excluded “history of” references for all conditions subject to these time frames. It appears that prevalence estimates based on medical records would be different in many cases from those based on household interviews. There is no uniform and easy answer to the question of where truth lies. This analysis uses the methodological differences described here as well as other factors to help explain differences by condition in reporting between the household and medical record. A second analysis examines differences in agreement be- tween the household and medical record by characteristics of the subject. The assumption in the latter analysis is that more agreement means better reporting in the interview, but the medical record is not routinely viewed as a validation mechanism for the interview report. Analysis methods As noted earlier, the HIES design was intended to allow evaluation of the reporting of chronic conditions as well as ambulatory medical care visits and hospital stays. Because this report focuses on the reporting of chronic conditions, this discussion of analysis methods is limited. Future reports will describe other aspects of the study. Once the medical record data were coded, edited, and the condition data reclassified into the NHIS recode, the two sources were compared for each of 23 chronic condi- tions, person by person. The checklist included probes for more than 23 conditions, and orthopedic impairments and deformities were collapsed into one category for HIES analysis. Blindness and other visual impairments were combined, as were deafness and other hearing impair- ments. Rheumatism and congenital heart disease were 20 Table L. Matrix for matching interview with medical record reports of chronic conditions Condition mentioned by respondent in interview Condition noted in medical record Yes No YO8 wv vnsmssnrws A Cc NBuows sismeemems B D NOTES: A is positive match, B is false positive, C is false negative, and D is negative match. Prevalence by Health Interview Evaluation Survey report calculated as (A + B)/N. Prevalence by medical record calculated as (A+ C)/N. Kappa value calculated as 2((AxD) — (BxC))/((A+B)x (B+D)+(C+D)x(A+C)). excluded from the analysis because of very low prevalence among the study sample. For each condition, a person was classified into one of four cells as shown in table L, depending upon whether the condition was present in the interview file and the medical record file. Multiple condi- tions for a person in one NHIS recode classification were counted the same as a single condition. Prevalence for the analytic sample was calculated using the formulae in the notes of table L. Comparison of prevalence from the two sources is an aggregate measure of agreement; it says nothing about how well individual cases matched. The Kappa statistic was used to analyze the agreement at a person level. Kappa is a weighted proportion, with possible values ranging from -1 (perfect disagreement) through 1 (perfect agreement). Regarding Kappa as a measure of interrater reliability, Landis and Koch (20) suggest that values of less than 0.4 represent poor-to-fair agreement, 0.4—-0.6 moder- ate agreement, 0.6-0.8 substantial agreement, and 0.8-1.0 almost perfect agreement. It is possible for two sources to produce identical prevalence with very low agreement at the individual level. If the medical record was considered “the truth” in such a case, the rates of interview overre- porting and underreporting would both be high and about the same. Results Survey statistical differences As described earlier, the HIES sample was drawn from the membership rolls of an HMO in the Washington, D.C., area. Oversamples were drawn of older persons, those with recent hospital stays and those with recent doctor visits. In addition, the sample was by design drawn from medical centers serving communities with large black populations. Therefore, black people are much more heavily represented in the HIES sample than in the general population. These features of the HIES design limit direct comparisons to the NHIS. As shown in table 1, the HIES sample is older and contains more black people, people of higher income, and more highly educated people than the U.S. population. Because of the oversample of persons with recent medical utilization, the HIES sample is also probably sicker than the U.S. population. In addition, the sample is limited to one geographic area, and all HIES sample persons have health insurance coverage and access to basic health care services. The combination of these factors, some of which can be controlled for in examining HIES results and some of which cannot, should have significant effects on preva- lence of chronic conditions derived from HIES household interviews when compared with the same rates derived from the NHIS. Also, NHIS prevalence estimates are weighted to the U.S. noninstitutionalized civilian popula- tion; HIES data are presented unweighted. The prevalence of chronic conditions reported in the interview for list-sample persons is about twice that of the general U.S. population according to the 1989 NHIS. Adding household members (the “supplementary sam- ple”) reduces the differential to 60-percent greater preva- lence overall in the HIES. Much of this difference is attributable to the HIES oversample of persons 65 years of age and over. Tables 2-5 compare the NHIS and HIES prevalence of the chronic conditions studied by age group. Column 1 of these tables presents the weighted NHIS estimates (prevalence per 1,000 population) for the partic- ular age group for 1989, and column 2 shows the compa- rable prevalence rate for the HIES, including both list- sample persons and household members. Column 3 is the raw difference between the two rates, and column 4 is the percent difference between columns 1 and 2. In table 2, for example, the HIES prevalence of arthritis is 13 persons per 1,000, or 27 percent higher, than the NHIS estimate. The “All conditions” row shows, in columns 1 and 2, the total number of conditions listed that are reported per 1,000 persons by the NHIS and HIES, respectively. Column 4 of tables 2-5 shows that the HIES had an overall prevalence of the selected chronic conditions that was 41 percent higher than the NHIS for persons 18-44 years of age, 27 percent higher for those aged 45-64 years, 14 percent higher for persons aged 65-74 years, and 11 per- cent higher for persons aged 75 years or over. This decrease by age may be related to oversampling for HIES persons with recent doctor visits (assumed to be sicker); the effect of this sampling strategy on prevalence may be lessened with older persons. Some of the remaining dis- crepancy between the two sources is attributable to partic- ularly high rates of hypertension and diabetes, which are more prevalent among black people than white people, because the HIES sample is disproportionately composed of black people. However, the NHIS shows lower preva- lence of most other conditions studied here for black persons. In summary, the HIES sample has reported more chronic conditions than one would expect from a nation- ally representative sample, even when HIES age and race oversampling are taken into account. The list sample’s greater likelihood of having a recent doctor visit probably accounts for some of this difference. Also, all HIES sample persons are insured and have access to health care. People who do not seek medical care may be less likely to report chronic conditions that require professional diagno- sis. The effects of geographic clustering and differences in procedures between the HIES and NHIS are unknown, but procedural differences (summarized in figure 1) were minimized and probably have little effect. Interview reporting compared with medical record, by condition Tables 6 and 7 present the results of matching inter- view and medical record reports for the 23 chronic condi- tions studied for list-sample persons and household members, respectively. The conditions are arranged in order of their NHIS recodes, with summary lines for heart conditions in general and for heart rhythm disorders. Following the typology of table L, the tables show num- bers of cases falling into positive match (type A), negative 21 match (type D), apparent interview overreport (type B), and apparent interview underreport (type C). They also present prevalence calculated from HIES interview and medical records using the formulae in table L, and com- pare these rates by showing net and proportional overre- porting by the interview as opposed to the medical record. Finally, the tables present the Kappa values describing person-level agreement between the interview and medi- cal record. Tables 8 and 9 present the same data, without the heart condition summary rows, in descending order of Kappa values. As shown in tables 6 and 7, about two-thirds of the conditions were overreported in the HIES interview, and these were divided into two roughly equal groups (by proportional net overreport among list-sample persons) of conditions overreported by 200 percent or more and con- ditions overreported by about 100 percent or less. Kappa values ranged from around 0 to about 0.82 for list-sample persons and were generally slightly lower for household members. In tables 8 and 9, the conditions with the highest Kappa values are all underreported by the interview. The proportional differences from tables 6 and 7 are presented in figure 3, with the conditions arranged from highest to lowest proportional net overreporting. Figure 3 reveals a fairly consistent pattern of higher interview reporting versus the medical record for list-sample per- sons than for household members. The interview reported noticeably higher rates for list-sample persons for varicose veins, allergic rhinitis without asthma, chronic sinusitis, constipation, and orthopedic impairment. Higher relative rates for household members were reported for heart murmurs, migraine headache, hardening of the arteries, hemorrhoids, and ischemic heart disease. Although the HIES did not include an experimental design to examine the effects of proxy reporting, it is reasonable to speculate that much of the difference between list-sample persons and household members in the proportion of overreport- ing by condition is attributable to proxy response, because all list-sample persons were self-respondents, but many household members were not. Comparing proportional overreporting controls for many of the artifactual differ- ences between the two populations—the list-sample Heart murmurs v] List=sample persons Other household members Tinnitus Varicose veins Allergic rhinitis Chronic sinusitis Visual impairment Hearing impairment Migraine headache Constipation Orthopedic impairment Arthritis Atherosclerosis Hemorrhoids Chronic bronchitis Tachycardia High blood pressure Asthma Diabetes Ischemic heart disease Cataracts Dermatitis Other heart disease Other heart rhythm disorders Percent under reported 1 1 1 1 1 1 100 150 200 250 300 350 Percent over reported Figure 3. Percent overreport by interview compared with medical record in the Health Interview Evaluation Survey 22 persons are older, sicker, and have more recent doctor visits than the household members. Previous studies have used other measures of agree- ment. In particular, Harlow and Linet (17) accumulated findings from a number of studies using only positive responses to measure agreement. (They also presented percent agreement and Kappa values where these could be calculated.) Their measures were “percent of positive reports in records matched by interviews” (A/(A+C) in the terminology of table L) and “percent of positive re- ports in interviews matched by records” (A/(A+B) in table L terms). Similar measures were used by Madow (2) as “rate of underreporting” and “rate of overreporting,” respectively. Table 10 compares the Harlow and Linet statistics for the Health Insurance Plan (HIP) and Kaiser Permanente (KP) studies and the HIES. It also calculates the HIES measure “net overreporting” from these statistics. Ta- ble 10 contains striking similarities (between KP and HIES for arthritis and hearing impairments and between KP and HIP for asthma) and striking differences (between all three studies for visual impairments). The three studies were done in HMO settings but in different geographic areas, in different times, and using somewhat different procedures. For example, the HIP and KP studies used 1 year’s medical records, but the HIES used 3 years’; the KP study included experiments in questionnaire design, the HIES and HIP did not. The NHIS questionnaire and coding procedures changed considerably between the times HIP, KP, and HIES studies were conducted. The HIP and KP studies used the seventh revision of the International Classification of Diseases, but the HIES used the ninth; the chronic condition recodes and condition checklists on which the HIES was based were considerably expanded following the time of the earlier studies. The HIP and KP studies used one recode for most heart conditions, but the HIES used six of the eight current NHIS recodes. The KP study combined asthma and allergic rhinitis, and the HIES separated them. Thus, the comparison of specific condi- tions across these three studies must be done with care and must be limited to fairly broad generalities. Possible reasons for mismatches Tables 6 and 7 indicate considerable variation in the rates of agreement across conditions, as well as in the differences in prevalence estimates between the sources. There are several plausible explanations for both type B and type C mismatches. Possible reasons for interview reports not confirmed by the medical record (mismatch type B) include (with examples from the list in tables 6 and 7): ® Medical treatment was not sought for the reported condition, either because it was not thought to be serious, the person was averse to seeking treatment, the condition was felt to be embarrassing, or the condition (e.g., hemorrhoids, constipation, chronic si- nusitis, hearing impairment) was treated by patent medication or other nonprofessional means. ® The reported condition (e.g., tinnitus or constipation) is a symptom of a more serious condition and is not recorded in the medical record because it was sub- sumed by the causative condition or not felt to be worthy of note. e® The reported condition (e.g., varicose veins or heart murmur) is stable and requires no ongoing treatment. ® An impairment (e.g., orthopedic) has not necessitated treatment. ® Respondents may confuse two condition names or misdiagnose a condition. Possible reasons for conditions appearing in the med- ical record but not being reported in the interview (mis- match type C) include: ® The condition name is not familiar to the patient and therefore is not remembered or recognized in the interview. e The condition (e.g., cataracts) is not salient to the respondent, perhaps because it is at a threshold level and has not caused any discomfort or worry. ® The respondent’s definition of a condition in the checklist is different from that intended by the study. ® The respondent can only describe a condition in a vague way that is not included in the NHIS definition for the prevalence estimates. ® The medical provider did not tell the patient about the condition. ® The respondent is aware of the condition (e.g., impair- ment), but denies its presence. ® The respondent does not recall having the condition because of cognitive limitations. Thus, one would expect agreement between interview and medical records for conditions that (a) are fairly well defined from both the clinical and lay perspectives, (b) require ongoing treatment, (c) have commonly recognized names, and (d) are salient to the respondent because they cause discomfort or worry. The conditions with the highest Kappa values — diabetes, high blood pressure, asthma, and ischemic heart disease (“heart attack” and angina) —meet these criteria. Condition-level prevalence The prevalence and match ratios between interview reports and medical records described so far have been conducted at a person level —that is, if a person reports one or more conditions within an NHIS recode group, that person is counted as one occurrence. However, the NHIS prevalence estimates are prepared at a condition level. Each mention of an ICD-9-CM condition within an NHIS recode group counts as one occurrence (but multiple reports of the same ICD-9-CM condition for a person only count as one). For most of the conditions examined in this report, the person-level prevalence and condition- 23 level prevalence are virtually identical. The exception is heart conditions; multiple conditions within one NHIS recode are often present for one person. Table 11 presents the prevalence for heart conditions in the study sample at both the person level and the condition level. The NHIS recode groups examined are ischemic heart disease, tachy- cardia or rapid heartbeat, heart murmurs, other heart rhythm disorders, and “other selected diseases of the heart.” Not included because of their relatively low prev- alence are NHIS recode groups for rheumatic heart dis- ease and congenital heart disease. Effect of broader condition typologies Several of the possible reasons for type C mismatches and one of the reasons for type B mismatches given earlier relate to interview respondents’ not knowing, misremem- bering, or confusing condition names. Cox and Iachan (10) found that agreement between survey respondents and medical records in reasons for visit was considerably higher for higher levels of aggregation in condition coding. Madow (3) used a “loose match” that the author admitted was “not well specified” as well as a “tight match” in comparing interview reports with physician reports of chronic conditions. The idea is that interview respondents often know generally what is wrong but cannot specify a condition in sufficient detail for agreement with medical records with relatively highly differentiated classification schemes. From the perspective of evaluating prevalence esti- mates on the NHIS, the loose-match concept has some limited applicability. The aggregation of ICD-9-CM codes into the NHIS recode for the purpose of making preva- lence estimates may be viewed as somewhat arbitrary, more so for some conditions than for others. One concept of loose match, then, is that the NHIS recode classifica- tions could be expanded to include conditions that are clinically equivalent to those in the existing group. In other words, would a typical physician classify the person as having the broad clinical entity (such as ‘“arthritis™) based upon the information available from the medical record? Constructing an extension of the NHIS recode groups under consideration here began with a review of mis- matches. Based upon other conditions reported by the interview or medical record for mismatches (types B and C), additional clinically equivalent ICD-9-CM codes were added to the following chronic conditions: ® Arthritis ® Dermatitis ® Hardening of the arteries ® Chronic bronchitis In addition, allergic rhinitis and chronic sinusitis were combined and expanded into one upper respiratory cate- gory. The analysis based on these broadly defined condi- tion groupings will be referred to as the “loose match.” Details of the loose-match condition map are presented in appendix III 24 The revised map was used to evaluate the mismatches from the NHIS recode-level analysis. First, only persons classified as type B or type C mismatches for the specified conditions were evaluated. The purpose of this analysis was essentially to evaluate the NHIS recode definitions by determining how often one source reported a chronic condition within the definitions when the other source reported a clinically equivalent but excluded condition. Then, the loose-match map was applied to the negative matches to examine the overall effect of the revised map on prevalence. The results are presented in table 12 for list-sample persons only. (Note that chronic bronchitis is not included in table 12; there were no changes as a result of the loose match for this condition.) Table 12 includes three groups of five columns. The first group presents the match classification and Kappa value when applying the NHIS recode definition of the specified condition. These columns also appear in table 6. The second group of columns, labeled “loose match 1,” presents the results of applying the expanded condition definitions to cases originally falling into mismatch types B and C. Comparing the first and second groups of columns indicates how much of the apparent discrepancy between the two data sources may be attributable to reporting equivalent conditions in different terms. The third group of columns, labeled “Loose match 2,” presents the results of applying the expanded definitions to negative matches (type D) as well as mismatches. Comparing this group of columns with the others shows the effect on prevalence if one were to use the expanded definitions in place of the NHIS recode C classification. Condition-specific results The discussion in this section is based on tables 2-5 for all conditions, on table 10 for conditions included in the earlier studies, on table 11 for heart conditions, and on table 12 for the loose-match conditions. Arthritis — When controlling for age, the prevalence of arthritis for all HIES sample persons is roughly compara- ble to NHIS estimates (tables 2-5). The largest difference is for persons 18-45 years of age (27 percent higher in the HIES). Arthritis was reported somewhat more frequently by households than it was recorded in medical records — 38 percent more for list-sample persons and 12 percent more for household members (tables 6 and 7). Agreement between the interview and medical record on the presence of arthritis was moderate (Kappa = 0.40 for list-sample persons, 0.48 for household members), perhaps surpris- ingly low for a well-known condition often treated by prescription drugs. The discrepancy between medical records and interviews appears to be the result of several factors: imprecise or erroneous use of the term “arthritis” by respondents, lack of physician visits for this affliction, physicians not recording arthritis, even if present, and the somewhat limited definition of arthritis in the NHIS recode. Persons with joint pain of unknown cause may be self-diagnosed as “arthritis.” Bursitis and tendinitis might also be reported as arthritis—these would be reporting errors. Such errors would make arthritis appear to be overreported. Medical records may not include mention of arthritis even if the patient has it. The condition may not be severe enough to be worth noting, the arthritis may never have been a reason for a visit, or the physician may have written something more specific, such as cervical radiculopathy, which is caused by arthritis, but is not considered arthritis in the NHIS recode definition. By the NHIS recode definition, arthritis includes pyo- genic arthritis, unspecified infective arthritis, crystal arthr- opathies, rheumatoid arthritis and other inflammatory polyarthropathies, osteoarthrosis and allied disorders, other and unspecified arthropathies, ankylosing spondylitis, and spondylosis and allied disorders. For the loose match, this definition has been expanded to include 13 other condi- tions that involve inflammation of the joint. They also either occur with such regular frequency that they can be considered part of the disease, as in the case of Sjogren’s syndrome, or they commonly occur as a principal result or sequela of arthritis, as in cervical radiculopathy and sciat- ica. The full list of conditions added for the loose match is: Sjogren’s syndrome, cervical radiculopathy, sciatica, spinal stenosis, neuritis or radiculitis, carpal tunnel, spondylitis, chondromalacia of the knee, periarthritis of the shoulder, costochondritis, disc disorder, lumbosacral or cervical de- generation, and gout. As shown in table 12, when applied only to previous mismatches, the loose match resulted in a 15-percent reduction in the number of type B mismatches (interview reports not confirmed by the medical record) and only a 4-percent reduction in type C mismatches (medical record reports only). Thus, it appears that interview respondents are somewhat likely to report clinically equivalent condi- tions as arthritis but fairly unlikely to do the reverse. Applying the loose match to negative matches as well as mismatches results in an overall 27-percent increase in type C mismatches but has almost no effect on type B mismatches relative to applying the loose match just to original mismatches. This result indicates that many per- sons with clinically equivalent conditions (but not arthri- tis) according to the medical record are not reporting arthritis in the interview. Only one additional positive match was created by extending the loose-match criteria to original negative matches. In summary, the loose match helps to explain some of the apparent overreporting of arthritis by interview respondents. If prevalence estimates of arthritis were to be made from medical records, an expansion of the NHIS recode might be appropriate. However, there is little evidence to support expanding the NHIS recode definition of arthritis for classifying inter- view responses. Some diagnoses may be confused with arthritis and may accompany it but are not invariably associated with it. These kinds of conditions were not included in the loose match. A case-by-case review of mismatched interviews and medical records revealed a large number of these conditions, many pertaining to restricted mobility and painful joints and backs. Tendinitis, for example, may accompany arthritis, but just as often may not be associ- ated with it. Similarly, tenosynovitis, myositis, and tendini- tis do not always, or even frequently, involve the joint. All of these conditions involve inflammation around the joint, and could be confused with the diagnosis of arthritis. Pain in joints was not considered specific enough to be consid- ered arthritis. Many type B mismatches remained type B mismatches after the loose match because they involved conditions not considered clinically equivalent. Arthritis is more prevalent among the elderly than the nonelderly, and physician contacts also increase with age. Because physician contacts increase with age, one might expect agreement between the medical record and inter- view to improve with age. However, the Kappa values vary only from a low of 0.26 for the group 45-64 years of age to a high of 0.39 for the group aged 65-74 years among list-sample persons. These are all in the poor-to-fair agreement range. Percent net overreport increases with age except for those 75 years of age and over. List-sample persons and household members reported similarly. Dermatitis — Dermatitis is more prevalent among the HIES sample than would be expected from NHIS esti- mates in all age groups except for persons 18-44 years of age (tables 2-5). Prevalence from the medical record is considerably higher than from the HIES interview for both list-sample persons and household members (tables 6 and 7), and agreement between the two sources is low (Kappa = 0.23 for list-sample persons and 0.17 for house- hold members). The presence of chronic dermatitis is perhaps the most ill-defined of any of the conditions studied. Determi- nation of chronicity (presence of a condition for 3 months or longer) is very difficult from the medical record because much of the apparent underreport may be the result of acute episodes of dermatitis in the medical record. Sun- burn and poison ivy, which are included in the NHIS recode definition of dermatitis, were excluded from the definition for classifying medical record conditions be- cause they are unlikely to last 3 months or more. As defined by the NHIS recode, dermatitis includes the following ICD-9-CM codes: 690, Erythema- tosquamous dermatosis; 691, Atopic dermatitis and re- lated conditions; 692 Contact dermatitis and other eczema; 693, Dermatitis due to substances taken internally; and 694, Bullous dermatoses. However, a more liberal defini- tion of dermatitis (inflammation of the skin), is practical. Match results using this expanded definition are presented in table 12. The conditions included in the loose match are presented in appendix III. Ten of the 33 type B mismatches (interview reports not confirmed by the medical record), but only 2 of the 81 type C mismatches (medical record report only), were matched using the loose-match criteria. These additional matches increased the Kappa value from 0.23 to 0.36. However, applying the loose-match criteria to negative 25 matches resulted in an overwhelming increase in type C mismatches. Although a substantial proportion of inter- view respondents reporting originally unverified cases of dermatitis were confirmed by the loose match, there is no evidence of interview respondents reporting dermatitis as any clinically equivalent condition in the loose match. Impairments — Impairments represent a different set of issues when comparing interview reports and medical records than morbidity conditions do. Two attributes of impairments make this true. One is that the perception of an impairment is different for the patient and the physi- cian to the extent that they may disagree as to whether the patient has an impairment. The other is that impairments do not always necessitate physician visits. This analysis included reviews of the following impair- ments: blindness, other visual impairments, deafness, other hearing impairments, and deformity and orthopedic im- pairments. Because of limitations in the HIES design and the relatively low prevalence of some types of impair- ments, the NHIS recodes for blindness and other visual impairments (201 and 202), for deafness and other hear- ing impairments (203 and 204), and for orthopedic impair- ments and deformities (228-240) were collapsed for matching and analysis. Within the deformity and orthopedic-impairment group, any code on the medical record could thus match any other within that classifica- tion on the interview. Thus, hammertoe could be matched to chronic elbow pain because these are both within the collapsed categories. As shown in tables 6 and 7, visual and hearing impairments, including tinnitus, had high net overreports in the interview, ranging from 100 percent to 250 percent for list-sample persons, and within the same range for household members. Agreement was low for tinnitus and visual impairment, but in the fair range for hearing impair- ment. Tinnitus is entirely a subjective phenomenon, but hearing impairments may be noted by physicians in the course of seeing patients for any reason. Agreement tended to be highest for persons 75 years of age or over for these conditions. For orthopedic impairments and deformities, the in- terview and medical record produced very similar preva- lence, but with very low agreement (Kappa = 0.17 for list-sample persons and 0.12 for household members) despite (or perhaps because of) the broad match criteria. Thus, the similarity in prevalence appears to be coincidental. The interview questions for impairment ask about the existence of a symptom falling into the previously named categories and about the cause of the impairment. In general, the pattern among the type B mismatches for impairment was that the impairment was reported to the interviewer but the cause was not. In many cases of type B mismatch, a probable underlying disease (cause) was on the medical record, and the resulting impairment was not on the medical record. This is particularly true for visual and orthopedic impairment where about half the type B mismatches have a probable cause on the medical record. 26 Thus, the same information may not be located in both places. Rather, the mismatches show the effect of using two different instruments —an interview and a med- ical record—to attempt to collect the same information. The interview reflects a response to a direct question about impairment. The medical record would only note an impairment if it was a reason for visit or clinically signifi- cant in itself. For example, many of the orthopedic impair- ments were probably caused by arthritis. However, the impairment was not noted on the medical record, and the cause of the impairment was not known to the patient and was therefore not reported in the interview. The cause is not always reported on the medical record because often the respondent has never seen a physician for an impairment. This pattern is notable for the hearing impairments, where there are five times more type B mismatches than type C mismatches. The type C mismatches may be the result of differing perceptions about impairment. The physician may note an impairment, but the patient may compensate for it so well that it does not seem worthy of reporting to an interviewer as an impairment. Type C mismatches may also result from different nomenclature used by physicians and pa- tients, as with all condition mismatches. As discussed earlier, it is possible that type C mismatches were the result of a report in the medical record for an acute episode, not a chronic condition. The mismatches also reflect coding instructions. For example, when the interviewer asks if anyone has blind- ness in one or both eyes, a positive response would be recoded to 201 or 202. However, the medical record would show the results of an acuity exam to be recoded as 201 or 202. The medical record might have an indication of presbyopia, the degeneration of sight because of age. Presbyopia is not matched to poor vision under the NHIS recode. The respondent may have said that “old age” caused their poor vision. However, old age is not specific enough to be matched to presbyopia. Case-by-case review of the mismatches reveals that approximately half of them would be matched if symptoms were matched to probable underlying diseases or medical conditions. However, the great number of positive- interview, negative-medical record combinations leads one to believe that impairments are frequently reported to interviewers but not to physicians. In addition, patients may have reported impairments to physicians, but the impairments were either not noted, because they were not diagnostically relevant, or were noted as “patient com- plains of ,” which would not have been coded. An interview may be a better source of data on impairment than a review of medical records. This is because an impairment reflects self-perception, which is unknown by the physician, and because the interview specifically asks about impairment, which physicians often do not. The HIES interview prevalence of impairment is gen- erally about what would be expected from NHIS estimates or slightly lower, except among persons 75 years of age and over (tables 2-5). The HIES sample’s greater access to medical care may be related to the lower rates of impairment. Tinnitus — Alone among the conditions studied, a diag- nosis of tinnitus, or ringing in the ears, is based solely on a patient’s report. Thus, one would not expect a high level of type C mismatches, that is, reports found only in the medical record. In fact, there were only nine type C mismatches for list-sample persons and two for household members (tables 6 and 7). There were also very few positive matches, but a fair number of type B mismatches (interview reports not confirmed by the medical record), leading to a low Kappa (0.17 for list-sample persons, 0.34 for household members), and indicating that tinnitus was often not reported to medical professionals or not re- corded if reported. Tinnitus was reported in the HIES at roughly the rates expected from NHIS estimates, except for persons 75 years of age and over (table 5). For these persons, the rates are almost double those of the NHIS estimates; the relatively greater extent of self-reporting for older persons in the HIES may partially explain this difference. It is likely that proxy reports of tinnitus would not be as comprehensive as self-reports. Tinnitus was reported much more often for list-sample persons than for household members (tables 6 and 7), which reflects the higher proportion of older persons in the list sample than among household members, but may also be affected by the presence of proxy reporting for household members. Cataracts — Cataracts were more prevalent by inter- view report among the study sample than would be ex- pected from NHIS estimates (tables 2-5). For persons 45-64 years of age, the relative prevalence in the HIES study sample is almost 120 percent higher than the NHIS estimate for that age group. Access to preventive care, including eye examinations, may contribute to this differ- ence; cataracts are often detected in routine exams long before they cause discomfort or loss of vision. Agreement between the interview and medical record (Kappa) was slightly higher for cataracts than the average across conditions for list-sample persons (table 6) but was somewhat lower than average for household members (table 7). In terms of agreement, cataracts seem to be more poorly reported by proxy than average for the conditions studied. In comparing the HIES interview prevalence with the medical record, cataracts appear relatively underreported by the interview. This finding is not surprising because it is unlikely a person would report cataracts unless they had been detected by a medical professional. Thirty-four per- cent more list-sample persons and 58 percent more house- hold members were shown as having cataracts in the medical record than were reported by the interview. Several factors may contribute to this difference. First, the NHIS asks about cataracts “in the past year.” If a person had cataract surgery more than 1 year before the interview date, the proper report would be “no.” The medical record review covered 3 years before the interview and included some persons recorded as having cataract sur- gery. A second, and probably more important, reason for the interview underreport is the likelihood of mention in the medical record of “early cataracts” that may not be mentioned to the patient, may be forgotten, or may not be considered as really having cataracts by the interview respondent. Type B mismatches were relatively uncommon; 27 of 83 interview reports were not confirmed by the medical record. However, a review of medical records for these apparently false positives indicates some possible confu- sion with similar (but not equivalent) conditions present in the records, such as eye floaters, diabetic retinopathy, uveitis, and dry eye syndrome. Some respondents who did not report cataracts mentioned in the medical record (type C mismatches) did report other eye problems. How- ever, these problems were typically confirmed by the record. Thus, any confusion by respondents about the definition of cataracts seems to contribute to overreport- ing rather than underreporting. Several indications from this study point to the likeli- hood of a significant underreport of cataracts in the NHIS: the relatively higher reported prevalence among a study population with good access to preventive eye care, the relative underreporting by the interview against the med- ical record, and the apparent additional underreporting by proxy respondents. Only an apparent slight tendency for definitional confusion to result in overreport counterbal- ances these factors. Constipation — Constipation (asked as “frequent con- stipation” in the past year in the NHIS checklist) was relatively much more prevalent (more than 100 percent for all study subjects) in the HIES sample than would be expected from the NHIS. The greatest difference is in the group 18-44 years of age (table 2). Part of this higher prevalence may be attributable to the higher proportion of black people in the HIES sample; prevalence from the NHIS is higher for black people than for white people (32). Prevalence from the interview and from medical records is similar for both list-sample persons and household members, but Kappa values for both are relatively low (tables 6 and 7). Thus, although there are about the same number of reports from the interview as from medical records, most reports from both sources are unconfirmed. The presence of apparently false positives is not surprising because constipation is somewhat subjective, afflicted per- sons may not seek care, and constipation as a symptom may not be recorded in the medical record when the condition causing it is. On the medical record side, both the difficulty of determining chronicity (3 months or more) and the timing (past year versus earlier) make the record potentially unreliable in providing reports that meet the NHIS definition. In summary, the similarity between prevalence ob- tained from the interview and prevalence in the medical record appears coincidental. Intuitively, it seems that the interview would be a better source for prevalence data 27 than the medical record. The fact that the Kappa value for household members (0.22) is twice as high as for list- sample persons may indicate that persons who tell family members about problems with constipation are more likely also to seek medical treatment. Diabetes — Diabetes has exceptionally good agreement between interviews and medical records: Kappa = 0.82 for list-sample persons and 0.74 for household members. Most of the mismatches are of type C—reported in the medical record but not in the interview. Thus, the inter- view had a net underreport of 23 percent for list-sample persons and 30 percent for household members. The high rate of agreement reflects the specific nature of the disease, and the general agreement between patients and physicians on the terminology used to describe it. Diabetes is also a condition that requires a specific test, and therefore physician visit, to diagnose, so it would be noted in the medical record. There were only three type B mismatches, where an interview report was not confirmed by the medical record. Review of these cases showed that two of the three had elevated blood sugar readings noted in their medical record, and the other one had a diagnosis of hyperglyce- mia. These three mismatches, then, are probably the result of some confusion by the patients as to their exact diagnoses. The type C mismatches, of which there are 40, are likely to be accurate indicators of underreporting by interview respondents. It is unlikely that the respondents reported this condition under any other name. Review of the medical records of these respondents showed that the majority had hypertension, and about one-fourth had heart disease. Most had a number of serious conditions but appeared to be reporting only hypertension or some other condition such as arthritis. It may be that respon- dents with multiple related conditions identify one as the source of their health problems — perhaps the most serious or the earliest diagnosed. Prevalence of diabetes in the HIES sample was con- sistently higher than would be expected from NHIS age- specific estimates (tables 2-5). However, this difference in most age groups is likely the result of the relative overrep- resentation of black people in the HIES sample; NHIS prevalence estimates of diabetes in black people are about twice that for white people, except among those under 45 years of age (32), for whom the prevalence among black people is less than 20 percent higher than that for white people. The HIES prevalence in the group aged 18-45 years is nearly 2.5 times the NHIS estimate for that age group. Given this pattern of reporting between the HIES interview and medical record and the high access to care among the HIES sample, the HIES findings suggest that the NHIS may significantly underestimate the prevalence of diabetes in younger persons. Migraine headache — Prevalence of migraine headache in the HIES sample was not significantly different than would be expected from the NHIS (tables 2-5). Compared with the medical record, the HIES interview overreported 28 migraine by 63 percent for list-sample persons and 100 per- cent for household members. Agreement between the interview and medical record was about the same for these groups, and a little worse than the average across condi- tions (tables 6 and 7). Although migraine is a fairly well-defined condition clinically, the popular concept of migraine may be indistinguishable from “bad headache.” Indeed, headaches were mentioned in the medical record for some type B mismatches (those with interview reports not confirmed by the medical record). Migraine is also a condition for which, once it is diagnosed, some patients seek no further treatment. Thus, there is good reason to suspect error in household reporting for migraine as well as in reliance on medical records for prevalence data. The HIP study showed a net underreport of “headache and migraine, chronic” and much lower agreement than the HIES for migraine (table 10), which may be because of the broader definition used in the earlier study. Heart conditions —Heart disease was reported much more often in the HIES interview than would be expected from NHIS prevalence estimates, particularly among younger persons. “Other selected diseases of the heart,” which includes vague reports such as “heart trouble,” showed the largest difference among those 18-45 years of age (table 2), and heart rhythm disorders showed the largest difference among older people (tables 4 and 5). Oversampling people for the HIES with recent doctor visits and hospital stays may have influenced these rates. With the exception of heart murmurs, heart condi- tions were more prevalent from the medical record than from the interview report. Heart murmurs represent a special case in this condition group because they are often detected early in one’s life and usually require no treat- ment. One reason for the relatively greater prevalence of heart conditions in the medical record is the phenomenon of individuals with multiple heart conditions (in different NHIS recode classifications). Often such people report only one or two of the conditions in the interview (some- times vaguely as “heart trouble”), apparently lumping together such diverse problems as angina and tachycardia. Only ischemic heart disease (including angina and myocar- dial infarction) showed a Kappa value above 0.40, at 0.62 for list-sample persons, and 0.68 for household members. The terms “angina” and “heart attack” are apparently among the most salient and least ambiguous to household respondents of the chronic conditions studied. In contrast to the HIES, the HIP and KP studies (table 10) both showed little net overreporting of heart conditions and also showed relatively good agreement between the interview and medical record or examination report. However, the KP study and some HIP tables used only one major category for heart conditions, and the classification of heart disease changed somewhat between the seventh and ninth revisions of the ICD. Gordon (16) found a considerable net overreport of heart conditions by self-report as opposed to medical records, but net under- report, compared with a physical examination. Just as some people have multiple heart problems classified into several NHIS recodes, some have multiple problems within an NHIS recode classification. Of the conditions studied, the heart conditions are nearly unique in this regard. The NHIS prevalence estimates are condition-level statistics (that is, they represent the num- ber of distinct conditions per 1,000 persons), so a person with two distinct conditions within a recode C classifica- tion would contribute two counts to the estimate. The HIES analysis has examined only person-level prevalence, where a person can only contribute one count within a recode C classification. Table 11 compares person-level prevalence with condition-level prevalence for heart con- ditions in the HIES, for list-sample persons only. The “person-level prevalence” columns of table 11 replicate the information in table 6 for heart conditions. The “condition-level prevalence” columns show prevalence us- ing normal NHIS rules. Aside from ischemic heart dis- ease, there is no significant difference between the person- level and condition-level rates for household reports. Again, the reporting for angina and heart attack (myocardial infarction) appear relatively good — interview respondents are somewhat able to distinguish the two conditions and report them separately. However, on the medical records, there are much bigger differentials between person-level and condition-level rates for ischemic heart disease and other selected diseases of the heart than in interview reports. Heart rhythm disorders (except heart murmurs) show a slight increase in prevalence at the condition level on the medical record side but no change on the interview side. Overall, relative underreporting of heart conditions in the interview jumps from 29 percent at the person level to 44 percent at the condition level. Excluding heart mur- murs, the rates of relative underreporting are 40 percent at the person level and 53 percent at the condition level. Case-by-case review of persons with reported heart conditions revealed that the interview respondent often mentioned one or two, perhaps ill-defined, heart ailments, but the medical record lists several specific problems, often falling into two or more NHIS recode groups. Of the mismatches in the NHIS recode group “other selected diseases of the heart,” 39 percent of the type B mis- matches fall into the “unspecified ill-defined” subcategory (e.g., “heart trouble”), but all of the type C mismatches are more specific ailments, further supporting this notion. The loose match for heart conditions consisted of a person-level collapsing of the three NHIS recodes under heart rhythm disorders into one and collapsing all heart conditions into one recode. Note that this approach vio- lates the “clinical equivalence” criterion for the loose match described earlier but follows the NHIS practice of presenting prevalence estimates for heart conditions in the aggregate categories. In table 12, the “loose match 1” row for heart rhythm disorders shows an increase in the number of positive matches over the original match by NHIS recode, and a Kappa value considerably higher than that for two of the three NHIS recodes individually. These results indicate that there may be some confusion within the heart rhythm disorder categories. Excluding heart murmurs from the loose match (loose match 2) further improves the Kappa value, mostly by eliminating more than half the type B mismatches from the totals. The loose match for all heart conditions reveals a similar pattern. The Kappa value for all heart conditions combined is 0.58, and 0.60 if heart murmurs are taken out. Again, this result indicates some confusion about the exact nature of heart trouble by some respondents. The improve- ment in agreement is also related to the earlier observa- tion of multiple heart conditions (across NHIS recodes) being more likely in the medical record than in the interview. Even at the aggregate level, however, there remains considerable underreporting of heart conditions in the interview as opposed to the medical record. Thus, the evidence from the HIES data suggests that NHIS prevalence estimates for heart conditions may be low for several reasons: Interview respondents may fail to mention a heart condition at all, people with conditions in multiple NHIS recodes may report only one or two, and people with multiple conditions within one NHIS recode ‘may report fewer than are delineated in the medical record. Hypertension — The reporting pattern for hypertension is similar to that for diabetes. After diabetes, hypertension had the highest rates of agreement of any of the chronic conditions reviewed (Kappa = 0.72 for both list-sample persons and household members), and a comparison of the prevalence between interview and medical record shows a slight net underreport by the interview. Like diabetes, hypertension requires a medical provider’s diag- nosis, so the net underreport is not surprising. However, some 59 list-sample persons reported hypertension that was not confirmed by the medical record. Some of these type B mismatches may be the result of patients receiving the diagnosis of hypertension before the 3-year period covered by abstracted medical records or by self-testing of their blood pressure. More than one-fourth of type C mismatches for hyper- tension had long medical records, indicating a poor health status. About one-third of the type C persons had some type of heart disease, which was often reported in inter- views. As discussed under heart conditions, such respon- dents may have felt they covered the topic by reporting the most salient of their circulatory problems or may have reported a general problem meant to encompass both heart disease and hypertension. Like diabetes, the HIES prevalence for hypertension far exceeds what would be expected from NHIS estimates and most notably for persons 18-45 years of age (table 2). Also like diabetes, hypertension is more prevalent among black than white people according to the NHIS (32) (about 69 percent higher for persons under age 45, with gradually decreasing differentials in older age groups). However, this does not explain the large differences in tables 2 and 3 (HIES 233 percent higher for persons 18-44 years of age, 90 percent higher for persons 45-64 years of age). Once again, the relatively higher access to care of 29 the HIES sample may be related to the higher-than- expected prevalence of hypertension. If this relationship does exist, the NHIS estimates of the prevalence of hypertension in the general population under 65 years of age may be considerably below the true prevalence. An- other way of stating the same thing is that the general population may have considerable undetected or unac- knowledged hypertension among the groups under 65 years of age, a supposition consistent with comparisons of medical histories and clinical examinations in the National Health Survey (15). Hardening of the arteries—The HIES interview and medical record showed very low prevalence of atheroscle- rosis, also called hardening of the arteries—19 interview reports and 14 from the medical record for list-sample persons. Only one of these reports matched, resulting in a Kappa near zero. From a clinical standpoint, hardening of the arteries requires a physician’s diagnosis and is a gradual process occurring in all persons (with no definitional threshold). The term may mean conditions other than atherosclerosis to some respondents, which could explain some of the type B mismatches. In addition, some type B mismatches might be the result of the patient reporting atherosclero- sis, a general condition, when they have developed more specific conditions as a result. The more specific condition would more likely be recorded on the medical record. Atherosclerosis, NHIS recode 510, consists of athero- sclerosis of any arteries. It would be clinically consistent to match ischemic heart disease and angina pectoris to ath- erosclerosis for analytical purposes. Ischemic heart dis- ease is a form of atherosclerosis. Atherosclerosis is the only cause of ischemic heart disease. Angina pectoris is also a sequela of atherosclerosis. Because they are clini- cally consistent, ischemic heart disease and angina pecto- ris are included in the loose match for atherosclerosis. This will result in a match for the patient who reported “hardening of the arteries,” but whose physician reported “angina.” Cerebral atherosclerosis is also clinically consistent with atherosclerosis, although not included in the NHIS recode, and is therefore included in the loose match. Because cerebral atherosclerosis is a more specific diagno- sis than hardening of the arteries, one would expect it to be on the medical record but not the interview. Claudica- tion, (angina in the leg) is also a common sequela of atherosclerosis and is thus also included in the loose match. As with the two previously described conditions that are included in the loose match, adding claudication should result in matching some previously denoted type B mismatches. As shown in table 12, applying the loose-match crite- ria to the type B and C mismatches results in a dramatic improvement in the match for atherosclerosis—a Kappa value of 0.686 as opposed to 0.045. Of the 16 type B and C mismatches that became loose matches, 10 were previ- ously matched on ischemic heart disease. These figures suggest two possible explanations for the low match rate 30 on atherosclerosis. First, persons with conditions of the circulatory system, including heart conditions and hyper- tension, may tend to summarize their complaints in one or two condition names. If so, atherosclerosis may be part of this phenomenon. Second, persons who have developed ischemic heart disease may report an earlier diagnosis of atherosclerosis, whether or not they report the more recent condition, although the medical record (limited to the past 3 years) makes no mention of the earlier condition. Extending the loose match to previous negative matches results in a great increase in the prevalence of atheroscle- rosis, both from the interview and medical record. Be- cause the loose match essentially combines atherosclerosis with ischemic heart disease, the second loose-match re- sults look very similar to the figures for ischemic heart disease. The loose-match results are relatively encouraging for the accuracy of NHIS estimates, in the sense that persons reporting atherosclerosis appear to be clinically correct much of the time. However, the loose match raises the question of how to define the “true” prevalence of athero- sclerosis. If some, but not most, people who have devel- oped ischemic heart disease report an earlier diagnosis of atherosclerosis, then the NHIS prevalence estimates may be either somewhat too high or considerably low, depend- ing upon how the prevalence rate is defined. Varicose veins of the lower extremities — Varicose veins were reported in the HIES at roughly comparable rates to the NHIS (tables 2-5). They were overreported in the interview compared with the medical record by more than 200 percent for list-sample persons and 100 percent for household members (tables 6 and 7). Persons afflicted with varicose veins may not seek medical advice or treat- ment for many years after a medical consultation. In the HIP and KP studies, varicose veins were not limited to those in the lower extremities. The HIP figures for varicose veins were similar to those in the HIES, but the KP study, although showing a comparable rate of false negatives to the other studies, did not show net overreport- ing by the interview. Hemorrhoids —Hemorrhoids may be considered a stig- matizing condition, and thus one would expect net under- reporting in an interview. However, the HIES interview showed a net overreport of 35 percent against the medical record for list-sample persons and 58 percent for house- hold members (tables 6 and 7). The less frequent report- ing by list-sample persons than by household members, which is contrary to the typical pattern, may reflect some respondent embarrassment at mentioning their own hem- orrhoids. Among adult household members, those present for the interview reported hemorrhoids at about the same rate as the medical record, but proxies reported almost three times more hemorrhoid cases than the medical record for persons not present in the interview. The general apparent overreport by the interview compared with the medical record may reflect self- medication, mis-self-diagnosis of hemorrhoids, or, as sug- gested by Marquis (12), possible disinclination of medical professionals to check for them in an examination. Fur- ther, hemorrhoids discovered during an examination for another condition, such as colorectal cancer, may not be noted in the medical record. These speculative reasons and the relatively low rate of agreement between the interview and medical record (Kappa = 0.27 for list- sample persons and 0.32 for household members), suggest that the actual prevalence of hemorrhoids in the sample population may be underestimated by both sources. Hemorrhoids were reported at about the same rate by HIES respondents as would be expected from the NHIS estimates, except among those 75 years of age and over (table 5), who reported hemorrhoids about 90 percent more than the NHIS estimate for persons in that age group. The HIP and KP studies showed similar net overre- porting of hemorrhoids (table 10), with the HIP study showing slightly less agreement between interview and examination report than HIES, and the KP study showed considerably more agreement. Chronic bronchitis — The agreement between the inter- view and medical record for chronic bronchitis is very low (Kappa = 0.09 for list-sample persons and 0.14 for house- hold members). However, the prevalence is similar be- tween the two sources, with a net interview overreport of 25 percent for list-sample persons and 31 percent for house- hold members (tables 6 and 7). There were only 5 positive matches for list-sample persons, but 40 type B mismatches (interview report only) and 31 type C mismatches (medical record report only). Part of the reason for the discrepancy between medi- cal records and interviews is the NHIS recode definition of chronic bronchitis. Half of the 31 type C mismatches are interpreted from the medical record as “bronchitis not specified as acute or chronic.” These cases may have been acute and would not have been reported in the interview. In 11 of the 40 type B mismatches, the respondent reported chronic bronchitis, and the medical record indi- cated acute bronchitis. It cannot be determined whether the medical record reflected an acute episode in a person with the chronic condition, which would mean that both sources were correct, or if one source misreported. The NHIS and HIES prevalence estimates for chronic bronchitis differ by as much as 58.1 percent for the group 65-74 years of age, and as little as 5.9 percent for those 75 years of age and over. Several related diseases occurred concurrent with chronic bronchitis in the HIES survey and medical records for some individuals. Possible related conditions include rhinitis, chronic cough, upper respiratory infection, sinusi- tis, asthmatic bronchitis, pneumonia, and chronic obstruc- tive pulmonary disease (COPD). Among these, only COPD is diagnostically consistent enough to be used for a loose match. There were no reports of COPD among list-sample persons in the HIES, however, so no loose-match analysis was performed for chronic bronchitis. Asthma — Asthma has fairly good agreement between medical records and household interviews, compared with other conditions, despite its relatively low prevalence among the study sample. The interview showed a net underreport of 20 percent for list-sample persons and 30 percent for household members, with both sample groups showing fairly high rates of agreement (Kappa = 0.55 for list-sample persons and 0.58 for household members). Asthma, like the other conditions with at least fair agreement (Kappa greater than 0.40), requires a physician visit for diagnosis. Severe asthma may require many phy- sician visits throughout the year, which would increase the likelihood of agreement between the interview and medi- cal record. On the other hand, several factors may account for cases where the two sources did not agree. The interview asks whether the respondent has had asthma in the past 12 months. The patient may have had it, but it could legitimately not be on the medical record if it did not require medical supervision. Alternatively, the medical record may have mentioned asthma more than 1 year before the interview, and the patient may not have suf- fered an attack in the interview reference period. Yet another possibility is that the patient could have actually had a similar but different condition, such as acute bron- chitis, and erroneously reported it as asthma. Bronchitis not specified as acute or chronic is included in the NHIS recode for asthma. Therefore, in some cases where the medical record appears to note asthma, but the interview does not, the medical record could be reflecting acute bronchitis, and not asthma. Also, the medical record could have a notation of asthma, but it might not have been recent or important enough for the respondent to report it in the interview. Although chronic bronchitis and asthma appear fre- quently together, they are distinct diseases and were not grouped together for a loose-match analysis. Upper respiratory conditions — Two chronic conditions, chronic sinusitis and allergic rhinitis without asthma, show similar patterns of reporting, with many more interview reports than medical record notations. The net overreport for the interview is 226 percent and 221 percent for chronic sinusitis and allergic rhinitis, respectively, for list-sample persons. The net overreport rates for household members are about one-half those for list-sample persons. These conditions have among the lowest match rates of all conditions studied, with Kappa values of only around 0.1 for list-sample persons. The Kappa values are higher, around 0.2, for household members. This difference, and the difference in net overreport for the two samples, suggests that respondents may report more cases, includ- ing perhaps less serious cases, for themselves than for others. More serious cases would be both more likely to receive medical attention and more likely to be noticed by other family members. Two attributes of these conditions also contribute to the reporting pattern described here. First, chronic sinusi- tis and allergic rhinitis may be easily confused. In the NHIS recode definitions, allergic rhinitis includes hay 31 fever, pollinosis, and spasmodic rhinorrhoea, but chronic sinusitis includes postnasal drip and sinus drainage. For example, some respondents, may consider postnasal drip to be a symptom of allergy rather than sinusitis. Second, both allergic rhinitis and chronic sinusitis are frequently self-treated. If they were never the reason for a medical visit or involved in a diagnosis in the 3 years covered by the medical record, they would probably not be in the record. Review of the mismatched medical records and interviews revealed that often these two conditions occur in tandem. In addition, the medical record showed many sample persons reporting these conditions having lower respiratory conditions. Chronic sinusitis is both a disease entity and a descrip- tion of specific symptoms. However, allergic rhinitis, is a disease that can manifest itself with a variety of symptoms, including sinusitis. Chronic rhinitis and chronic nasophar- yngitis are also sequelae of allergic rhinitis. Because of this relationship, the four conditions — allergic rhinitis, chronic sinusitis, chronic rhinitis, and chronic nasopharyngitis — were grouped together for a loose match. The results are presented in table 12. Under “loose match 1,” the row “Upper respiratory problems” shows the results of com- bining allergic rhinitis and chronic sinusitis; a small in- crease in agreement indicates some possible confusion of the two conditions, but the pattern of much higher preva- lence from the interview report persists. Similarly, adding chronic rhinitis and chronic nasopharyngitis to the loose match (“loose match 2”) increases the agreement slightly, but does not affect the overall pattern of mismatches. The HIES prevalence is consistently higher than the NHIS estimates for both conditions, more so for allergic rhinitis. This tendency may be related to the climate around Washington, D.C., which is damp and laden with pollen and other irritants much of the year. Effect of person characteristics on reporting As previously noted, the HIES sample was skewed in a number of ways when compared with the U.S. popula- tion: All persons in the study were HMO members, the population from which the sample was drawn included a much higher proportion of black people than the general U.S. population, and the design oversampled older people and people with recent doctor visits and hospital stays. Some effects of these design features were apparent when the relative prevalence of the studied chronic conditions was compared between NHIS estimates and the HIES sample. However, the analysis has not included differen- tials in reporting behavior between NHIS estimates and the HIES sample across different person characteristics. Tables 13 and 14 summarize the reporting of all chronic conditions together by various demographic and other person characteristics, for list-sample persons and household members, respectively. Rather than prevalence, as in tables 6 and 7, these tables present the mean number of NHIS recode conditions per person. As in the earlier 32 tables, a particular NHIS recode condition is only counted once per person, even though for some conditions a person may have more than one condition falling into the recode. The net and proportional overreport and Kappa columns are similar to those in tables 6 and 7; that is, they are computed from the total numbers of type A and D matches and type B and C mismatches across all conditions. Demographic characteristics Age — Distinct patterns of increasing number of condi- tions and decreasing net and percent of overreported conditions are apparent across increasing age groups among list-sample persons. Kappa values are markedly lower for the youngest (under age 45) and oldest (75 years of age and over) age groups. The increasing number of condi- tions reported with age is expected; the other patterns are attributable to different causes. The lower Kappa values and higher overreporting for persons under 45 years of age are in large part the result of the mix of conditions reported for this age group. The most common conditions among those studied include upper respiratory ailments and orthopedic impairments, which show generally lower- than-average agreement between the data sources and are among the most overreported conditions. Younger per- sons may also be less likely to seek treatment for relatively minor conditions. The oldest group are more likely to have heart conditions and cataracts, conditions that are gener- ally underreported in the interview and that have rela- tively low Kappa values. Older persons may also be more likely to have cognitive problems that interfere with accu- rate reporting and may be less likely to report less serious conditions because they have more conditions overall. Previous research has yielded mixed results with re- gard to reporting differences by age (table L), with both younger and older respondents appearing to be more in agreement with medical records in different situations. The HIES analysis suggests that these differences may be attributable in part to differences in what conditions were included in the respective analyses. Sex —List-sample women in both the age group under 65 years and the group 65 years and over were more likely to overreport compared with the medical record than men. Men showed only slightly higher agreement with the medical record. There was little difference in the number of conditions reported per person by the medical record between sexes. Women under 65 years of age were much more likely to report upper respiratory problems, migraine headaches, hemorrhoids, and heart murmurs not confirmed by the medical record than were men under age 65. Among those aged 65 years and over, women were more likely than men to report arthritis and varicose veins not confirmed by the medical record. Both sources showed considerably higher prevalence of these two conditions among women as well. Thus, greater reporting by women appears to be largely for conditions that might be viewed as embarrassing (hem- orrhoids, varicose veins, possibly arthritis) or that are usually relatively minor (heart murmurs, chronic bronchi- tis, allergic rhinitis, chronic sinusitis). For the latter set of conditions, the medical record shows little difference in prevalence between sexes, but the interview shows consid- erably greater prevalence among women. By and large, the NHIS estimates these conditions as significantly more prevalent among women as well. As with age, previous research has yielded mixed results on whether men or women have higher agreement with medical records. Again, this pattern may be attribut- able in part to the specific conditions studied. Race —The racial composition of the HIES sample allows comparisons only between black people and mem- bers of other races. Black persons both under age 65 and age 65 or over in the HIES overreported somewhat more compared with the medical record than did their counter- parts of white and other races. However, black people’s reports showed more agreement with the medical record than those of white people and people of other races. Overall, the patterns of reporting by condition between races were fairly comparable to those in the NHIS. Socioeconomic characteristics Employment status — People not currently employed reported considerably higher numbers of chronic condi- tions than did the employed in both the age group under 65 years and the group 65 years and over. The unem- ployed showed slightly higher agreement with the medical record in both age groups, but there was no pattern for overreporting against the medical record. Income — The total number of conditions per person declined as family income increased in the NHIS; family income may vary with age, affecting the number of condi- tions per person. Proportionate overreporting against the medical record was higher for those with family incomes under $30,000, although no particular pattern of agree- ment between the interview and medical record was ap- parent by family income. Education — Agreement between the interview report and medical record did not vary by education. College graduates did overreport noticeably less than those with less education, although the pattern for those with less than a college degree was that those with more education overreported more. Medical services utilization Two-week doctor visits — People with doctor visits in the 2-week reference period (according to the medical record) had somewhat more chronic conditions per person than those without such doctor visits in both the age group under 65 years and those 65 and over. Those with 2-week doctor visits showed considerably less overreporting against the medical record and slightly more agreement with the medical record in both age groups. This finding is consis- tent with that of the HIP study (1). Health assessment — GHA offers a comprehensive med- ical checkup called a health assessment to its members. People who had had health assessments in the 2 years before the interview date were comparable to those with- out health assessments in the number of chronic condi- tions in the medical record. However, people with recent health assessments overreported against the medical record at lower rates and showed somewhat higher agreement with the medical record than persons without recent health assessments. Thirteen-month hospital stay— As expected, persons with hospital stays within the 13-month reference period had more chronic conditions per person than those with- out recent hospital stays. Those with 13-month hospital stays overreported against the medical record less in both the age group under 65 years and those 65 years of age and over. However, those persons 65 years and over with 13-month hospital stays had lower rates of agreement with the medical record than those without stays, and the reverse pattern was true for persons under age 65. Self-perceived health status Both the interview report and the medical record showed a strong correlation between perceived health status and the number of chronic conditions per person. Persons reporting themselves in excellent health overre- ported the fewest conditions compared with the medical record but also had a noticeably lower rate of agreement with the medical record. These observations may indicate that persons reporting themselves in excellent health are less likely to report chronic conditions than those report- ing very good, good, fair, or poor health. Number of chronic conditions reported Tables 13 and 14 compare the reporting of chronic conditions for persons reporting different numbers of conditions included in the interview checklist. The propor- tion of overreporting increases with the number of condi- tions reported. However, the Kappa values for persons with four or more conditions are lower than those for persons reporting fewer conditions in both the list and supplementary samples. Persons reporting four or more conditions may be more prone to overreporting than others, or they may tend to report a higher proportion of conditions not likely to be. confirmed by the medical record. Response status The HIES was not designed as a formal experiment comparing self- and proxy reporting. However, the inclu- sion of household members in the analytic sample allows ad hoc comparison of responses for adult household members who were present during the interview (and presumably responded for themselves in most cases) and adult household members who were not present, for 33 whom proxy responses were obtained. The final rows of table 14 present totals for adult household members by whether they were present during the interview. Those not present for the interview had fewer condi- tions reported than those who were present but also had fewer conditions in the medical record, confirming the observation of Berk, Horgan, and Meyers (8) that persons not present for the interview appeared to be healthier (from the perspective of number of conditions) than those who were present. The Kappa values for the two groups are virtually the same, but persons present for the inter- view overreported, compared with the medical record, at a higher rate (19 percent) than those not present (5 per- cent). Household members present for the interview over- reported at about the same rate as list-sample persons (21 percent), who were also self-respondents. The number of reports for specific conditions is too small for meaning- ful analysis at the condition level between the self- and proxy reporters. Discussion The HIES was designed to evaluate the reporting of chronic conditions in the NHIS by comparing interview responses to medical records for the same individuals; it is the first such evaluation in nearly 20 years. The major strength of the evaluation is the use of a full study design in which both positive and negative reports from the interview and medical record can be compared for all study subjects. Some additional features that enhance the ability to examine chronic condition reports and focus on demographic subgroups of particular policy interest are oversampling of older persons and persons with recent health care visits and the selection of an HMO with a large minority membership. The study population comprised HMO members in- terviewed about themselves; with a moderate additional effort, interview and medical record data on household members were obtained, which replicated the findings and permitted some analysis of proxy reporting. HIES meth- ods and procedures followed those of the NHIS as closely as possible, so that HIES findings could be used to help evaluate the NHIS. The research described here from the HIES has supported previous studies’ observations that survey inter- views and medical records often provide very different pictures of the prevalence of chronic conditions in a population. The HIES design and analysis have notas- sumed the medical record to be a “gold standard” with regard to the presence of chronic conditions but rather have focused on interpreting the differences between the two data sources. Some of these differences are artifacts of the procedural differences in acquiring and interpreting reports from the two sources, but others are inherent in the definitions, manifestations, and need for professional medical care of the conditions studied. Regardless of the reason for the differences, their existence has ascertained the accuracy of survey-based prevalence estimates of chronic conditions. 34 It is helpful to classify chronic conditions in several ways. For the first group of conditions, consider those that require a physician’s diagnosis to identify and are likely to require ongoing medical care. Among the conditions stud- ied, the following may be considered in this group: diabe- tes, most heart conditions, high blood pressure, and asthma. Two conditions not included in this list are cataracts, which do require a physician’s diagnosis but do not re- quire ongoing care, and hardening of the arteries, which meets the criteria but may be subsumed in a more imme- diate condition. Once diagnosed, the presence of these conditions is likely to be noted in the medical record within a 3-year period. (Medical records examined in the HIES covered the 3 years prior to the date of the interview.) These conditions are also all considered “chronic by definition” by NHIS coding rules; that is, ever having the condition counts as having it at the time of the interview. For these conditions, the medical record may be considered as near a “gold standard” as is possible to find. Each of these conditions (with the exception of heart murmurs, a special case among heart conditions) was underreported by the HIES interview against the medical record, from a low among list-sample persons of 4 percent underreport for hypertension to a high of 48 percent underreport for “other selected diseases of the heart,” while most other conditions were apparently overreported. Diabetes, asthma, high blood pressure, and ischemic heart disease also had the highest rates of agreement among all conditions stud- ied, with Kappa values among list-sample persons ranging from 0.55 for asthma to 0.82 for diabetes. Thus, one may conclude that interview reports of these conditions are likely to be accurate, but that their prevalence may be underestimated by survey data. The problem of underes- timation may be particularly problematic for heart disease, where individuals with more than one condition (accord- ing to the medical record) often reported fewer conditions in the interview. The other conditions apparently underreported by HIES respondents were cataracts and dermatitis. Al- though the medical record may have overstated the prev- alence of cataracts (counting some that were surgically removed before the “past year”), it is likely that cataracts are underreported by survey respondents. Many notations of “beginning cataracts” or “early cataracts” were noted in the records; these cases may not be salient enough for respondents to remember or may not have even been mentioned by the provider discovering them. Dermatitis is a condition for which chronicity is difficult to determine from the medical record —the apparent HIES underreport likely does not indicate a corresponding underreport from the NHIS. At the other end of the spectrum from the first group of conditions are those that can only be diagnosed by patient report: Constipation and tinnitus from the list studied here meet this criterion. Both were significantly overreported by HIES list-sample persons, and both had very low rates of agreement with the medical record. For these conditions, the medical record reports shed almost no light on the accuracy of interview-based prevalence estimates. However, they do suggest that many people do not report these conditions to their physicians, so that medical records would underestimate prevalence. Another group of conditions is those that may be salient to the persons suffering from them but that may not require ongoing treatment and thus may not be in the medical record. These include orthopedic impairment, visual and hearing impairment, migraine headache, vari- cose veins, allergic rhinitis, and chronic sinusitis. These conditions were substantially overreported in HIES inter- views, but, with the exception of visual and hearing impair- ments, all had substantial numbers of type C mismatches (medical record report only) as well—more type C mis- matches than type A matches (reported in both interview and medical record). The presence of impairment is a subjective determination, whether by a provider or an individual; for other conditions in this group, some self- diagnosis probably occurs. The extent to which such self- diagnosis would conform to a physician’s opinion cannot be determined from the data, but undoubtedly some interview reports for conditions in this group (other than impairment) are false positives. Overall, medical records provide a different picture of prevalence for this group of conditions than do interviews and the rates from medical record data would likely be considerably lower. Some conditions studied are less well defined than others from the household respondents’ perspective and from a clinical perspective. These issues were discussed in the context of the “loose match” that grouped clinically equivalent conditions. Some interview reports of arthritis, although technically “false positives,” appear to match clinically equivalent conditions in the medical record. The extent to which other reports of arthritis may reflect more generalized joint pain could not be determined. Circula- tory conditions are a special case of definitional problems from a respondent’s perspective. The current research has provided some evidence that people with several heart or other circulatory conditions tend to group them under one heading. The loose-match analysis found evidence of this for heart conditions; it may be true for the larger family of circulatory conditions as well. That is, persons with heart disease may report “high blood pressure” as the global condition that encompasses all their circulatory problems. Potentially the most interesting of the HIES design features with regard to its effect on study findings is the universal access to health care and the emphasis on preventive care in an HMO setting. (An HMO population was selected for the HIES because an HMO is one of the few health care settings in which a full-design record check is feasible; it includes a complete set of provider records.) Evidence from our analysis and previous re- search indicates that people who get medical care are better able to report the presence of chronic conditions. This is true for the first group of conditions described earlier because a physician’s diagnosis is required for a person to know that he or she has the condition. Among the general population, many of whom have less access to medical care than the study sample, what would be the effect on reporting of chronic conditions and thus on prevalence rates? It may be that the conditions underre- ported in the HIES would be more underreported in a national sample —both because of people who have not had a diagnosis and because of people who have not sought medical care after receiving a diagnosis. The former would not know they had the condition, and the latter might forget or deny its existence. Conversely, self- diagnosed conditions might be more overreported among the general population than in the HMO study sample, as persons with limited access to care might have less chance to have their diagnoses refuted. Finally, proxy effects seem to be present in the report- ing of chronic conditions. Some of the differences between the list sample and household members are consistent with differential reporting by proxies. Generally, with the exception of embarrassing or stigmatizing conditions, one would expect estimates based on proxy reports to be lower than those based on self-reports. Overall, interview- derived prevalence for household members, some of whom were reported by proxies, was closer to that from medical records than it was for list-sample persons, who were all self-respondents. The comparison of self-responders and persons with proxy reports among household members indicated that proxy reports included considerably less overreporting, but agreement with the medical record was about the same for the two groups within the supplemen- tary sample. The net effect of proxy reports on NHIS prevalence estimates is difficult to determine from the analysis described here. 35 References 10. 11. 12. 13. 14. 15. 36 Balamuth E. Health interview responses compared with medical records. National Center for Health Statistics. Vital Health Stat 2(7). 1965. Madow W. Interview data on chronic conditions compared with information derived from medical records. National Center for Health Statistics. Vital Health Stat 2(23). 1967. Madow W. Net differences in interview data on chronic conditions and information derived from medical records. National Center for Health Statistics. Vital Health Stat 2(57). 1973. U.S. Public Health Service. The National Health Survey 1935-1936 scope and method. Public Health Bibliography Series 5. 1951. Trussell R, Elinson J. Chronic illness in a rural area. Cambridge: Harvard University Press. 1959. Cannell C, Fowler F. Comparison of hospitalization report- ing in three survey procedures. National Center for Health Statistics. Vital Health Stat 2(8). 1965. Daugherty V. Illness conditions. In: Anderson R, Kasper J, Frankel M, eds. Total survey error. San Francisco: Jossey Bass. 1979. Berk M, Horgan C, Meyers C. The reporting of stigmatizing conditions: A comparison of proxy and self-reporting. J Econ Soc Meas 14:197-205. 1986. Cox B, Cohen S. Methodological issues of health care surveys. New York: Marcel Dekker. 1985. Cox B, Iachan R. A comparison of household and provider reports of medical conditions. J Am Stat Assoc 82(400):1013- 18. 1987. Jabine T. Reporting chronic conditions in the National Health Interview Survey: A review of tendencies from evaluation studies and methodological tests. National Cen- ter for Health Statistics. Vital Health Stat 2(105). 1987. Marquis K. Record checks for sample surveys. In: Jabine T, Loftus E, Straf M, et al, eds. Cognitive aspects of survey methodology: Building a bridge between disciplines. Wash- ington, DC: National Academy Press. 1984. Edwards W, Cantor D. Toward a response model in estab- lishment surveys. In: Biemer P, Forsman G, Lyberg L, Mathiowetz N, et al, eds. Measurement errors in surveys. New York: John Wiley and Sons, Inc. 1991. Tourangeau R. Cognitive sciences and survey methods. In: Jabine T, Loftus E, Straf M, et al., eds. Cognitive aspects of survey methodology: Building a bridge between disciplines. Washington DC: National Academy Press. 1984. Roberts J, Rowland M. Hypertension in adults 25-74 years of age, United States 1971-1975. National Center for Health Statistics. Vital Health Stat 11(221). 1981. 16. 17. 18. 19. 20. 21 22. 23. 24. 28. 26. 27. 28. 29. Gordon T. Three views of hypertension and heart disease. National Center for Health Statistics. Vital Health Stat 2(22). 1974. Harlow S, Linet M. Agreement between questionnaire data and medical records. Am J Epidemiol 129(2):233-47. 1989. Harlow S, Linet M. The authors reply. Am J Epidemiol 409-10. 1991. Hertz-Picciotto I. RE: Agreement between questionnaire data and medical records: The evidence for accuracy of recall (editorial). Am J Epidemiol 408-09. 1991. Landis J, Koch G. The measurement of observer agreement for categorical data. Biometrics 159-74. 1977. Maclure M, Willett W. Misinterpretation and misuse of the Kappa statistic. Am J Epidemiol 126(2):161-9. 1987. Barlow W, Lai M, Azen S. A comparison of methods for calculating a stratified Kappa. Stat Med 10:1465-72. 1991. Linet M, Harlow S, McLaughlin J, et al. A comparison of interview data and medical records for previous medical conditions and surgery. J Clin Epidemiol 42(12):1207-13. 1989. Kovar M, Wright R. An experiment with alternate respon- dent rules in the National Health Interview Survey. Proceed- ings of the American Statistical Association Social Statistics Section. Alexandria, VA: American Statistical Association. 311-16. 1973. Haase K, Wilson R. The study design of an experiment to measure the effects of using proxies resources in the Na- tional Health Interview Survey. Proceedings of the Ameri- can Statistical Association Social Statistics Section. Alexandria, VA: American Statistical Association. 289-93. 1972. White A, Massey J. Selective reduction of proxy response bias in the Household Interview Survey. Proceedings of the American Statistical Association Social Statistics Section. Alexandria, VA: American Statistical Association. 211-16. 1981. Laurent A, Cannell C, Marquis K, et al. Reporting health events in household interviews, effects of an extensive ques- tionnaire and a diary procedure. National Center for Health Statistics. Vital Health Stat 2(49). 1972. Gleeson G. Interviewing methods in the Health Interview Survey. National Center for Health Statistics. Vital Health Stat 2(48). 1972. Birnbaum Z, Sirken M. Design of sample surveys to esti- mate the prevalence of rare diseases, three unbiased esti- mates. National Center for Health Statistics. Vital Health Stat 1(11). 1965. 30. 31. Cannell CF. Reporting of hospitalization in the Health Interview Survey. Health Statistics from the U.S. National Health Survey. Washington, D.C.: U.S. Department of Health, Education, and Welfare. (Series D-No. 4). 1961. Krueger D. Measurement of prevalence of chronic disease by household interviews and clinical evaluations. Am J Public Health 47(8):960. 1957. 32. Adams P, Benson V. Current estimates from the National Health Interview Survey, 1989. National Center for Health Statistics. Vital Health Stat 10(176). 1990. 37 List of Detailed Tables 38 . Number and percent of persons in Health Interview Evaluation analytic samples, by selected characteristics . . . . . Prevalence of selected chronic conditions in persons 18-44 years of age from National Health Interview Survey estimates and from Health Interview Evalua- tion Survey household reports, by condition. ....... . Prevalence of selected chronic conditions in persons 45-64 years of age from National Health Interview Survey estimates and from Health Interview Evalua- tion Survey household reports, by condition . . ...... . Prevalence of selected chronic conditions in persons 65-74 years of age from National Health Interview Survey estimates and from Health Interview Evalua- tion Survey household reports, by condition. ....... Prevalence of selected chronic conditions in persons 75 years of age and over from National Health Inter- view Survey estimates and from Health Interview Eval- uation Survey household reports, by condition . . .... . Comparison of chronic condition reports for list- sample persons from Health Interview Evaluation Sur- vey interviews, medical records, and Kappa values, by condition coeds 20m LURES STE NE Bae E TEE HLGOE . Comparison of chronic condition reports for house- hold members form Health Interview Evaluation Sur- vey interviews, medical records, and Kappa values, by CONAMION + vv sve rmrrermurs sesms tases sie sas ses 30 40 41 42 43 44 10. 11 12. 13. 14. . Comparison of chronic condition reports for list- sample persons from Health Interview Evaluation Sur- vey interviews, medical records, and Kappa values, by CONBILION wins s nema rme wd s¥ SRM LR SRE s SHH a . Comparison of chronic condition reports for house- hold members from Health Interview Evaluation Sur- vey interviews, medical records, and Kappa values, by CONAILION : vv50 ev wa wiamate ba sens WEES Hawise £8 Percent of matching positive reports for selected con- ditions in three studies, by condition. ............. Prevalence of heart conditions in list-sample persons from the Health Interview Evaluation Survey, by source of information and cONBION. « «yi vos viv saa vir sven Number of conditions and Kappa values for two stud- ies and the Health Interview Evaluation Survey, by type of match and condition «+. c.vsv sews vsvrsnans Number of conditions, number and percent of overre- ports, and Kappa values from Health Interview evalu- ation Survey responses and medical records for list- sample persons, by selected characteristics . ........ Number of conditions, number and percent of overre- ports, and Kappa values from Health Interview Eval- uation Survey responses and medical records for household members, by selected characteristics . . . .. 46 47 48 48 49 50 Table 1. Number and percent of persons in Health Interview Evaluation analytic samples, by selected characteristics us. List Sample Household Member Combined Sample population Person characteristic number percent number percent number percent percent Age Under 18 years. . «we ve smesh cvs 0 0.0 285 40.5 285 16.7 26.3 18-44 YBEIS. « vu swusn ena s sms 309 30.7 173 24.6 482 28.2 42.8 AB-BAYOAIS, « co vw kv pin nE vin 373 37.1 138 19.6 511 29.9 18.9 BB-7AYOUS:. ss wv sn swan sus 193 19.2 71 10.1 264 15.5 73 75 years and Over . . vox v www sss 130 12.9 36 5.1 166 9.7 4.7 Sex and age Female: 65yearsandover. . .......... 179 17.8 63 9.0 242 14.2 7.0 Under 1 year-64 years. . . ...... 366 36.4 330 46.9 696 40.7 44.5 Male: 65Years anti Over. . . w «vss 0 s+ 144 14.3 44 6.3 188 11.0 5.0 Under 1 year-64 years. . . ...... 316 31.4 266 37.8 582 34.1 43.5 Race and age Black: BSysars and Over, . .« cx rsp aus 156 155 37 5.3 193 11.3 1.0 Under 1 year-64 years. . . ...... 518 515 452 64.3 970 56.8 11.3 White or other: 65 years and’ Over, : seus vm w in 167 16.6 70 10.0 237 13.9 11.0 Under 1 year-64 years. . . ...... 164 16.3 144 20.5 308 18.0 76.7 Employment and age Employed: 63 years and over. , . «us ve wre in 74 23 3.3 94 55 1.6 18-BAYOAIS « «vv vx pws www 595 59.3 244 34.8 839 49.2 46.6 Unemployed: 65yearsandover. ........... 251 25.0 84 12.0 335 19.6 10.4 18-64vyears ............... 86 8.6 66 9.4 152 8.9 15.2 Income! B0-S19,999 . ov «ven nnn naw wa 148 17.1 32 4.9 180 11.8 27.6 $20,000-$29,999. . ............ 119 13.8 51 7.8 170 11.2 23.3 $30,000-$49,999. . . ........... 263 30.4 198 30.3 461 30.3 (1) $50,000 and over ............. 335 38.7 373 57.0 708 46.6 32.9 Education? Less than high school. . . . ....... 167 16.7 69 16.6 236 16.7 24.4 High school graduate. . . ........ 307 30.7 138 33.2 445 31.4 38.7 Somecollege. . .............. 204 20.4 81 19.5 285 20.1 174 College graduate. . . ........... 321 32.1 128 30.8 449 31.7 19.9 SOURCES: U.S. Bureau of the Census, Statistical Abstract of the United States: 1989 (109th edition). Washington, D.C., 1990. 1990 (NCHS) unpublished data and (32). 1For U.S. population, categories are $0-$19,999, $20,000-$34,999, $35,000 and over. 2pPersons 18 years of age and over; for U.S. population, persons 25 years and over. 39 Table 2. Prevalence of selected chronic conditions in persons 18-44 years of age from National Health Interview Survey estimates and from Health Interview Evaluation Survey household reports, by condition Number of Percent overreports per overreports 100 persons, by HIES NHIS HIES HIES compared compared Condition name and NHIS recode number prevalence prevalence with NHIS with NHIS ANcoNAIONS ; 3 sv sos wrmus ws ess sR We PRES Eos 84 wns 925.8 1,302.9 377.1 40.7 oles Tr TITTY 101 48.9 62.2 13.3 27.3 Dormalilis « cvcovsrsws var wr spies mens vous wees vs 113 36.0 31.1 -4.9 -13.6 Blindness or other visual impairment . . ................ 201 27.2 18.7 -8.5 -31.4 Deafness or other hearing impairment. . . . .............. 203 47.8 41.5 -6.3 -13.2 Deformity or orthopedic impairment . . . ................ 228 138.3 114.1 -24.2 -17.5 TIAARUS: os + wv mn @ moms cas me Ga vw Pas ws 6a RE E80 240 14.4 20.7 6.3 44.1 Calaraols. . «cut vrms rs rac ramsey vase er 241 35 4.1 0.6 18.6 CORBUPGIINN, cris: x: vn 0 snows: so: rae sensi 6s 1 BUR: BE 0 VRE 6 2 RE 8 314 11.9 37.3 25.4 213.8 DIBIEEBE «iv #0 mom ons wins mis oo mim wE w k 403 10.7 37.3 26.6 249.0 MIGIAING NBAGBOKG: vv: vv six vw www ww rin ws mio in wm 406 57.2 64.3 71 12.4 Heartdisease. . . .............iiinnnnnnnns — 36.1 89.2 53.1 147.1 Ischemic heartdisease . . . . ................. cv... 502 4.1 2.3 -2.0 -49.4 Heartrhythmdisorders . . . . .............. civ Sten 25.3 49.8 24.5 96.8 Tachycardia or rapid heartbeat . . . . .................. 503 5.3 4.1 -1.2 -217 PGA TOUTING + 5 = 6s 0 ib # 5 80% 0 2 S00 U8 ec Et 0 hy 4065 4 J 90 il (8 0 504 17.1 37.3 20.2 118.4 Other heart rhythm disorders . . . . ................... 505 2.9 8.3 5.4 186.2 Other selected diseases of heart. . . .................. 507 6.7 37.3 30.6 457.4 High DIOOU PIeSSUre. & «5. ws Ha sme Mas sd we ®ve ms us 508 56.0 186.7 130.7 233.4 Hardening of theiarieries. : «sx cus musws sme mr we sme as 510 0.1 2.1 2.0 1,974.7 Varicose veins, lower extremities . . . . ................. 513 24.8 27.0 2.2 8.8 Hemorrhoids + vs svi vniwrepims er imran swe en ems dy 514 57.2 83.0 25.8 45.1 Chronic bronchilis. ..c +» sms sp ems wma ws ns Wa Un sw os 601 44.5 43.6 -0.9 -2.1 AID vn 5 wes v2:9 1 5 vi ty 5 5 op 2 7 i SY 0 oe 0 a TF 2 6 602 41.3 45.6 4.3 10.5 Allergic rhintiswithout asthma .. . «.. vv vc vive ws vmv vs 603 108.8 209.5 100.7 92.6 Chronicsinusitis . . ........... citi rnnnnnns 605 161.1 184.6 23.5 14.6 NOTE: NHIS is National Health Interview Survey and HIES is Health Interview Evaluation Survey. 40 Table 3. Prevalence of selected chronic conditions in persons 45-64 years of age from National Health Interview Survey estimates and from Health Interview Evaluation Survey household reports, by condition Number of Percent overreports per overreports 100 persons, by HIES NHIS HIES HIES compared compared Condition name and NHIS recode number prevalence prevalence with NHIS with NHIS AHCONAIIONS . : crs rsnvpssnsmesvs ns sims sin 1,657.8 2,099.8 442.0 26.7 BANGS. 55 5 50 56 5 00 30s m4 Dem be RE wes» 101 253.8 254.4 0.6 0.2 DOIBIIIS « s svensn rs samre menses vans es 113 30.6 54.8 24.2 79.1 Blindness or other visual impairment . ........ 201 45.1 35.2 -9.9 -21.9 Deafness or other hearing impairment. . . . . .... 203 127.7 97.8 -29.9 -23.4 Deformity or orthopedic impairment . . . ....... 228 155.5 148.7 -6.8 —4.4 TILE. + ooo 50 0 500) 00 1 06 9 Ef wf 0 240 45.8 37.2 -8.6 -18.8 CAAIBOIS. 05 50 vie 96 I HE RE 241 16.1 35.2 19.1 118.8 ConSURAUON. « 5 « wis 3 3 ose 0% 5 Gi B® EE 0 314 20.9 31.3 10.4 49.8 DYADIBEOR ote. 0 slow i 5 3s 0 (0) 1 403 58.2 135.0 76.8 132.0 Migraine headache ... «cvs vvowiwwvmans 406 51.2 41.1 -10.1 -19.7 HEA QISEA88.: vw « vi 40 bw wu wv won es EHF EE EE 118.9 199.6 80.7 67.9 Ischemicheart disease. . .. .... ov s0 0 sons 502 54.5 72.4 17.9 32.9 Heart thythm SOBs. « 5.x v ot wo mk 45 2 0d ik © al ww 40.1 97.8 57.7 144.0 Tachycardia or rapid heartbeat . . . . ......... 503 14.9 35.2 20.3 136.4 Heart MUMS ; » cov v4 Se sn smb ae gmake 4 504 16.4 45.0 28.6 174.4 Other heart rhythm disorders . . . ........... 505 8.8 17.6 8.8 100.1 Other selected diseases of heart. . . ......... 507 24.3 29.4 5.1 20.8 High BICOA Pressure. . . . «vss vms ws cms om 508 229.1 434.4 205.3 89.6 Hardening of the arteries. . . . ............. 510 16.1 15.7 -0.4 -2.8 Varicose veins, lower extremities. . . ......... 513 57.8 93.9 36.1 62.5 HEMOINOUS. «= 6 #v 00 win #00 oie ds hn ios 5 4 514 74.9 88.1 13.2 17.6 Chronic bronchitis... . . « cv crm rvs su x wnns 601 53.7 37.2 -16.5 -30.8 ASHRMID & + nis i wiv 30 0% me 5 RR oe 602 41.5 39.1 -2.4 -5.7 Allergic rhinitis without asthma . . ........... 603 87.4 119.4 32.0 36.6 CHIONICISITILSING « vu viv 51: ier) 5 100 Sos: Ww kimiihe th 605 173.5 201.6 28.1 16.2 NOTE: NHIS is National Health Interview Survey and HIES is Health Interview Evaluation Survey. 41 Table 4. Prevalence of selected chronic conditions in persons 65-74 years of age from National Health Interview Survey estimates and from Health Interview Evaluation Survey household reports, by condition Number of Percent overreports per overreports 100 persons, by HIES NHIS HIES HIES compared compared Condition name and NHIS recode numbers prevalence prevalence with NHIS with NHIS AY CORGIIONS.... . « vv0s 0 0.6 Bs BRIE Bg wank sah wow wise 2,393.4 2,734.8 341.4 14.3 AHRIS. ov vs sw vsmy wos mans pain musasns 101 437.3 473.5 36.2 8.3 DBIVEHIS vo. a «ons coves noon i gil h Bea 6 9 81% cu 113 33.5 45.5 12.0 35.7 Blindness or other visual impairment . ........ 201 69.3 79.5 10.2 14.8 Deafness or other hearing impairment. . . . ..... 203 239.4 212.1 -273 -11.4 Deformity or orthopedic impairment . . . ....... 228 141.4 140.2 -1.2 -0.9 THDRUS: «os 2 vs mmin smsimn moms mim #3 85m Ren 84 240 76.4 75.8 -0.6 -0.8 CBIBIACIS. , oo vmomn sims nara ms sm bibe Sm 241 107.4 132.6 25.2 23.4 CONBUPAION.. or: x: + 0 2 coi im 50 500 2 5000550 0 0 To Bo 0 314 42.2 53.0 10.8 25.7 DHABIOS 1 co. ots to 0 0m owe ec 0 oh 4 0 3 is £00 800 403 89.7 140.2 50.5 56.2 Migraine headache .................... 406 29.8 26.5 -3.3 -11.0 HESRTEISBABE. «1. wi 5 wiv bow mek a se B54 16 Ti R BRA 231.6 359.8 128.2 55.4 Ischemic heart disease. . .. . ...c.-scesvas 502 112.7 132.6 19.9 17.6 Heart thythm disorders. .. cc + « wiv s ns nv 0 04 hws sus 63.8 132.6 68.8 107.8 Tachycardia or rapid heartbeat . . . . ......... 503 19.5 56.8 37.3 191.4 Hoar IMIITVNG. 5 52 ws of iin iin 4 Bans i # Bima ae do 504 19.2 45.5 26.3 136.7 Other heart rhythm disorders . . . ........... 505 25.1 30.3 5.2 20.7 Other selected diseases of heart. . . ......... 507 55.1 75.8 20.7 37.5 Highblood pressure. . . oc vu ss svmsmsnis ves 508 383.8 503.8 120.0 31.3 Hardening ofthe arteries. . . .............. 510 28.9 22.7 -6.2 -21.4 Varicose veins, lower extremities. . . ......... 513 72.6 75.8 3.2 4.3 HEMOFAOIAS + + «ois ims ass oun oe as ios 514 77.4 83.3 5.9 7.7 CHrOBIC.BIONCHILIS . «v5 205 in sis sim vais wus 601 54.2 22.7 -31.5 -58.1 AEG. vs vnu nm smn as ESTE Hasan o 602 57.3 45.5 -11.8 -20.7 Allergic rhinitis without asthma . . ........... 603 69.4 98.5 29.1 41.9 Chronic sinusitis . . . ............. 0... 605 151.8 162.9 11.41 7.3 NOTE: NHIS is National Health Interview Survey and and HIES is Health Interview Evaluation Survey. 42 Table 5. Prevalence of selected chronic conditions in persons 75 years of age and over from National Health Interview Survey estimates and from Health Interview Evaluation Survey household reports, by condition Number of Percent overreports per overreports 100 persons, by HIES NHIS HIES HIES compared compared Condition name and NHIS recode number prevalence prevalence with NHIS with NHIS ACONBIIONS . os ws sm rws sms wd ama ms smims 25% 2,986.5 3,319.3 332.8 11.1 5TH N 101 554.5 475.9 -78.6 -14.2 DSBS 0 vs is 55 wm 00 3 8 wns #7 31% 434 03 113 32.9 78.3 45.4 138.0 Blindness or other visual impairment . . . ...... 201 101.7 96.4 -53 -5.2 Deafness or other hearing impairment. . . . . .... 203 360.3 433.7 73.4 20.4 Deformity or orthopedic impairment . . . ....... 228 177.0 247.0 70.0 39.5 TIONS. « ms abe 5 «6 B05 eg +5 GLE @1 8 vse i) £ 268 Ft Bye 0 240 68.9 144.6 75.7 109.8 CHAOS, vs worms trons ame mph AE®s HE om 241 234.3 265.1 30.8 13.1 ICONSHPAION. a1 + 5s tw 54:7 0 oi 8 0 800 a5 0 a 4 Io 314 92.2 90.4 -1.8 -2.0 DIBDBIBE & ais oo woe i 4 05 50 5 005 5 5 als om 40% M08 of 01 5 403 85.7 108.4 22.7 26.5 Migraine headache ... cs vss ms ws smewns« 406 11.8 18.1 6.3 53.2 HERI IBOB8B . i» ws vi mos wis ie al or 5k are 20 353.0 373.5 20.5 5.8 Ischemic heart disease . . ................ 502 173.0 132.5 -40.5 -23.4 Heart thythm disorders. . .. cx «vs wv vis wn sme ans 89.1 144.6 55.5 62.3 Tachycardia or rapid heartbeat . . .. ......... 503 28.1 54.2 26.1 92.9 Heart MUMMIES » . «ic cs si nna os mo aan s 504 31.1 42.2 114 35.6 Other heart rhythm disorders . . . ........... 505 20.9 48.2 18.3 61.2 Other selected diseases of heart. . . . ........ 507 90.9 108.4 17.5 19.3 High Blood DIBSBUIS. u:i. « wis « wa wher wits # 0h #18 srw in 508 375.6 403.6 28.0 7.5 Hardening ofthe arteries. ... .. ..... ov vans 510 73.3 60.2 -13.1 -17.8 Varicose veins, lower extremities. . . ......... 513 86.6 90.4 3.8 4.3 HOmMOMNOIS « ..c s+ vos us ws sur sme ws dass 514 575 108.4 50.9 88.6 Chronic BIONCHIBL.. . 5.5 cis 4 sw wv 3 0 wih 4 wud 601 57.6 54.2 -3.4 -5.9 ASHAINB i 0% 5: loon 5 E15 fs 2 Sip 8 fo S00 5 of beuoh 8 602 42.3 12.0 -30.3 -71.5 Allergic rhinitis without asthma . . ........... 603 65.5 78.3 12.8 19.6 CRIONICSINUSIHE +o « vote vv was 76 £00k asa 605 155.8 168.7 12.9 8.3 NOTE: NHIS is National Health Interview Survey and HIES is Health Interview Evaluation Survey. 43 Table 6. Comparison of chronic condition reports for list-sample persons from Health Interview Evaluation Survey interviews and medical records and Kappa values, by condition Overreport by NHIS HIESS prevalance interview compared chronic Matching status according to— with medical record condition recode Positive False False Negative Medical Kappa Condition name number match! positive? negative? match® Interview record Net Percent value All conditions . . . ........... Han 1,055 1,325 906 19,829 2,368.2 1,951.2 416.9 21.4 0.433 Alls, ovnnsn mr msm ve 101 141 155 73 636 294.5 212.9 81.6 38.3 0.406 Dermatitis . . .............. 113 23 33 82 867 4 104.5 -48.8 —46.7 0.230 Blindness or other visual impairment. . . .. sss 0s seen 201 12 44 6 943 55.7 17.9 37.8 21141 0.305 Deafness or other hearing impairment. . ............. 203 53 102 22 828 154.2 74.6 79.6 106.7 0.401 Deformity or orthopedic Impairment: = +s sass ss vas 228 39 127 72 767 165.2 110.4 54.7 49.5 0.172 Tinnitus. . ....... LLL 240 7 49 9 940 85.7 15.9 39.8 250.0 0.174 Cataracts. . ov + «+ ava wn gms 241 56 27 71 851 82.6 126.4 —43.8 -34.6 0.482 Constipation. « : «+ sss criss an 314 6 44 27 928 49.8 32.8 16.9 51.5 0.109 Diabetes. . ............... 403 118 3 40 844 120.4 157.2 -36.8 -23.4 0.822 Migraine headache . ......... 406 14 35 16 940 48.8 29.9 18.9 63.3 0.330 Heart disease. ............. cn da ae ps sone 281.7 350.2 -98.5 -28.1 Blt Ischemic heart disease . . . . .. .. 502 63 14 52 876 76.6 114.4 -37.8 -33.0 0.622 Heart rhythm disorders . . . . . . .. $a shh rin vad is 109.5 110.4 -1.0 -0.9 sais Tachycardia or rapid heartbeat . . . 503 15 23 23 944 37.8 37.8 0.0 0.0 0.371 Heart mums . «5 swn ss ess 504 4 43 8 950 46.8 11.9 34.8 291.7 0.119 Other heart rhythm disorders . . . . 505 7 18 54 926 249 60.7 -35.8 -59.0 0.132 Other selected diseases of the healt. . wows ss sms seas 20s 507 40 26 86 853 65.7 125.4 -59.7 —47.6 0.362 High blood pressure. . . ....... 508 346 59 75 525 403.0 418.9 -15.9 -3.8 0.725 Hardening of the arteries. . . . . . . 510 1 18 13 973 18.9 13.9 5.0 35.7 0.045 Varicose veins, lower extremities . . 13 10 70 14 911 79.6 23.9 55.7 233.3 0.162 Hemorrhoids . ............. 514 26 64 41 874 89.6 66.7 22.9 34.3 0.276 Chronic bronchilis y+ i + 5 ws 2 5 5.5 601 5 40 31 929 44.8 35.8 9.0 25.0 0.087 Asthma. ................. 602 25 14 24 942 38.8 48.8 -10.0 -20.4 0.549 Allergic rhinitis without asthma . . . 603 18 132 28 827 149.3 45.8 103.5 226.1 0.122 Chronic sinusitis . . . ......... 605 26 185 39 755 210.0 64.7 145.3 224.6 0.099 Positive match means that both the interview and medical report were positive. 2False positive means that the interview was positive but the medical record negative. False negative means that the interview was negative but the medical record positive. 4Negative match means that both the interview and medical record were negative. SHIES is Health Interview Evaluation Survey. 44 Table 7. Comparison of chronic condition reports for household members from Health Interview Evaluation Survey interviews and medical records, and Kappa values, by condition Overreport by NHIS HIES® prevalance interview compared chronic Matching status according to— with medical record condition recode Positive False False Negative Medical Kappa Condition name number match! positive? negative? match* Interview record Net Percent value All conditions . . . ........... yah 295 412 393 15,069 1,005.7 978.7 27.0 2.8 0.397 AAS. 4 5s ww pis aw swe ww 101 34 34 28 607 96.7 88.2 85 9.7 0.475 Dermatitis . . .............. 113 11 10 74 608 29.9 120.9 -91.0 -75.3 0.168 Blindness or other visual IMPAIMBME. ; os ws mw sms ww 201 1 8 3 691 12.8 5.7 74 125.0 0.147 Deafness or other hearing impairment. . ............. 203 17 31 3 652 68.3 28.4 39.8 140.0 0.479 Deformity or orthopedic IMPAIMeNt, : +5 20s wma s wwe 228 8 39 36 620 66.9 62.6 4.3 6.8 0.119 Tinnitus. . . . LL 240 4 13 2 684 24.2 8.5 15.6 183.3 0.339 Calaracts: «ino cvs ms snsmns 241 10 6 27 660 22.8 52.6 -29.9 -56.8 0.357 Constipation. . . ............ 314 4 12 12 675 22.8 22.8 0.0 0.0 0.233 Diabetes . sw « ix ux wiv mms mie we » 403 19 2 11 671 29.9 42.7 -12.8 -30.0 0.736 Migraine headache . ......... 406 4 12 4 683 22.8 11.4 11.4 100.0 0.328 Heart disease. ............. rn JAN wl ro th 78.2 108.1 -29.9 -27.6 lid Ischemic heart disease . . . . .... 502 13 5 9 676 25.6 31.3 -5.7 -18.2 0.640 Heart rhythm disorders . . . . . . .. re ins wn $a Go 39.8 35.6 4.3 12.0 wwe Tachycardia or rapid heartbeat . . . 503 2 5 7 689 10.0 12.8 -2.8 -22.2 0.242 Heart murmurs; «1s «+ se sx smn: 504 0 17 4 682 24.2 5.7 18.5 325.0 -0.009 Other heart rhythm disorders . . . . 505 1 3 1 688 57 174 -11.4 -66.7 0.117 Other selected diseases of the heart, ash 9s ms SHER 5 20S 507 5 4 24 670 12.8 41.3 -28.4 -69.0 0.248 High bIGod pressure. . «.. +. « vv . 508 85 22 31 565 152.2 165.0 -12.8 -7.8 0.718 Hardening of the arteries. . . . . . . 510 1 5 1 696 8.5 2.8 5.7 200.0 0.247 Varicose veins, lower extremities . . 513 2 14 6 681 22.8 11.4 11.4 100.0 0.154 Hemomholds : . vv us sway wes 514 10 27 13 653 52.6 32.7 19.9 60.9 0.305 Chronic bronchitis . . ......... 601 3 18 13 669 29.9 22.8 74 31.3 0.140 Asthma, sews mv wos ww wwe wes 602 32 i 30 630 61.2 88.2 -27.0 -30.6 0.579 Allergic rhinitis without asthma . . . 603 14 65 21 603 112.4 49.8 62.6 125.7 0.190 Chronic: SIBUSHIS . +. vow w aw 605 15 49 23 616 91.0 54.1 37.0 68.4 0.243 Positive match means that both the interview and medical report were positive. 2False positive means that the interview was positive but the medical record negative. 3False negative means that the interview was negative but the medical record positive. 4Negative match means that both the interview and medical record were negative. SHIES is Health Interview Evaluation Survey. 45 Table 8. Comparison of chronic condition reports for list-sample persons from Health Interview Evaluation Survey interviews and medical records sorted by Kappa values, by condition Overreport by NHIS HIES® prevalance interview compared chronic Matching status according to— with medical record condition recode Positive False False Negative Medical Kappa Condition name number match! positive? negative® match Interview record Net Percent value Alcondifions . . : «vx cvv wn sums tig 1,055 1,325 906 19,829 2,368.2 1,951.2 416.92 21.4 0.433 DEBBIE. cous empresa rzas 403 118 3 40 844 120.4 157.2 -36.82 -23.4 0.822 High blood pressure. . . ....... 508 346 59 75 525 403.0 418.9 -15.92 -3.8 0.725 Ischemic heart disease . . . . .. .. 502 63 14 52 876 76.6 114.4 -37.81 -33.0 0.622 Asthma. ................. 602 25 14 24 942 38.8 48.8 -9.95 -20.4 0.549 CRarapls. «vss vans nase’ 241 56 27 71 851 82.6 126.4 —43.78 -34.6 0.482 Arthritis. . ................ 101 141 155 73 636 294.5 2129 81.59 38.3 0.406 Deafness or other hearing HPAL. vv. cers Anais o 203 53 102 22 828 154.2 74.6 79.60 106.7 0.401 Tachycardia or rapid heartbeat . . . 503 15 23 23 944 37.8 37.8 0.00 0.0 0.371 Other selected diseases of the NEB: ok 205 20 als ameRs BE 507 40 26 86 853 65.7 125.4 -59.70 -47.6 0.362 Migraine headache .......... 406 14 35 16 940 48.8 29.9 18.91 63.3 0.330 Blindness or other visual INpaMent. «sess aW sl £8 201 12 44 6 943 55.7 17.9 37.81 211.9 0.305 Hemorrhoids «co vasa vavue ves 514 26 64 41 874 89.6 66.7 22.89 34.3 0.276 Dermatitis «x vs sea vss sunk is 113 23 33 82 867 55.7 104.5 -48.76 -46.7 0.230 TinnlUS. « or ch om din Bd lh 240 7 49 9 940 55.7 15.9 39.80 250.0 0.174 Deformity or orthopedic impairment. .. . «cons wrviainn 228 39 127 72 767 165.2 110.4 54.73 49.5 0.172 Varicose veins, lower extremities . . 513 10 70 14 911 79.6 23.9 585.72 233.3 0.162 Other heart rhythm disorders . . . . 505 7 18 54 926 24.9 60.7 -35.82 -59.0 0.132 Allergic rhinitis without asthma . . . 603 18 132 28 827 149.3 45.8 103.48 226.1 0.122 Heart MUIMUIS « + « « vo v vv vie wv 504 4 43 8 950 46.8 11.9 34.83 201.7 0.119 Constipation. « : =.» «vas. sive oa 314 6 44 27 928 49.8 32.8 16.92 51.5 0.109 Chronicsinusitis . . .......... 605 26 185 39 755 210.0 64.7 145.27 224.6 0.099 Chronic bronchitis . . . . ccs. + 601 5 40 31 929 44.8 35.8 8.96 25.0 0.087 Hardening of the arteries. . . . . . . 510 1 18 13 973 18.9 13.9 4.98 35.7 0.045 Positive match means that both the interview and medical report were positive. 2False positive means that the interview was positive but the medical record negative. 3False negative means that the interview was negative but the medical record positive. 4Negative match means that both the interview and medical record were negative. SHIES is Health Interview Evaluation Survey. 46 Table 9. Comparison of chronic condition reports for household members from Health Interview Evaluation Survey interviews and medical records, sorted by Kappa values, by condition Overreport by NHIS HIESS prevalance interview compared chronic Matching status according to— with medical record condition recode Positive False False Negative Medical Kappa Condition name number match’ positive? negative® match* Interview record Net Percent value All conditions . . . ........... 295 412 393 15,069 1,005.7 978.7 27.0 2.8 0.397 Diabetes... .............. 403 19 2 i] 671 29.9 42.7 -12.8 -30.0 0.736 High blood pressure. . . ....... 508 85 22 31 565 152.2 165.0 -12.8 -7.8 0.718 Ischemic heart disease . . . . .... 502 13 5 9 676 25.6 31.3 -5.7 -18.2 0.640 ASE. ov sus n ses en smrna 602 32 11 30 630 61.2 88.2 -27.0 -30.6 0.579 Deafness or other hearing impairment. . ... cetera sen 203 17 31 3 652 68.3 28.4 39.8 140.0 0.479 ARIUS. «vc ons smrre ns nes 101 34 34 28 607 96.7 88.2 8.5 9.7 0.475 Calaractss » 15 v5 5 4 6o8a 35 0s 241 10 6 27 660 22.8 52.6 -29.9 -56.8 0.357 TIONS. vo vivre om we ow wim 240 4 13 2 684 24.2 85 15.6 183.3 0.339 Migraine headache . ......... 406 4 12 4 683 22.8 11.4 11.4 100.0 0.323 HEMOINOIIS 5 vv + ws wee wins 514 10 27 13 653 52.6 32.7 19.9 60.9 0.305 Other selected diseases of the HEE: wo vin swe s pT Aa 507 5 4 24 670 12.8 41.3 -28.4 -69.0 0.248 Hardening of the arteries. . . . . . . 510 1 5 1 696 8.5 2.8 5.7 200.0 0.247 Chronic: Sinusitis « . + + ws x24 5 2s 605 15 49 23 616 91.0 54.1 37.0 68.4 0.243 Tachycardia or rapid heartbeat . . . 503 2 5 7 689 10.0 12.8 -2.8 -22.2 0.242 Constipation. « «+ «+s saws ss + 4 314 4 12 12 675 22.8 22.8 0.0 0.0 0.233 Allergic rhinitis without asthma . . . 603 14 65 21 603 112.4 49.8 62.6 125.7 0.190 Dermatitis . cu sv vwn sms wwe 113 11 10 74 608 29.9 120.9 -91.0 -75.3 0.168 Varicose veins, lower extremities . . 513 2 14 6 681 22.8 11.4 11.4 100.0 0.154 Blindness or other visual Impaiment. . « . «ox auniinses 201 1 8 3 691 12.8 5.7 74 125.0 0.147 Chronic bronchitis . . ......... 601 3 18 13 669 29.9 22.8 2a 31.3 0.140 Deformity or orthopedic IMPBIMBM vo. ¢ 00 wie mie «wes 228 8 39 36 620 66.9 62.6 4.3 6.8 0.119 Other heart rhythm disorders . . . . 505 1 3 11 688 57 17.1 -11.4 -66.7 0.117 HEE MUAMUIS . . «viv oes iim 3 504 0 17 4 682 24.2 57 18.5 325.0 -0.009 Positive match means that both the interview and medical report were positive. 2False positive means that the interview was positive but the medical record negative. False negative means that the interview was negative but the medical record positive. 4Negative match means that both the interview and medical record were negative. SHIES is Health Interview Evaluation Survey. 47 Table 10. Percent of matching positive reports for selected conditions in three studies, by condition Kaiser Permanente study’ Health Insurance Plan study? Health Interview Evaluation Survey Percent postive Percent positive Percent positive reports in— reports in— reports in— Records Interview Records Interview Records Interview matched matched Net matched matched Net matched matched Net by by over- by by over- by by over- Condition name interview records report interview records report interview records report Arthritis and chronic rheumatism . . . . 68.5 51.3 33.7 33.2 33.2 -26.4 65.9 47.6 38.3 Chronic skin diseases. . . ........ 34.5 75.9 -54.5 19.5 19.5 -54.5 21.9 41.1 —46.7 Severe or other visual impairment . . . 72.0 57.3 25.6 33.3 33.3 -15.3 66.7 21.4 211.3 Hearing impairment . . . ......... 72.0 35.0 106.0 41.2 41.2 83.9 70.7 34.2 106.7 Deformity or orthopedic impairment . . 57.8 47.3 22.3 334 33.4 25.1 35.1 23.5 49.5 Diabetes . . . ................ 80.7 98.6 -18.2 61.7 61.7 -11.6 74.7 97.5 -23.4 Headache and migraine, chronic. . . . 62.2 47.1 32.2 14.9 14.9 -10.8 46.7 28.6 63.3 Diseases of the heart, NEC . . ..... 79.4 77.3 29 60.5 60.5 7.5 36.6 51.0 -28.1 Hypertension, NEC . ........... 81.1 64.6 25.6 45.8 45.8 -0.7 82.2 85.4 -3.8 Rheumatic fever, arteriosclerosis, NEC, and other chronic circulatory CONIIONS wvvp sims vm cpm sms 39.4 27.1 45.5 on. en en. 371 35.3 335.7 VariCoSB VEINS sus os sis we v5 ow 48.1 47.6 1.2 42.3 42.3 135.0 41.7 12.5 233.3 Hemorrhoids . ............... 66.4 45.3 46.6 38.2 38.2 93.9 38.8 28.9 34.3 Chronic bronchitis . . . .......... 79.2 31.1 154.2 65.0 65.0 306.3 13.9 11.1 25.0 ASHIMR. sins vmsumawsmapmuns 69.2 49.1 41.0 76.2 76.2 857.4 51.0 64.1 -20.4 Allergic rhinitis without asthma . . . . . 73.2 52.6 39.0 “4 “4 “4 39.1 12.0 226.1 Chronic sinusitis . . ............ 100.0 20.9 378.9 48.4 48.4 160.2 40.0 12.3 224.6 1Balamuth (1), Harlow and Linet (17). 2Madow (2) (3), Harlow and Linet (17). includes only arteriosclerosis. 4Combined with asthma. NOTE: NEC is not elsewhere classified. Table 11. Prevalence of heart conditions in list-sample persons from the Health Interview Evaluation Survey, by source of information and condition Person-level prevalence Condition-level prevalence Interview Medical Net Interview Medical Net Condition report repord difference report record difference Heart di8eass. o sucws sasme smsun 251.7 349.3 -27.9 264.7 471.6 —43.9 Heart disease (without heart murmurs) . . 205.0 337.4 -39.2 217.9 459.7 -52.6 Ischemic heart disease . . .......... 76.6 114.4 -33.0 88.6 1771 -50.0 Heart rhythm disorders . . . . ........ 109.5 109.5 0.0 109.5 125.4 -12.7 Tachycardia or rapid heartbeat . . . . . . . 37.8 36.9 2.6 37.8 41.8 -9.5 Heart MUITIIIS © wv win sams suis ma « 46.8 1.9 291.7 46.8 11.9 291.7 Other heart rhythm disorders . . . . .... 24.9 60.7 -59.0 24.9 71.6 -65.3 Other selected diseases of heart. . . . . . 65.7 125.4 -47.6 66.7 169.2 -60.6 48 -sau0ba)ed pauIquIoD ay} WoL SINULNW ueay sdoip g yojew asoo| ‘SUOIPUOD Ueay 104 "9p0dal SIHN Jayla ul Jou Sapod WO-6-00I OMI SPPE g Ydjew asoo| ‘suonipuod Alojendsas saddn 104 "sayojewsiw Se |[am se sayojew aAnebau 0) euao yorew papuedxa ay) saydde z yolew 2s007g 'sep0o2al SIHN Buiuiqwoo jo ynsa1 ay) si | yojew asoo) ‘suonipuod Alojeundsas seddn pue ueay io4 ,,'aAnebau asfey,, pue ,aansod aspey,, sadA} yojewsiw 0) Aluo endo yojew papuedxa ay) seydde | yojew 85007, ‘Aang mainselu| yyesH [BUOHEN S! SIHN; 1010 SGL 6€ G81 9¢ siisnuis dluoIyy es * 2432 2210 128 82 eet 81 us aiBsalY ¥91°0 or9 0S 0s2 S9 10 9v9 sv vse LS ak B24 Be Te He et swe|qoid Aiojendsal seddn st sree sre ces cee si oe ij re a 2960 esa og 9z op eee saseasip Heay 18410 ’ 2EL0 926 vs 81 Zs aa fe st 2 oe siapiosip wyikys Jay : ! 6LL°0 0S6 8 114 |Z sinuunw pes ce Hap 1LE0 v6 £2 154 St "© 1eaqueay pidel Jo eipedkyoe] L0v°0 888 SS 8c ve 8ve€0 v8 1S 29 cy Tr tr TU oo re si1apiosip whys peasy ee wr Ee ng at righ ak sn heel Se 2290 9.8 25 vl £9 ree aseasip 1eay Iwas] 8650 veL 96 9€ 6€1 0850 904 68 8S cst tas atl EER pie oo rrr SuOIpUOd pesH 2€9'0 698 9S LH 69 9890 2cl6 8 L 81 Sv0°0 cle El 8} I ssuaue ay) jo Buiuspiey rife bth ses Th ree 8580 198 08 cz ce 2620 198 28 ee £2 rreeeeeses snewsaq 6EY°0 019 £6 vel 891 98v°0 9€9 0s cel 291 vOov'0 9€9 €L GS1 Wwe corre A i shiiypy anjea yojew anjebau aasod yoew anfea yojew anjebeu amyisod yojew anjea yoyew anjebau aaysod yojew uonipuo) eddey aneban ase ase aAINISOd eddey annebsp as|e4 asfe4 BAINISO4 eddey) anneban as|e4 as|e4 BAISOd ¢ Z Yojew 8soo 21 Yojew 8soo $8p00aI | SIHN uonipuod pue ysew jo adh Aq ‘AaAing uonenjeas MaiaIalu] YieaH ay} pue saipnjs OM} 10} sanjeA eddey} pue suoiIpuod Jo Jaqunp "Z| aiqel 49 Table 13. Number of conditions, number and percent of overreports, and Kappa values from Health Interview Evaluation Survey responses and medical records for list-sample persons, by selected characteristics Conditions per person Interview overrport Number of By HIES! By medical Net Percent Kappa Person characteristic persons interview record overreport overreport value Age URC 45 YBBIS. » vi cv: + 51:5 ws v5 wie od ws wi 6 iw 309 1.60 0.98 0.62 63.6 0.348 ABBY YOBIB: iv ¢ ins + # 150, + 006 i Be 8 1 aides & GL oc 4 373 2.30 1.85 0.45 24.3 0.464 BETA YCBIB:: vc cv + ivi 5 wie nor» oo) Win 50 3 om wary vi 0 oo hw 193 2.95 2.63 0.33 12.4 0.466 7BYears and Over . . vcs spe cnr mis ER ERB EES 130 3.53 3.56 -0.03 -0.9 0.409 Sex and age Female: BE YEAS AOL OVBF. vo vs eww 5 vin mom mn mew 179 3.37 3.07 0.30 9.8 0.434 VYE-BAYOAIS 4 vv ww wv os iv 0 isis Ho 4 51% eB 366 2.23 1.48 0.74 50.1 0.420 Male: BE YAS ANA OVBE. 1. sc vv mis wi wv erm ww iw sb 144 2.96 2.92 0.03 1.2 0.450 VB-BAYEBIE ovis wi # min if & od 30s orion 00 ot hed 316 1.70 1.42 0.28 19.6 0.435 Race and age Black: BBYSars and Over. « » «saws wm swig was saad 156 3.37 3.07 0.30 9.8 0.474 18=BA YEAS . ovis nines inn oft Gr 8 mit 0 418 is. ke 518 1.96 1.43 0.54 37.6 0.433 White or other: BE YORIS BIVAOVBF. «5 vivid sms msm vine fine 167 3.01 2.94 0.07 24 0.408 VB-BAYCRIB + vv viva vr rms comms smsmenss 164 2.05 155 0.50 32.3 0.406 Employment status and age Employed: BBYoas and Over. «cc ccc n snr EE mEs 71 2.66 2.69 -0.03 -1.0 0.428 18-64 years 595 1.91 1.38 0.53 38.2 0.417 Unemployed: B5yearsand over. . ......... 0... 251 3.35 3.10 0.24 7.8 0.443 18-BAYRAIS i» vs ws svs wisi s nw sven 86 2.50 1.98 0.52 26.5 0.473 Income ) BO-BVG,900 . «vivir mw iv 4 ene ww Bin sh on 148 3.07 2.46 0.61 25.0 0.434 $20,000-529.999. . \ ox wine wn pga 8 EES EES 119 2.61 2.02 0.60 29.6 0.423 $30,000-$49,999. . . . .. 263 2.34 1.96 0.38 19.6 0.452 $80,000 and OV8Y ov wn was ss wae wy vey ve wns 335 2.15 1.73 0.42 241 0.431 Education Less than high shook. « co « ss wis ws mvs os sus 167 2.75 2.31 0.44 19.2 0.445 Highschoolgraduate. . ..........cccno:v.. 307 2.21 1.77 0.45 25.2 0.436 Some college. . «. . «ciara 204 2.44 1.76 0.68 38.3 0.434 College graduate: ois + + woe 35 Bwg 1s Hae sr Ewe 321 2.25 2.02 0.23 11.4 0.434 Whether 2-week doctor visit and age 2-week doctor visit: B5 YBATS BNTIOVEE. 7 «viv ve min win wie 31 wen esi 170 2.87 3.07 -0.20 -6.5 0.452 18-BA YEAS «vw wwe 8 wae ny wEs va BEE 280 2.18 1.71 0.47 27.6 0.445 No doctor visit: BBYBaIs and OVeY. . « «vc vv wr ws pus vw wan 153 3.54 2.93 0.61 20.8 0.431 AB—BHYBBIS «iv. ek 50a 8 800 a 0 thie 402 1.85 1.28 0.57 44.3 0.409 Whether 13-month hospital stay and age Inpatient stay in past 13 months: BS years ANA OVBT. . « «cn vv vow oo vis wn smins 108 3.37 3.52 -0.15 -4.2 -0.398 18-BAYBAIS «vpn 5 nye 65 poms Baws ous 138 2.35 1.83 0.52 28.6 0.459 No inpatient stay in past 13 months: BEYyears ang OVeL. «« cn w sx vam rr www yw 2 wun 215 3.09 2.74 0.35 127 0.465 VBBRAYBEIS 0 45 ives ihe of irk tn) Fld 5 8 oe ie 544 1.89 1.36 0.53 38.9 0.415 Whether health assessment past 2 years and age Health assessment past 2 years: Boyes BNO OVEL. . « vos sx raw vir sev ax sus 134 3.00 3.01 -0.01 -0.5 0.461 Under 1year-64vyears. ................. 246 1.97 1.55 0.42 273 0.437 No health assessment past 2 years: BBYears BNO OVEY. . . cuc ssn um sa wae a5 EE 189 3.32 2.99 0.32 10.8 0.427 Under 1year-64vyears. . ................ 436 1.99 1.40 0.59 41.8 0.420 THIES is Health Interview Evaluation Survey. 50 Table 13. Number of conditions, number and percent of overreports, and Kappa values from Health Interview Evaluation Survey responses and medical records for list-sample persons, by selected characteristics — Con. Conditions per person Interview overrport Number of By HIES? By medical Net Percent Kappa Person characteristic persons interview record overreport overreport value Self-perceived health status Excellent. . vo ms ns vvrmsrasmsnssnne nse 203 1.41 1.31 0.10 7.5 0.395 VEIY QOD « 5s wis ints mis 1a 0 4 68 8 0c 2 0a oh i 6 293 2.13 1.65 0.48 29.1 0.432 BOOH ..ienr wns mammens mass ea pms wash 296 2.62 2.11 0.51 24.4 0.445 FU OrPODl sv: vases emi mtwnintepmemewmemn 201 3.31 2.80 0.51 18.1 0.446 Number of chronic conditions NONE: 2; satms snibanrmens catni hmsms so 139 0.00 0.81 -0.81 -100.0 0.000 ORB. ox vs mms ws wn mms RAE RR BREE CR 235 1.00 1.34 -0.34 -25.6 0.412 TWO vs ra onic ns i Gemma BRrimn sams Imes 248 2.00 1.74 0.26 14.8 0.483 THIER. vv vvemu nmr comms susie emma ay 160 3.00 2.26 0.74 33.0 0.447 FOUTOTMOI® , vos visms su ems sms bs cmsms ms 223 5.25 3.32 1.93 58.0 0.405 51 Table 14. Number of conditions, number and percent of overreports, and Kappa values from Health Interview Evaluation Survey responses and medical records for household members, by selected characteristics Conditions per person Interview overreport Number of By HIES By medical Net Percent Kappa Person characteristic persons interview record overreport overreport value Age O-17 YORE. un sn stn 20% 2 pT LYE BWW E DAE 285 0.40 0.59 -0.19 -32.3 0.345 1B-BAYBAIS. «rons 4 divin Bw tsi $08 mh 08) Re 173 0.77 0.53 0.25 47.3 0.348 ABBR YOGA 10x 50 400 30 Bar SE 8 Re 138 1.56 1.52 0.04 2.4 0.401 BETAYOBIB. sc oni win wim darwin Yims Bg 08 71 221 1.82 0.39 21.7 0.483 7BYeaISANAIOVBE « vooiw » vw bin) #0 919 tw pike moms Rite iwi 36 2.56 2.69 -0.14 -5.2 0.344 Sex and age Female: BB Years ANC OVBF. . « « «x cox cv sw x nw aim ves 63 2.13 2.05 0.08 3.9 0.398 UNder1 year-B4A Years. « « + «sv vv vw vw ww wwe 330 0.86 0.86 0.00 0.0 0.378 Male: BBYCAIS BNO OVBE. «os vs vx 2 4 9h nus wp dn woe 44 2.61 2.20 0.41 18.6 0.468 Underiyear-64years. .................. 266 0.67 0.70 -0.02 -3.2 0.373 Race and age Black: BB Yoars ant OVBY. «s+ + svi + wos boa sini ow aie 37 2.38 1.78 0.59 33.3 0.505 Under 1year-64 years. . ................. 452 0.81 0.81 -0.00 -0.3 0.355 White or other: BEYOAS BNO OVBY., is vu vin hk 3 088% Faia snes 70 2.30 2.29 0.01 0.6 0.393 Under 1year-BA Years. ...us vv ssa nns prs .144 0.68 0.72 -0.03 -4.9 0.452 Employment status and age Employed: BE YSMS ANCOVEE. vs vans + Win tm a was aw ae 23 2.09 1.96 0.13 8.7 0.380 ME=BANYOBIS 1 + 000 9000, 0 540 10 1 0 Wiis, 1 Lola 244 1.00 0.89 0.11 12.4 0.385 Unemployed: 65yearsand over. . ...........00n nnn 84 2.39 2.15 0.24 11.0 0.441 UBB YBBIS i «16s v0 wi 00 ike oy 0 0 Bow Jo i 690 fo or 66 1.56 1.24 0.32 25.6 0.388 Income B0-519,900 vic 5 wv 5 wm eR 0 BR 32 1.72 1.25 0.47 37.5 0.296 $20,000-329.990. - +5 7m sien HS 2k Bis Bn ple 51 1.16 1.06 0.10 9.3 0.380 $30.000-849909..... . x = 5.5 40 «0 5 avi wi 4a Bn 8m 198 0.97 1.05 -0.08 -7.2 0.431 $B0, 000A OVBY sv 5s 5 vs ms @ a ® 6» % ww 355 373 0.95 0.91 0.04 4.4 0.384 Education’ Lessthan high School. « . . «vu cs rsa sw ave sas 69 2.04 1.49 0.55 36.9 0.401 High school graduate. . . . « «x oc ven ens en eames 138 1.17 0.97 0.20 20.1 0.366 SOMB CONBGB.: « 4m se = 1 iwi wal Wire 5, 48 win 16 0m WilW 81 1.26 1.12 0.14 12.1 0.451 Collegs graduale. i » vx tw +088 Faia 8 9m eds © 0 wie 128 1.49 1.53 -0.04 -2.6 0.420 Whether 2-week doctor visit and age 2-week doctor visit: BEBYSIBANCIOVEE. . « vv nn nn wir iv v3 swim ao wn 26 2.50 2.81 -0.31 -11.0 0.433 Underiyear-64years. .................. 77 1.25 1.34 -0.09 -6.8 0.380 No doctor visit: BEBYearand OVE « co svuwe rummy avvn ene 81 227 1.89 0.38 20.3 0.428 Under 1 year-BA Years. . «crus vs ss ans smu 519 0.71 0.70 0.00 0.3 -0.377 Whether 13-month hospital stay and age Inpatient stay in past 13 months: BS years aNAIOVEP. .. « «uw si © « m sinie wor kin® § wb 11 3.55 3.45 0.09 2.6 0.351 Underiyear-84years. ..... v0 v vee mans 42 1.50 1.57 -0.07 -4.5 0.364 No inpatient stay in past 13 months: BBYears an OVI, « c: vr sa sms vs sua sm anes 96 2.19 1.96 0.23 11.7 0.441 Under year-B4 years, . . . «es cv rus wwe wen 554 0.72 0.73 -0.01 -0.7 0.376 Whether health assessment past 2 years and age Health assessment past 2 years: BBYear8 ANCOVBE, , uv vs viv v wa giv rsa eine wns 52 2.38 2.54 -0.15 -6.1 0.506 Underiyear-B4vyears. . ............0004. 17 0.91 0.92 -0.01 -0.9 0.377 No health assessment past 2 years: BEBYSarS BNO OVBE., uv. vv sins wv sins mn ows «nw 55 2.27 1.71 0.56 33.0 0.341 UNOSEY yBar-BAYBaIS. . « «vst vn v mr vw ewes 479 0.74 0.54 0.20 36.5 0.445 Persons 18 years of age and over only. 52 Table 14. Number of conditions, number and percent of overreports, and Kappa values from Health Interview Evaluation Survey responses and medical records for household members, by selected characteristics —Con. Conditions per person Interview overreport Number of By HIES By medical Net Percent Kappa Person characteristic persons interview record overreport overreport value Self-perceived health status Excalient. . «i: vs unis ns rns sR IMIR BE 274 0.53 0.55 -0.02 -4.0 0.330 VOY GOOG & sii 1 i 4 or id ii 0 oh § 4 0h 2 okie moms om com aw 210 0.86 0.92 -0.07 -7.2 0.390 BOO. . vs msm prs fe vs @s 3 BEBNE HWS 159 1.37 1.26 0.11 8.5 0.393 FPRIT OF POE w « oc visimn seis ampme nme wn nies mn 56 2.96 257 0.39 15.3 0.449 Number of chronic conditions NNOIG. 4.55 5 05 £6 0 305 wa ter # i 0 0 #0 030 ah fo 3 oi wt 337 0.00 0.45 -0.45 -100.0 0.000 ONB.s vss s ons Biwi dina taal 3 smu 4s WEE WEEN 68% 181 1.00 1.04 -0.04 -3.7 0.460 TI ics ia:00 5: 50 [31 55 2100 00 00 mn tno ve 104 2.00 1.38 0.63 45.5 0.406 TOTEG os 4 iw 000 i mi 9 0 6 4 7 8) $8 2 8 UE ADEE 3 39 3.00 2.64 0.36 13.6 0.467 FOUL OTIAOIS ov 000s nn ti 003 4h Koto 0 im om 42 4.88 2.62 2.26 86.4 0.379 Response status Adult present for interview . . . ....... LLL. 245 1.69 1.42 0.27 18.9 0.411 Adult not present for interview. . . ............. 183 1.03 0.98 0.05 5.0 0.404 53 Appendixes Contents 1. Health Interview Evaluation Survey QUESHOINGIIE evs uwsis sew view om ain 040 006 90.0 300 550 010 310.8 900 915 0 450 30 974 2 03% wi 8. 90 55 II. Health Interview Evaluation Survey abstracting procedures medical record coding guidelines. ................ 101 ITI. Loose match recommendations. . . «ov vttt tt tt treet ettteeeereeeeeeeeeeneeeeeeeeeeeeeeeeeeeeeeeeeenns 118 TV. Definitions of terms used 11 this TEPOTL: « vs + us wiswms win sis a semis wos sis mies ws 4.5.5 016 FERED 53:8 SH 54 0508 4% 5 $d 50s 120 Appendix | Health Interview Evaluation Survey Questionnaire CDC 64.01 OMB No. 0920-0239: Approval Expires 12/31/90 NOTICE - Information contained on this form which would permit of any or has been with a that it will be held in strict confidence, will be used only for purposes stated for this study, and will not be disclosed or released 10 others without the consent of the individual or the establishment in accordance with section 308(d) of the Public Health Service Act (42 USC 242m). Public} reporting burden for this collection of information Is estimated to vary from 18 to 35 minutes per response, with an average of 28 minutes per response. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to PHS Reports Clearance Officer, ATTN: PRA; Hubert H. Humphrey Bidg., Rm. 721-H, 200 Independence Avenue, SW; Washington, DC. 20201; and to the Office of Management and Budget, Paperwork Reduction Project (0920-0239), Washington, DC 20503. rorm HIS-1 (Evaluation) (2-1-90) WESTAT, INC. ACTING AS COLLECTING AGENT FOR THE U.S. PUBLIC HEALTH SERVICE HEALTH INTERVIEW EVALUATION SURVEY 1. Book of ___ books 2-5. Not applicable this form 6a. What is your exact address? (Include House No., Apt. No., or other identification, 14. Noninterview reason county and ZIP Code) TYPEA 01 JRefusal — Describe in footnotes ET pre pp ae Lo 02[JNo one at home, repeated calls City | State [County 7 ZIP Code 03] Temporarily absent — Footnote I | | 0a] other (Specify) 3 | ! ! Fill items b. Is this your mailing address? (Mark box or specify if different. [J] Same as 6a ! 0. 2 Include county and ZIP Code.) 12-15 city TT state TCounty ~~ 1ZIP Code J15. Record of calls | 1 | ! 1 1 1 Com- Items 7, 8, and 9 not applicable this form. Month ! Date Beginning Ending pleted 10. CLASSIFICATION OF LIVING QUARTERS — Mark by observation | 1X) 1 Items 10 a and b not applicable this form. P ai; ai c. HOUSING unit (Mark one, THEN page 2) 3 : 7 pm: pri 010 House, apartment, flat y P a.m. a.m. 02] HU in nontransient hotel, motel, etc. 2 | p.m. p.m. 03] HU — permanent in transient hotel, motel, etc. i 04[J HU in rooming house 3 | P a.m. 4.m- 05] Mobile home or trailer with no permanent room added ! T RM: pn. 06] Mobile home or trailer with one or more permanent rooms added ! P a.m. am 070] HU not specified above — Describe in footnotes a | 7 iy BR. d. OTHER unit (Mark one) , a.m. am. 5 1 08[J Quarters not HU in rooming or boarding house | T pm. p.m. 09] Unit not permanent in transient hotel, motel, etc. p ah. am: 10d Unoccupied site for mobile home, trailer, or tent 6 | p.m. p.m. 11 Student quarters in college dormitory L I 12] Other unit not specified above — Describe in footnotes 16. Not applicable this form 17. Record of additional contacts T GO TO HOUSEHOLD COMPOSITION PAGE Month! Date| Beginning Ending | sited | time time Mark — | (X) # a 1 What is the telephone | Area code/number 12. Interview observed? 1 ! p an a; number here? | ! T p.m. p.m. ! Y 0 None | 10ves 2[0No 2 P a.m. a.m. | | T p.m. p.m. 13a. Interviewer’s name | Code | b. Language of interview | P ar oy | | 3 T p.m. p.m. | | 1 [CJenglish 3 Osoth English and Spanish ) : 2 [Jspanish 8 [Jother | P a.m. a.m. \ | 4 | T p.m. p.m. 55 56 0 ow age A. HOUSEHOLD COMPOSITION PAGE 1 [O Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) d. Do all of the persons you have named usually live here? Probe if necessary: Does — — usually live somewhere else? Ask for all persons beginning with column 2: 2. What is — — relationship to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD ar 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. A3 Refer to ages of all related HH mambers. 4a. Are any of the persons in this family now on full-time active duty with the armed forces? Yes J No (5) b. Whoisthis? ~~~ Coon TooToo mmm Delete column number(s) —_—_ by an *’X"’ from 1—C2. c. Anyone else? [J Yes (Reask 4» and c) ONo Ask for each person in armed forces: d. Where does — — usually live and sleep, here or somewhere else? Mark box in person’s column. If related persons 17 and over are listed in addition to the respondent and are not present, say: 5. We would like to have all adult family members who are at home take part in the interview. Are (names of persons 17 and over) at home now? If “’Yes, ’’ ask: Could they join us? (Allow time) Read to respondent(s): This survey is being conducted to collect information on the nation’s health. | will ask about hospitalizations, disability, visits to doctors, illness in the family, and other health related items. HOSPITAL PROBE 6a. Since (13-month hospital date) a year ago, was — — a patient in a hospital OVERNIGHT? b. How many different times did — — stay in any hospital overnight or longer since (13-month hospital date) a year ago? 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name °*H b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes, enter A y fiames in columns « | Relationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies orsmalichildren? . .................. cc. i iii, a 0 : tof binh Date Year — any lodgers, b s, Or p you employ who livehere? . .............. ef 3 O : ; - anyolie Who LISUALLY Jyes here but is now away from home oO 0 HOSP. | WORK RD [2-WK.DV travelingorinahospital? . ................... iii iia N vo BRYONO BIO BIAYINGMBIOP . « sivivs s+ 0sninn vs 0 s@ stasis ssn sessasinesss «| 0 0 C1 pol] None 10wa [10 ves ool] None Number 200 we [200 No Number A3 | I rod 1 1 Ll 1 i A LA TRA T1DV_ TINJ.TCCLTRI HSTCOND. 1 | 1d 1 I L i le 1 1 LA IRA IDV IINJ.|CLLTRI HS|COND. | 1 ol I 1 1 1 1 1 A [LA ~ TRA T1DV TiNJ.TCCLTRI HSTCON DV |INJ. | CLLTRI HSICOND. ! I 1 | | | Can persons 65 and over (5) O other 14) pada 0 Living at home J Not living at home 1 Oves 2 [J No (Mark “HOSP. ** box, THEN NP) ““HOSP."’ box (Make entry in THEN NP) | ‘Number of times _ Ask for each child under one: 7a. 7a. Was — — bor in a hospital? Ask for mother and child: TTT TTT TTT b. b. Have you included this hospitalization in the ber you gave me for — —? 1 ves 2 ONo ive) O ves ve) [J No (Correct 6 and “HOSP. box) FOOTNOTES ‘FORM HIS-1 (Evaluation) (2-1-90) Page 2 Ooi age Cod age [J oid age [CJold age 4. | First name Mid. init. FAge First name Mid. init. ge 1. [First name Mid. init. ge First name Mid. init. fAge Last name ex Last name 8X Last name ex Last name ex 10m 10m 10m 10m Oe 200F 20F 20°F 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth «+ [Date of birth Date of birth Month | Date Year Month J Date | Year Month |Date [Year Month ' Date | Year 1 J oo 1 1 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00] None 10wal10 Yes 00] None |oo[J None 10wa [1 ves 00 None c1 00] None 10wa [10 ves [J None foo [J None 10Owa [10 Yes oo] None Number 200ws [200No Number | Number 20we 20INo Number Number 20we 12000 Number | Number 20we |20Ne | reer LA ~ TRA ~|DV |INJ. [CLLTR{HS T|COND.|TA~ ~ [RA T|DV |INJ.” | CLLTRIHS [COND. 1 I I 1 | 1 1 1 | 1 1 1 ro 1 | 1 11 1 I | 1 1 1 ocd 1 Jed, L Lenina 1 1-3 1 ded MA ~ TRE ~ BV INU" aL CONG. | TA ~ [RA™ TOV TIND. | TLITRHE [CONG CA™ ~)RA~ TDV [INJ. |TLLTATHS JCONG|LA ~ THA —| DV NJ. TCCLTR HS | COND. | [I 1 1 ro 1 1 1 1 1 [I | [I | ol 1 Ll 1 1 A 1 1 1 11 | 1 | Ld 1 Ll [LA = TRA ~|6V INT [nas ny TA 7 {RA™ TOV TINJ. 7 | TLITRTHS [TOND. CA™ T|RA™ TOV (INT. T|CLITRTHS — LA ~ THA 7 DV INJ. TCCLTR HS 7] COND. 1 ro | | 1 1 ro | ro | 1 | [I | [I | [| 1 1 1 1 1 1 1 1 1 1 3 1 1 1 l 1 1 1 | -l J 1 - ~ TRA TOV INU [CLTTRIHS COND. |TA™ ~ {RA TOV INI. 7 | CLITATHS = CAT TRA” TDV (INT. T|TLLTRATHS [CONO.JUA — THA | DV INJ. TCCLTR|HS 7 COND. 1 i 1 ro 1 [| | ro 1 1 | | I ro | [I 1 Li 1 1 | 1 | 1 fed 1 fe Eo 1 et LA = TRA “IDV INJT TCLTTRIHS COND. [TA~ ~ [RA TOV TINJ.”| CLITRIHS [COND. CA” TIRA™ TOV TINJ. TICLLTRIHS {CONGJUA ~ TRA ~| DV] INJ. TCLLTR|HS TT COND. | | | 1 bu 1) | 1 | | mn | [a | [I | j_ 1 1 [I | 1 | [I 1 | oe -— a : * 4d. [J Living at home [J Living at home 4d. oO Living at home [J Living at home Not living at home OI Not living at home OI Not living at home DO Not living at home — — | 1 6a.| 1[0ves 10 Yes 6a.| 10ves 10ves 20No (Mark *"HOSP.’* box, THEN NP) 2 0 No (Mark ‘HOSP.’ box, THEN NP) 200No (Mark ‘HOSP.’ box, THEN NP) 200 No (Mark “"HOSP."’ box, THEN NP) oo Sp Se MR i Hi [os smsomtion seins wii iis 50 tk iy om et FTA IAR Gn: irs mae a rt, mR ST (Make entry in (Make entry in (Make entry in (Make entry in Bb “’HOSP."* box, ‘“HOSP."’ box, b ’HOSP."’ box, “HOSP.” box, * | Number of times THEN NP) Number of times THEN NP) *| Number of times THEN NP) Number of times THEN NP) F — 7a.| ves 10 Yes 78.| ves 10 ves 20 No vp) 200No vp) 200No ive) 200No (vp) 0 Jot gyn ests ese ef JV A RS SE BEE wh omit epee vee sms ee AR Cr eR Ft et pe TES ee b. O ves vp) O ves ve) | Oves we O ves vp) 0 No (Correct 6 and “HOSP.” box) OJ No (Correct 6 and “HOSP.” box) OI No (Correct 6 and “HOSP.” box) OI No (Correct 6 and “HOSP. box) FOOTNOTES FORM TS 1 Evaluation) (2-1-4 301 Page 3 58 O ou age A. HOUSEHOLD COMPOSITION PAGE 1 Probe if necessary: Does — — usually live somewhere else? d. Do all of the persons you have named usually live here? [ Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an *’X"* from 1—C2 and enter reason.) Ask for all persons beginning with column 2: What is — — relationshi to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD A1 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. 1a. What are the names of all persons living or staying here? Start with the name of the or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name "5 b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes, enter 20 M names in 2. [Relationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies orsmallchildren? ................c.0vviiirininiiinnnnnnnan 0 £l 3. |Qate of birth IDate Year — any lodgers, boarders, or persons you employ who live here? . . ....... FE 0 0 ; ' — anyone who USUALLY Nyas here but is now away from home 0 0 HOSP. | WORK RD |2-WK.DV traveling or ina hospital? .........ccos cc ivrvnrsnrnnnvnesprEEEc eves ss — anyone else staying here? . . ..............ueeeeennnneaennnnanennnnan iY 0 (C1 poliNone 10wa [100 ves 00[J None 20wb Number Number 2] No LA IRA IDV [INJ.TCLLTRI HSICOND. | | i | i 1 1 i adumnk LA ~ TRA IDV TINJ.TCLLTRI HSTCON IINJ. 1 CLLTRI HSICOND. ! 1 | I I I | | | | | i A3 Refer to ages of all related HH members. A3 CO an persons 65 and over (5) OJ other (4) B. LIMITATION OF ACTIVITIES PAGE Refer to age. B1 B1 1018-691) 2 [J Other vp) 1. What was — — doing MOST OF THE PAST 12 MONTHS; working at a job or business, 1. 1+ O working (2) keeping house, going to school, or something else? 2 _J Keeping house (3) Priority if 2 or more activities reported: (1) Spent the most time doing; (2) Considers the most important. 3 [J Going to school (5) «J Something else (5) 2a. Does any impairment or heaith problem NOW keep — — from working at a job or business? 2a. | ,[Jves (7 OO No b. Is — — limited in the kind OR amount of work — — can do b of any impai or health problem? b. [| 200ves 7 3D No 6) 3a. Does any impairment or health problem NOW keep — — from doing any housework at all? 3a aD ves a "Ono b is == anhad in the kind OR amount of housework — — can do because of any impairment b. [ves ray eno 15 4a. What (other) condition causes this? Ask if injury or operation: When did [the (injury) occur?/— — have the operation?] 4a. (Enter condition in C2, THEN 4b) Ask if operation over 3 months ago: For what condition did — — have the operation? If pregnancy delivery or O— 3 months injury or operation — . 1 [J oid age (Mark “Old age” box, Reask question 3 where limitation reported, saying: Except for — — (condition), . . .? THEN 4c) OR reask 4b/c. oo ETT... b. Besides (condition) is there any other condition that causes this limitation? b. [J Yes (Reask 4a and b) ONo (4a) c. Is this limitation caused by any (other) specific condition? c. [J Yes (Reask 4a and b) No . Ge BEE we RT 4 0 St a, 2a i ft Ni Gg Agi) i ET Mark box if only one condition. d. Cony 1 condition d. Which of these conditions would you say is the MAIN cause of this limitation? Main cause 5a. Does any impairment or heaith problem keep — — from working at a job or business? Sa 1 Oves (7) One b. Is — — limited in the kind OR amount of work — — could do because of any impairment or health problem? b. 2 Oves (7) I 30 No oo B2 Refer to questions 3a and 3b. 1 [J "Yes" in 3a or 3b (NP) 2 [5 other (6) d. Which of these conditions would you say is the MAIN cause of this limitation? 6a. ls — — limited in ANY WAY in any of ani irment or health problem? 6a. | 1 [ves 20 No inp) b. In what way is — — limited? Record limitation, not condition. b. Limitation 7a. What {other) condition causes this? 7a Ask if injury or operation When did [the (injury) occur?/—- — have the operation?) : (Enter condition in C2. THEN 7b) Ask if operation over 3 months ago. For what condition did — — have the operation? 1 (Jotd age (Mark "Old age’ box, If pregnancy delivery or O— 3 months injury or operation — THEN 7c) Reask question 2. 5. or 6 where limitation reported, saying: Except for — — (condition), . . .? OR reask 7b.c. b. Besides (condition! is there any other condition that causes this limitation? b. [J Yes (Reask 7a and b) Ono (701 c. Is this limitation caused by any (other) specific condition? Sc. [ves (Reask 7a and bl | No Mark box if only one condition. d. Conny veonsiven Main cause FORM HIS. 1 (Evaluation) (2 1-90) Page 4 Ooid age Cod age [J old age [Joid age 1. | First name Mid. init. Age First name Mid. init. JAge 1. [First name Mid. init. §Age First name Mid. init. FAge Last name ex Last name ex Last name ex Last name ex 10m 10m 10m 10m 20F 200F 200F 2[0F 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth | Month ! Date Year Month ' Date | Year Month |Date [vear Month Date | Year 1 1 1 | 1 1 1 HOSP. | WORK | RD _|2-WK.DV | HOSP. | WORK | RD |2-WK.DV HOSP. | WORK | RD [2-WK.DV| HOSP. | WORK] RD | 2-WK.DV c1 oo None 10Owa 10] Yes 00] None [00] None 10wa [100 ves 00] None c1 oo] None 10Owa [10 Yes oo [J None [oo (J None 10Owa [10 Yes ool] None Number 200ws 20] No Number Number 2U0we no Number 20ws [2[0No Number | Number 200we |200No Number "LA — TRA —|DV INJT [CLTTR|HS TCOND.|TA ~ [RA [DV [INJ.” | CLLTRJHS [COND. CA™ T|RA_ DV |INJ. |CLLTRIHS [COND.JUA ~ TRA ~| DV INJ. |CLLTR|HS | COND. | Io | [I | Io | rol | io [I I od I Io 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 [LA ~ TRA —|BV INIT [CLTTAHS TCONG.|TA™ — jRA™ OV JING.” TLITAHS [COND CA~ —|RA™ TDV JINJ. CLITR{HS [CONDJUA ~ TRA ~| DV] INJ TCCLTR AS | COND. | Po Io I Io I Io | Io i | vd | Io 1 1 1 1 1 1 1 i ok 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 MLA ~— TRE —|6V INU. [CLTTRIHS TCONG.|TA~ ~ [RA™ TOV TINJ. | CLOTRTHS > CA— —|RA~ TDV TINT. T|CLLTRIHS [CONDJUA — THA | DV] INJ. TCCLTRAS COND. [I Io I [I | i 4 | | TE I [| | il ' 1 1 J 1 1 1 1 L 1 1 X 1 1 1 1 1 1 1 1 l ! 1 | he ~ TRA ~|BV INST [CLTIAHS COND. |TA~ ~ {RA™ TOV TIND. 7) TLITATHS N CA— ~|RA~ TDV JINJ. TTLLTRTHS [CONO|LA ~ THA —| DV] INJT TCCLTR|HS COND. | Io | [I I [I I [I | [I or | od I od | Ld | | J I | Lo | Ll | 1 LA ~ TRA ~ IDV INJT [CLLTRIHS COND.|TA ~— [RA TDV TINJ.” | CLLTRIHS COND. CA~ “IRA DV [INJ. |CLLTRIHS [CONDJUA ~ TRA “| DV] INJ. TCLLTRIHS | COND. | Fa | [I | {nd | 1 | | | | oo | yo | LJ I [I | | | -_ | [I a " [I | 1 1018-69 (1) 1018-69 (1) 1018-69 (1) B1| 1018-601 2 [J other (NP) 2 [J other (vp) 2 [J other (NP) 2 [J other (vp) 1. 1 [J Working (2) 1. 1 J Working (2) 1. 1 [J Working (2) 1. 1 [J Working (2) 2 a Keeping house (3) 20 Keeping house (3) 20 Keeping house (3) 20 Keeping house (3) 3 [J Going to school (5) 3 [J Going to school (5) 3 [J Going to school (5) 3 [J Going to school (5) 4 [J something else (5) 4 [J something else (5) 4 [J Something else (5) 4 [J something else (5) 2a. 1Oves (7) One 2a. 10 ves (7) Ono 2a.| [ves (7 Ono 2a. | [ves Ono b. 20 Yes (7) 20 ves (7) 30 No 8) b. 20 ves (7) 3d No (8) b. | 2[00ves (7 30 no (6) 3a. 40 ves (4) 4 ves (4) One 3a. 400ves (a One 3a. | aves (a One mn fe ee mess iii to Sr. HERA A i et ii oe emo i eV ms. roms. mont tonto b. 5 Oves 4) 6 [J No (5) b. 5 Oves (4) 6 J No 15) b. 5 Oves (4) 6 (No (5) b. | 5 ves 4) 6 [J No (5) 4a. (Enter condition in C2, THEN 4b) 4a. (Enter condition in C2, THEN 4b) 4a. (Enter condition in C2, THEN 4b) 4a. (Enter condition in C2, THEN 4b) 1 [J oid age (Mark “Old age” box, 1 [J oid age (Mark “*Oid age’ box, 1 [J oid age (Mark “Old age” box, 1 [J oid age (Mark “Old age’ box, THEN 4c) THEN 4c) THEN 4c) THEN 4c) b. [1 Yes (Reask 4a and b) b. [J Yes (Reask 4a and b) b. [J Yes (Reask 4a and b) b. [J Yes (Reask 4a and b) O No (4a) Ono (4a) OI No (4d) O No (4a) fost rages sm etic os ns mim fe mm meee ses sete ee mt time tian compet rm ot pte mg ems Sm tt sc Cc. [ves (Reask 4a and b) Cc. [J Yes (Reask 4a and b) Cc. [J Yes (Reask 4a and b) Cc. [Yes (Reask 4a and b) Ono One Ono Ono d. only 1 condition d. only 1 condition d. only 1 condition d. only 1 condition Main cause Main cause Main cause Main cause 5a. 1 Oves (2) Ono 5a. 1 Oves (7) Ono 5a. 1 Oves (7) One 5a. | 1 [ves 7 Ono oi om rere oe ea See or Sef fe nt i a Chall EVIL path tl REE pe Hid Py ef CRE tg. STRRE b. 2 Oves (7) 30 No b. 2 (ves (7) 30 No b. 2 Oves (7) 30 No b. | 2 [ves (7 30 no B2 1 [J “Yes'* in 3a or 3b (NP) B2 1 [J “Yes in 3a or 3b (NP) B2 1 [J “Yes in 3a or 3b (NP) B2 1 [J "Yes" in 3a or 3b (NP) 2 [J other (6) 2 [J other (6) 2 [J Other (6) 2 [J other (6) 6a. 1 Oves 200 No vp) | Ba. 1 Oves 200 No vp) | Ba. 1 Oves 200No vp) | Ba. | 1 ves 2 No np) Ce. | TTT Ce. | TTT EN Bw. TTT TTT Limitation Limitation Limitation Limitation 7a. (Enter condition in C2, THEN 7b) | 74. (Enter condition in C2, THEN 7b) | 78. (Enter condition in C2, THEN 7b) | 78. | (Enter condition in C2, THEN 7b) 1 CJ oid age (Mark “Old age” box, 1 Dow age (Mark “Old age"’ box, 1 oid age (Mark “Oid age”” box, 1 Cold age (Mark “Old age” box, THEN 7c) THEN 7c) THEN 7c) THEN 7c) b. [J Yes (Reask 7a and b) b. [Yes (Reask 7a and b) b. [J Yes (Reask 7a and b) b. [1 Yes (Reask 7a and b) Ono (7a) Ono (7a) OI No (7a) I ONo (7a) LL [J Yes (Reask 7a and b) 8. [J Yes (Reask 7a and b) es [J Yes (Reask 7a and b) 0, [J Yes (Reask 7a and b) Ono No Ll ONeo No d. Conly 1 condition d. Oonly 1 condition d. Oonly 1 condition d. CJonly 1 condition Main cause Main cause Main cause Main cause FORM HIS-1 (Evaluation) (2-1-0) Page 5 59 60 0 Old age A. HOUSEHOLD COMPOSITION PAGE 1 [J No (APPLY HOUSEHOLD MEMBERSHIP 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | If "Yes," enter id y names in columns Relationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON ot — any babies or small children? . ..............c.c.0iiernnnneennnnnneennnn 0 0 Date cibinh, ie IVear — any lodgers, boarders, or persons you employ who livehere? ................ Od Od ; : — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK.DV travelingorinahospital? . ..............cc0viinnnnenenennnennn . O Od EEN OE —anyoneelsestayinghere? ...............cuuitiininnrinnrnnrnnennnnan a 0 C1 one 0 wa [100 Yes one Number 20wb 200 No Number d. Do all of the persons you have named usually live here? [J Yes (2) prin RULES. Delete nonhousehold members Probe if necessary: by an *’X’* from 1—C2 and enter reason.) Does — — usually live somewhere else? C2 Ask for all persons beginning with column 2: 2. Whatis —— rel. hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD A1 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. B. LIMITATION OF ACTIVITIES PAGE, Continued Refer to age. B3 LA JRA “IDV TINJ.TCLLTRI HSTCOND | 11 1 oe 1 l L L LA ~ TRA IDV TINJ.TCLLTRI HSTCOND | | i 1 i | 1 1 4 LA" ~ TRA IDV TINJ.TCLLTRI HSTCOND | | | Io 1 d 1 1 1 & [LA ~ TRA IDV 1INJ.TCLLTRI HSICON B3 DV |INJ. I CLLTRI HSICOND. ! | i 3 oJ under 5 (10) 2 [J 18-69 (NP) d. Is — — limited in school attendance because of — — health? 108-1711 3[J70and over (8) 8. What was — — doing MOST OF THE PAST 12 MONTHS; working at a job or business, keeping 8. 1 [J Working house, going to schooi, or something else? 2 CJ keeping house Priority if 2 or more activities reported: (1) Spent the most time doing; (2) Considers the most important. 3 J Going to school 4 0 Something else 9a. B of any img or health probl does — — need the help of other persons with 9a. 0 0 == personal care needs, such as eating, bathing, dressing, or getting around this home? re te b. Because of any impairment or health problera, d does — — need the help of other persons in dridling b. — — routine needs, such as everyday h , doing yb 20 Yes (13) 300No (12) getting around for other purposes? 10a. Is — — able to take part AT ALL in the usual kinds of play activities done by most children — — age? |10a. Cves one 11) b. Is —— limited in the kind OR amount of play activitiss — — can do because of any impairment | b.| — | or health problem? 1 Oves (13) 20No (12) 11a. Does any impairment or health problem NON keep — — from attending school? 11a. 1 Oves (13) One Ask if injury or operation: When did [the (injury) occur?/— — have the operation?) Ask if operation over 3 months ago: For what condition did — — have the operation? If pregnancy/delivery or 0— 3 months injury or operation — Reask question where limitation reported, saying: Except for —— (condition), . . .? OR reask 13b/c. Mark box if only one condition. d. Which of these conditions would you say is the MAIN cause of this limitation? 4 (Over sONo 12a. 1s — — limited in ANY WAY in any activities because of an impairment or health problem? 12a. + [JVes 2 ING INP) b.Inwhatwayis —— limited? ~~ Record limitation, not condition. | eT] Limitation 13a. What (other) condition causes this? 13a. (Enter condition in C2, THEN 13b) 1 Dowd age (Mark ‘Old age’’ box, THEN 13c) b. [Yes (Reask 13a and b) ONo (130) [Jes (Reask 13a and b) Ono Oonly 1 condition Main cause FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 6 | | | | | | 1 1 1 1 1 1 | | | | | 1 1 1 1. | | | | | | 1 1 1 1 I 1 oud age Cog age Cord age Cow age 4. | First name Mid. init. Age First name Mid. init. fAge 1. |First name Mid. init. Age First name Mid. init. JAge Last name ex Last name ex Last name Last name ex 10m 10m 10m 200k 200F 20F 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth Month Date Year Month Date | Year Month |Date Year Month |Date | Year 1 1 | 1 1 | | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD [2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00 None 1Owal100 Yes 00] None |oo [J None 10wa 0 es 00] None c1 00] None 10Owa [100 ves loo [J None [oo [J None 1Owa [10 Yes 00] None ~omger | 20wWb|20No | foes | Romper | 20We RON | Romper “omer |20Wb [20No0 | amber | Nomper [20We [20 No | Romper C2 [LA ~— TRA ~|DV [INJ. [CLLTR|HS |COND.|LA ~ [RA [DV [INJ.” | CLLTR|HS [COND. CA™ T|RA DV |INJ. |CLLTRIHS [COND.JUA™ TRA ~| DV] INJ. [CLLTR|HS | COND. I I I I | I | | | | | | | | I | | | | | I | | 1 11 1 Led 1 {— 1 1 | 1 CA 1 1 1 1 ] | 1 do [LA ~ TRA —|BV INU [CLTTR|HS COND. TIND. | TLLTATHS [COND. CA~ —|RA~ DV JINJ. |CLITATHS [CONG|LUA ~ TRA —| DV NJ. TCCLTR HS | COND | | | | | | | | | | | | | | COND. | | | | | | | | | | | | LA" ~ TRA ~|DV |INJ. [CLTTRIHS T[COND.|TA ~ (RA [DV [INJ.” | CLLTRIHS [COND. CA™ 7 |RA [DV [INJ. "|CLLTRIHS [COND.JLA — TRA ~| DV] INJ. [CLLTR|HS | COND | ro | yi | 1 | Io | i | [I | [I | | 1 1 1 1 1d. —l | 1 1 1 o 1 L | 1 1 1 1 1 Ll 1 1 | he ~ TRA T|DV INJT [CLTTRIHS COND. [TA™ ~ [RA [DV TINJ.” | CLITRIHS ~~ CA™ 7|RA™ DV [INJ. "|CLLTRIHS [COND|UA ~ TRA ~| DV] INJ. TCLLTR|HS | COND | i | i | 1 | i | I | | ! | [I | i | Ed L f= | le 1 Lo} 1 Ll 1 be 1 1 1 fll IDV INU. [CULTRIHS COND. |[TA™ ~ [RA™ [DV TINJ.” | CLLTRIHS COND. CA™ IRA DV TINJ. T|CLLTRIHS [COND.JUA ~ TRA ~| DV] INJ. TCLLTRIHS COND | | | | | B3 o Ounders (10) 2[J 18-69 (vp) od unders (100 2] 18-69 (NP) B3 od unders (100 2018-69 (NP) oJ unders (10) 2018-69 (vp) 10s-17¢1 3070and 10s-1701n 3070and 10s-1701 3070and 10s-17¢1 3070and over (8) over (8) over (8) over (8) 8. 1 Od Working 1 O Working 8. 1 O Working 1 ad Working 20 Keeping house 20 Keeping house 20 Keeping house 20 Keeping house sO Going to school 30 Going to school 30 Going to school 30 Going to school +0 Something else +0 Something else «0 Something else «OJ Something else 93 | Mves 13 Ono 10 Yes (13) ONo 92. | ves 13) Ono 10 ves (13) Ono b. b. 2 Oves (13) 30No (12) 20 Yes (13) 30No (12) 20 ves (13) 30No (12) 20 Yes (13) 3l0No (12) 10a 10a. O ves o[dNo (13) Oves 0 ONo (13) [yes o[INo (13) Dyes o[JNo (13) bl Oves (13 20No (12) 10 ves (13) 20No (12) b.| Oves 13 20No (12) 10 ves (13) 2[0No (12) 11a. 11a. 1 O Yes (13) Ono 10 Yes (13) Ono 10 ves (13) Ono 10 Yes (13) Ono by Oves 13) Ono 20 ves (13) Ono bl Oves 13) Ono 20 Yes (13) Ono Ce mean Pe ame Pe i ear = [Te mm Te 3 ves (13) Ono 3 ves (13) Ono 30 ves (13) Ono 30 ves (13) Ono NE WS, SO Pr mm mn mim in fm im te im ne d. 5 4 ves (13) s[JNo 40 ves (13) s[JNo aves (13) 5 JNo a ves (13) s[JNo 12a. 12a. 1 ves 20 No vp) 10 ves 20No (NP) 10 ves 20 No vp) 10 Yes 20 No (vp) | ., 1 Limitation Limitation Limitation Limitation 13a. a Co 13a. a = (Enter condition in C2, THEN 13b) (Enter condition in C2, THEN 13b) (Enter condition in C2, THEN 13b) (Enter condition in C2, THEN 13b) 10 oid age (Mark ‘Old age’’ box, 10 oud age (Mark ‘Old age’’ box, 10 od age (Mark “Old age’’ box, 10 oid age (Mark “’Old age’’ box, THEN 13c) THEN 13c) THEN 13c) THEN 13c) ii ort eo ene EA TN Sd EB RY oe te ret pint mo ei mere mime rcs b. OJ Yes (Reask 13a and b) 0 Yes (Reask 13a and b) b. [J ves (Reask 13a and b) [J ves (Reask 13a and b) Ono | ONo13 [J No (130) [J No (13d) c. [J Yes (Reask 13a and b) [J Yes (Reask 13a and b) c.| [J Yes (Reask 13a and b) [J Yes (Reask 13a and b) a No No No No d. 3 Only 1 condition O Only 1 condition d. 0 Only 1 condition a Only 1 condition Main cause Main cause Main cause Main cause FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 7 61 62 O Old age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes, enter names in columns c. | have listed (read names). Have | missed: Yes No — any babies or smallchildren? . ...................¢c¢c¢ciiiirrinnnnnnnnn — any lodgers, boarders, or persons you employ who livehere? ................ — anyone who USUALLY lives here but is now away from home traveling OF INA NOSPIRAIT « c.oouu i + 5 wise 4 4 3 siwiwins 4 5 329505 » & 3 Slarmine + » siwsine — anyoneelsestayinghere? . ................cc0itiniinnnnnrnnnnanannns goo 00 LE 0 d. Do all of the persons you have named usually live here? J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP Probe if necessary: RULES. Delete nonhousehold members Does — — usually live somewhere else? First name Mid. init. JAge Last name Sex 10m 2[ JF Relationship REFERENCE PERSON Date of birth Month ; Date Year 1 by an *’X"" from 1—C2 and enter reason.) Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD A1 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. 1 HOSP. WORK RD 2-WK. DV oo[J None Owe 1 [Tves 00] None 20wb [200 No Number Number LA JRA “IDV TINJ.TCLLTR| HSTCOND. | | Lo | 1 1 1 l l 1 ILA" = TRA "IDV TINJ.TCLLTRI HSTCOND. | | Ha | 1 1 l 1 1 1 ILA" = TRA TIDV TINJ.TCCLTRI HSTCOND. | | | | | 1 1 1 ! x [LA ~ IRA IDV 1INJ.TCLLTRI HSICOND. | | [I 1 1 1 1 1 1 1 LA IRA IDV [INJ.| CLLTRI HSICOND. | | Io | | | | pl B. LIMITATION OF ACTIVITIES PAGE, Continued B4 | Refer to age. B4| oO unders me 2 Oeo—s9 (14) 3[J70and over (NP) 10 5-59 (85) B 5 Refer to “’Old age’’ and LA’ boxes. Mark first appropriate box. BS [J 01d age’ box marked (14) [J Entry in "LA" box (14) Ask if injury or operation: When did [the (injury) occur?/— — have the operation?] Ask if operation over 3 months ago: For what condition did — — have the operation? If pregnancy/delivery or 0— 3 months injury or operation — Reask question 14 where limitation reported, saying: Except for — — (condition), . . .? OR reask 15b/c. b. Besides (condition) is there any other condition that causes this limitation? c. Is this limitation caused by any (other) specific condition? Mark box if only one condition. d. Which of these conditions would you say is the MAIN cause of this limitation? [J other (NP) 14a. Because of any impairment or health problem, does — — need the help of other persons with 14a. — — personal care needs, such as eating, bathing, dressing, or getting around this home? 10 Yes (15) Ono If under 18, skip to next person; otherwise ask: . | oo ) oo oo h b. CTT TTT b. Because of any impairment or health problem, does — — need the help of other persons in handling 0 O — — routine needs, such as everyday household chores, doing necessary business, shopping, or 2l tyes 3 LINo VP) getting around for other purposes? 15a. What (other) condition causes this? 15a. (Enter condition in C2, THEN 15b) 10 ow age (Mark ‘Old age’ box, THEN 15c) [J Yes (Reask 15a and b) Ono (150) [J Yes (Reask 15a and b) Ono . [J only 1 condition Main cause FOOTNOTES FORM HIS-1 (Evaluation) (21-90) Page 8 Cod age Cloud age Cloud age oud age 1. | First name Mid. init. FAge First name Mid. init. fAge 1. [First name Mid. init. fAge First name Mid. init. fAge Last name ex Last name ex Last name ex Last name eX 10m 10m 10m 10m 20F 20F 2[0F 20°F 2. | Relationship Relationship 2. [Relationship |Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth Month | Date Year Month Date | Year Month Date Year Month Date | Year 1 1 1 1 1 1 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV | HOSP. WORK RD 2-WK. DV c1 00] None 10Owal10 Yes 00] None Joo] None 10wa 0 Yes 00] None c1 00] None 10wa [100 Yes loo None foo] None 10Owa [10 Yes 00] None Number 20wo 2] No Number Number 20we Uno Number Number 20we 200No Number | Number [No Number c2 MA JRA ~ IDV |INJ. [CLLTR|HS T|COND. |LA JRA T|DV [INJ.” | CLLTRIHS [COND. CA~ “JRA DV |INJ. |CLLTRIHS |COND.JUA ~ TRA ~| DV | INJ. [CLLTR/HS | COND. | [| | Io | el | 1 | [I | [I | Io | [I 1 Ll 1 Lo 1 Lod 1 | 1 Ll 1 ro | feud | dorm [LA — TRA — BV INJ. [CLLTR|HS CONG. |TA~ ~ {RA OV JINJ. | CLLTR{HS [COND [A~ T|RA_ DV [INJ. |CLLTRTHS [CONDJLA ~ TRA —| DV INJT TCCLTR HS | COND. | [I | Li | HL | Io | [I | ol | Io | Io 1 Ly 1 Lon 1 Lk ccs fe) | Ld I I) | _ 1 feel MLA — TRA — DV jINJ. [CLTTRIHS TCOND.|TA~ — [RA [OV TINJ.~ AoE pa CA~ ~|RA~ DV [INJ. |CLLTRTHS [CONDJUA — TRA —| BV] INJ. JCLLTR|HS J COND. | Io | [I | Io | [| | i | ol | Io 1 1 1 1 | 1 1 l l 1 l 1 1 l | l l LL 1 1 1 hz “TRA TOV INJ. [CLTTRIHS COND. [TA ~ [RA [DV 7INJ.” | CLLTRIHS COND. CA™ 7 |RA™ DV INJ. "|CLLTRHS [COND.JLA~ TRA —| DV] INJ. TCCLTR|HS |] COND | I | [I | i | 1 | yd | It | [I | Io 1 L { 1 Lo | | Lo 1 IL | 1 {J 1 IL 1 ) CA” JRA” DV JINJ. "|CLLTRIHS |COND.JLA ~~ TRA ~ | DV INJ. TCLLTR|HS | COND. | I | | I | | | | | | | | | | | | 4 | | | | | | | | | 1 1 | | 1 1 1 | | | | | | | | 1 Lond 1 1 | LA ~ TRA ~|DV INJ [CLLTRIHS T[COND.|TA ~ [RA [DV [INJ.” | CLLTRIHS COND. CA” T|RA™ DV [INJ. TICLLTRIHS [CONDJUA ~ TRA ~ | DV] INJ. [CLLTR|HS | COND. | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | DR VISIT 2 DR VISIT 3 DR VISIT 4 PERSON NUMBER PERSON NUMBER PERSON NUMBER ________ F1 LJ under 14 (1b) F1 [J under 14 (16) F1 [J under 14 (1b) O 14 and over (1a) 0 14 and over (1a) 0 14 and over (1a) 1a, o 77770] Last week a o 77770] Last week 1a, oR 77770] Last week b. Month Date 8888] Week before b. Month Date 8888] Week before , Month Date 8888] Week before Cc. 1 [J Yes (Reask 1a orb and c) Cc. 1 [J Yes (Reask 1a or band c) c. 1 Yes (Reask 1a orbandc) I 2 [1] No (Ask 2—6 for each visit) 2 [J No (Ask 2—6 for each visit) 2] No (Ask 2—6 for each visit) 2a. 1 [J GHA Med. Center (5) 3 C] Phone call to GHA (b) 2a. 1 [CJ GHA Med. Center (b) 3 J Phone call to GHA (b) 2a. 1 [J GHA Med. Center (b) 3 [CJ Phone call to GHA (b) 2] Somewhere else (c) 4 [] Phone call some- 2] Somewhere else (c) 4 [J Phone call some- 2] somewhere else (c) 4] Phone call some- where else (c) where else (c) where else (c) b. (3) Cc. a. | Notinnosphar Hospital: o Not in hospital: Hospital: a hosp : “| 020 Home 08 [J 0.p. clinic “| 02[d Home 08 [J 0.p. clinic 2 08[] 0.p. clinic 03[] Doctor's office os [] Emergency room 03] Doctor's office os [] Emergency room 03 [J Doctor's office 09L_J Emergency room 04] Co. or Ind. clinic 10 [J Doctor's office 04[_| Co. or Ind. clinic 10 [J Doctor's office 04 [] Co. or Ind. clinic 10L_J Doctor's office 05] Other clinic 110 ab 05] Other clinic 110 ab 05 [] Other clinic 10 Lab orl] Lab 1200 Overnight ii esl] Lab 12 [] Overnight patient 06 = Lab 120 Overight patient 07] Other (Specify) — (Next doctor visit) 07[_] Other (Specify) — (Next doctor visit) 07 [_] Other (Specify) — (Next doctor visit) Pe 88] Other (Specify) Pei) sa] Other (Specity) : PAVE 88] Other (Speci) 3a.| [yess sokitMD. Gol 3a. 10 ves3) sa [Jokif M.D. (30) 3a.| [Ovesps sCokitmo. 30 land land and b. 200No 3) 9 [CJoK who was seen (4) b. | 200Nowse o[Jokwhowssseent b. | 2[0Noe 9 [JK whowasseen (4) "Oop 2s 4a. 1 [J Condition (item C2, THEN 4g) 4a. 1 OJ Condition (item C2, THEN 4g) 4a. 1 [J Condition (item C2, THEN 4g) and 2 LJ Pregnancy (4e) and 200 Pregnancy (de) and 200 Pregnancy (de) b. 3] Testis) or examination (4c) b. 3 LJ Test(s) or examination (4c) b. 3 LL] Test(s) or examination (4c) || s0Jotherispecity ~ (4g) | | _slJoterrspecity 4g) | | 8] other (specify) - (4g) le | Oven Ono 1S Owe Cle 1d | Ovesian UNowwg 1d. | Clvesn Chong (e-] Cle Novag [of Des ONowg f. (Item C2, i (Item C2, THEN 4g) THEN 4g) 19-1 Llves | Chow 9-1 lve Oho 9. Der . Pr 4 Pr h. sgosncy fel (Item C2, bh CL] Pregnancy (se (Item C2, bh. LJ] pregnancy re tem C2, THEN 4g) THEN 4g) THEN 4g) 6a.| ol ]Telephonein2 Next 1 [ves 2[JNos 5a. oJ Telephone in2 (Next 1 (ves 2 [No (6) Sa.| ol[]Telephonein2 Next 1 [ves 2 [No (6) Dr. visit) Dr. visit) Dr. visit) [or on. Sermon, moore ees FETS ETE RAT FRA Th ims se AE Be +5 A] Tet FT [yale Sem Sate ain, FT Ferre Sofi ies b. mn . m b. mn Lo B mmrr——— Ll _ 4° — | _{ *— Cc. [] Yes (Reask 5b and c) [no Cc. [] Yes (Reask 5b and c) CIno C. [J Yes (Reask 5b and c) Jno 6. City/County — [~~~ 6. City/County [~~ 6. City/County — J ~~ State/ZPCode — [~~~ State/ZPCode — [~~ State/ZPCode — [~~ FORM HIS-1 (Evaluation) (2-1-90) Page 19 [J oid age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. b. What are the names of all other persons living or staying here? Enter names in columns. | If “Yes, "" enter names in columns c. | have listed (read names). Have | missed: Yes No — any babies or small children? . . .............cccuvunennnnennenennnnnnn 0 Od — any lodgers, boarders, or persons you employ who livehere? ................ O O — anyone who USUALLY lives here but is now away from home UAVOIING Or INANOBPIAIT . ov vv isevrsessvrnsvssssmbassssamansesneis Od 0 — anyoneelsestayinghere? ...............c.0iiiiiiiiiriiiatnaarannan O 0 [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) d. Do all of the persons you have named usually live here? Probe if necessary: Does — — usually live somewhere else? 1. |First name Mid. init. JAge Last name Sex LM 2[1F Relationship REFERENCE PERSON « [Date of birth Month Date Year 1 HOSP. WORK RD 2-WK. DV On C1 loo[J None 1Cwa i ves 00 one Number 20 ws 200 No Number Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD NS 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. w—— l 1 1 LA ~ TRA “IDV TiNJ.TCLLTRI HSTCOND. | | Io Io | 1 Ll i ILA" ~ IRA “IDV TINJ.TCLLTRI HSTCOND. | | | i 1 1 1 on 1 1 [LA ~ TRA IDV 1INJ.TCLLTRI HSICOND. | | I id | | deel LA TRA IDV TINJ.TCLLTRI HSICON DV |INJ. | CLLTRI HSICOND. ! | | G. HEALTH INDICATOR PAGE 1a. During the 2-week period outlined in red on that calendar, has anyone in the family had an injury from an accident or other cause that you have not yet told me about? Oves ONo (2) b. Who was this? Mark “‘Injury’’ box in person’s column. c. What was —— injury? Enter injury (ies) in person’s column. d. Did anyone have any other injuries during that period? [Yes (Reask 1b, c, and d) ONo Ask for each injury in 1c: O Yes (Enter injury in C2, THEN e. e. As a result of the (injury in 1c) did [— —/anyone] see or talk to a medical doctor or assistant Te for next injury) (about — —) or did — — cut down on — — usual activities for more than half of a day? OI No (1e for next injury) 2. During the past 12 months, {that is, since (12-month date) a year ago} ABOUT how many days did 2. 000 None illness or injury keep — — in bed more than half of the day? (Include days while an overnight patient in a hospital.) No. of days 3a. During the past 12 months, ABOUT how many times did [— —/anyone] see or talk to a medical 3a. | 00 None (3b) doctor or assistant (about — —)? (Do not count doctors seen while an overnight patient in a cool Joni when oversight hospital.) (Include the (number in 2-WK DV box) visit(s) you already told me about.) patient in hospital _ No. of visits b. About how long has it been since [— —/anyone] last saw or talked to a dical d or i b. 1 interview week (Reask 3b) (about — —)? Include doctors seen while a patient in a hospital. 20 Less than 1 yr. (Reask 3a) 3 O 1 yr., less than 2 yrs. 4 Od 2 yrs., less than 5 yrs. 5 a 5 yrs. or more o[INever i od 4. Would you say — — health in general is excellent, very good, good, fair, or poor? 4. 1 excotent 4 JFair 2 Overy good sO Poor 3[JGood Mark box if under 18. 5a. [J under 18 (NP) 5a. About how tall is — — without shoes? Feet Inches b. About how much does — — weigh without shoes? b. Pounds FOOTNOTES FORM HIS-1 (Evaluation) (21-90) Page 20 73 74 oid age oid age Joid age old age 1. | First name Mid. init. re First name Mid. init. fAge 1. [First name Mid. init. fAge First name Mid. init. fAge Last name ex Last name ex Last name Last name S6X 1Om 10m 10m 200F 200F 200F 2. | Relationship Relationship 2. |Relationship |Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth | Month |Date Year Month Date | Year Month |Date IVear Month Date | Year 1 1 I 1 1 A 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| ‘HOSP. WORK RD 2-WK. DV N c1 old None, ia |i] ves [000] None joo CI None |, | ves 00 None c1 00lINone | ua ly [ves POE] None Joo LI None |, va |] ves [000] None Number 200wo|2lINo Number | Number 200we pONo Number Number 2Uwe Number | Number 2lINo Number | | | | | 1 1 1 1 1 I | | | I | 1 1 1 1 J IRA |DV |INJ. | | l 1 1 |CLLTRIHS | COND. 1 1 | | | | 1 1 1 L [LA ~ TRA —|OV INT [CLTTR]HS CONG. I I I | I | 1 1 1 1 1 1 TA — [RA OV JNJ.” CLTTAHS COND. | | | I | | 1 I 1 1 1 1 CA~ ~|RA~ TDV JINJ. CLLTRTHS [COND] | | | 1 1 1 A ~ TRA | BV] INJT TCCLTR| AS | COND) | | | | | | 1 1 1 0 1 1 [LA ~ TRA ~|OV INJT [CLTTRIHS IY TA 7 [RA TOV TINJ.T | CLOTAIHS i 1 1 TIRAT TDV [INJ. T|CLLTRTHS [COND | | | | | | 1 1 | 1 A™ A ~ TRA | DV INJ. TCLLTR|HS 7] COND. | | | | | | | | | | | | | | | | | | | 1 Ll | 1 L | A | I \ | | 1 | Lol [LA — TRA ~ OV INU JCLUTTA|HS CONG. |TA~ ~ (RA™ TOV TIND. | TLITATHS N CA™ TRA TDV TINJ. T|TLLTRATHS [CONOY|LA ~ THA ~| DV INJ. TCCLTR| AS 7 COND. | I | | | | | | | | | | | | | | | ! | | | | | | 1 1 | 1 1 | | 1 1 1 1 1 1 1 1 1 1 | | ii) 1 1 ] [LA ~ TRA — IDV IINJT [CLLTTRIHS COND. |TA~ ~ [RA [DV [INJ.” | CLLTRIHS [COND. CA™ TIRA TDV TINJ. TICLLTRIHS [COND.JLA ~ TRA ~| DV] INJ. TCLLTRIHS | COND. | | | | I | | | | | | | | I | | | | | | | | | | | | | | | | | | | | | | | Od Injury Od Injury 1b. 0 Injury Od Injury sti sm oa SS oa [fe gi A SS Sg Le Be cs sf i ei Se i i i i SS A iC. Injury Injury i [J Yes (enter injury in C2, THEN [J ves (enter injury in C2, THEN [J ves (enter injury in C2, THEN [J Yes (enter injury in C2, THEN 1e for next injury) 1e for next injury) 1e for next injury) 1e for next injury) 0 No (1e for next injury) Od No (1e for next injury) a No (1e for next injury) Od No (1e for next injury) 1 2. 000] None 000] None 2. 000] None 000] None No. of days No. of days No. of days No. of days 3a. | 4000] None (3b) 000] None (3b) 3a. | 000] None (36) 000] None (3b) ooo] Only when overnight ooo] Only when overnight ooo] Only when overnight ooo] Only when overnight patient in hospital patient in hospital patient in hospital patient in hospital (NP) (NP) (NP) (NP) No. of visits No. of visits No. of visits No. of visits b. 100 Interview week (Reask 3b) 100 Interview week (Reask 3b) 10 Interview week (Reask 3b) 10 Interview week (Reask 3b) 2[ Less than 1 yr. (Reask 3a) 2[J Less than 1 yr. (Reask 3a) 2[ Less than 1 yr. (Reask 3a) 2[ Less than 1 yr. (Reask 3a) 30d 1 yr., less than 2 yrs. 3d 1yr., less than 2 yrs. i 30 1 yr., less than 2 yrs. 3d 1 yr., less than 2 yrs. «02 yrs., less than 5 yrs. a2 yrs., less than 5 yrs. a2 yrs., less than 5 yrs. a2 yrs., less than 5 yrs. s(] 5 yrs. or more s(J 5 yrs. or more s[] 5 yrs. or more s(J 5 yrs. or more ol] Never ol] Never ol] Never ol] Never 4. 4. 100 Excellent 40 Fair 100 Excellent 4] Fair 100 Excellent 4[J Fair 1 Excellent 4 Fair 20 Very good 5 [J Poor 200 Very good 5 [J Poor 20 Very good 5 [J Poor 200 Very good 5 [J Poor 3] Good 3d Good 3[J Good 30] Good 6a. OJ under 18 (vp) [OJ under 18 (vp) 5a. [OJ under 18 (vp) O under 18 (vp) Feet Inches Feet Inches Feet Inches Feet Inches b. ATT TT TTT TTT TT Pounds Pounds Pounds Pounds FOOTNOTES FORM HIS- 1 (Evaluation) (2-1-90) Page 21 O oud age A. HOUSEHOLD COMPOSITION PAGE 1 b. What are the names of all other persons living or staying here? Enter names in columns. 0 . I have listed (read names). Have | missed: — any babies orsmallchildren? ..................cciiiiinnnrrrrnnaanans — any lodgers, boarders, or persons you employ who livehere? . ............... — anyone who USUALLY lives here but is now away from home Avaveling Or IA ROSPIEAIT « ......y co 4ovv inials Some ad wenn sl bias win oe 50 ewe ~ ANYONB BISO SLAYING NOPBZ « vu co viuiv cous win ww sn vas usin & ess sions senses d. Do all of the persons you have named usually live here? [J Yes (2) Probe if necessary: Does — — usually live somewhere else? 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) Ask for all persons beginning with column 2: What is — — relati N hip to (reference person)? First name Mid. init. JAge Last name ty LIM If “Yes, ’" enter 20 names in columns . [Relationship - Yes No REFERENCE PERSON « | Date of birth 0 a Month ' |Date Year O O : : 0 0 HOSP. WORK RD 2-WK. DV O 0 C1 loo] None Iwai ves 00] None Number 200 wo [200 No Number LA IRA "IDV TINJ.TCLLTR| HSICOND. i | 11 1 1 1 1 1 1 3. What is — — date of birth? (Enter date and age and mark sex.) LA IRA IDV [INJ.TCLLTRI HSTCOND. Il 1 1 1 I 1 L REFERENCE PERIODS 2-WEEK PERIOD IY = mites ston mim ess it mi oe io em ie 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. DV [INJ. I CLLTRI HSICOND. ! | {| H. CONDITION LISTS Read to respondent: you have mentioned it before. Now | am going to read you several lists of medical conditions. Tell me if anyone in the family has each condition | read, even if 1a. Does anyone in the family {read names} NOW HAVE — If “’Yes,’’ ask 1b and 1c. b. Who is this? b 1 c. Does anyone else NOW have — 3 c. Enter condition and letter in appropriate person’s column. family have — If ‘Yes,’ ask 3b and 3c. . Who was this? 3a. DURING THE PAST 12 MONTHS, did anyone in the DURING THE PAST 12 MONTHS, did anyone else have — Enter condition and letter in appropriate person’s column. A. PERMANENT E. Any other trouble A. stiffness or any hearing with one or deformity of the both ears? foot, leg, or back? Tari i ite maim B. F. Tinnitus or ringing in B. PERMANENT the ears? stiffness or any C. deformityofthe | | --——---—-—---—--——+ ie] fingers, hand, or D. arm? G. Blindness in one or | __ |] both eyes? C. Any condition caused | Epa a fe E. by an accident or injury which H. Cataracts? F. happened morethan | {| ~~ . three months ago? 1 G 1. Any other trouble seeing with one or Damaged heart valves? Tachycardia or rapid heart? Any other heart trouble? Hemorrhoids or piles? Arthritis or any kind of rheumatism? I. FREQUENT constipation? 2a.Has anyone in the family EVER HAD — If “Yes,” ask 2b and 2c. b. Who was this? c.Has anyone else EVER had — Enter condition and letter in appropriate person’s column. A. Hardening of the arteries or arteriosclerosis? D.Hypertension, sometimes called high blood pressure? B. Congenital heart disease? F. A myocardial infarction? C. Coronary heart disease? G.Any other heart attack? D. Deafness in one or both eyes EVEN H. Dermatitis or any both ears? yo woaiing other skin trouble? 0. Sinus trouble? FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 22 75 76 J oid age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. b. What are the names of all other persons living or staying here? Enter names in columns. | If “Yes,” enter names in columns c. I have listed (read names). Have | missed: Yes No — any babies or smallchildren? ..................cciiiirnnnninnennnnnn 0 O — any lodgers, boarders, or persons you employ who livehere? . ............... Od Od — anyone who USUALLY lives here but is now away from home AVEliNg Or INBIOBPIMAIP «so cs vnr vss rnnms dns ammensirvs eres sssnres O 0 — anyone elsestaying here? . ...............c.vuitiiirinrnenenennnnenns Od Od d. Do all of the persons you have named usually live here? [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP Probe if necessary: RULES. Delete nonhousehold members by an *’X"’ from 1—C2 and enter reason.) Does — — usually live somewhere else? 1. [First name Mid. init. JAge Last name Sex 10m 2[0F Relationship REFERENCE PERSON Date of birth Month |Date Year | HOSP. WORK RD C1 loo[J None 10 wa [10 Yes 00[J None Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. Number 20 ws 20 no el [LA° ~ TRA “IDV TINJ IRN | | i 1 1 1 1 1 1 L ILA" ~ IRA “IDV TINJ.TCLLTRI HSTCOND. | | [| 1 1 J 1 1 1 [LA" = TRA "IDV TINJ.TCLLTRI HSTCOND | | | | IA l 1 1 | 1 | [LA ~ TRA IDV 1INJ.TCLLTRI HSICOND | 1 | 1 1 1 1 1 1 [LA ~ TRA IDV TINJ.TCLLTRI HSTCON DV |INJ. | CLLTRI HSICOND. | | Lol | | I J. HOSPITAL PAGE HOSPITAL STAY 1 ® For initial *’No condition’’ ask: Why did — — enter the hospital? ® For tests, ask: What were the results of the tests? If no results, ask: Why were the tests performed? ® For delivery ask: ® For newborn ask: Was this a normal delivery? Was the baby normal at birth? If “No,” ask: If “No,” ask: What was the matter? What was the matter? 1. Referto C1, "HOSP." box. 1. PERSON NUMBER 2. You said earlier that — — was a patient in the hospital since (1.3-month hospital date) a year Month Date Year ago. On what date did — — enter the hospital ([the last time/the time before that])? Record each entry date in a separate Hospital Stay column. 2. 19 3. How many nights was — — in the hospital? 3. | 0000] None (Next HS) Nights 4. For what condition did — — enter the hospital? 4. 2] Normal at birth 3[J No condition [J condition 2 1 0 Normal delivery’ (5) J 1 Refer to questions 2, 3, and 2-week reference period. [J Atleast one night in 2-week reference period (Enter condition in C2, THEN 5) 0 No nights in 2-week reference period (5) 5a. Did — — have any kind of surgery or operation during this stay in the hospital, 5a. including bone settings and stitches? 100 ves 2[No (6) b. What was the name of the surgery or operation? b. 0 If name of operation not known, describe what was done. 2) 3) . i thisstay? R c dT me oe oo em c. Was there any other surgery or operation during this stay C1. Ves Rossh Ss ort Oe 6. What is the name and address of this hospital? 6. [Name Number and street City or County State FOOTNOTES FORM HIS 1 (Evaluation) (2-1-90) Page 24 Cod age Cloud age Cod age Cloud age 1. | First name Mid. init. FAge First name Mid. init. fAge 1. |First name Mid. init. fAge First name Mid. init. Age Last name Sex Last name ex Last name ex Last name Sex 10m 10m 10m 10m 20JF 200F 2(JF 2(F 2. | Relationship Relationship 2. [Relationship |Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth | Month |Date Year Month Date | Year Month |Date Year Month |Date | Year 1 1 ol | 1 1 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00 None yOwal10 Yes 00] None |oo [J None 10wa hO ves 00 None c1 00 None 10wa [10 ves 0] None [oo [] None 10Owa [100 Yes 00] None Number 2Dwol2lINo Number Number 2Uwo No Number Number 20wb [ono Number | Number [dw |200No Number c2 A ~ JRA DV INJT [CLLTR|HS |COND.|[LA ~ [RA [DV [INJ.” | CLLTRIHS [COND. CA~ T|RA DV |INJ. |CLLTRIHS [COND.JUA~ TRA | DV|iNJ. JCLLTR/HS | COND. | | | | I I | | | | | I I | | I I | | | | | 1 Li 1 Ll 1 Ll 1 IL | Ll 1 | ! 1 Ny 1 Vereen [LA — TRA —|0V INJ [CLTTRHS JCOND.|TA~ — {RA OV TINJ. | CLLTRHS [CONG. CA~ ~|RA~ DV JINJ. CLLTRA{HS [CONDJLA ~ TRA ~| DV NJ. CLLTR|HS | COND | | | | | | | | | | | | | | | | | | | | | I | | 1 I 1 | 1 a | fone) 1 bd | 1 | 1 I | (ES [LA ~ TRA ~|DV |INJ [CLTTR|HS COND.|TA ~ [RA [DV [INJ. | CLLTRIHS COND. CA™ 7|RA~ DV JINJ. T|CLITRTHS [CONDJUA ~ TRA —| DV] INJ. TCCLTR|HS | COND. | | I | | I | | | | | | I I | | I | | I | I | I 1 1 1 fois 1 Lod | J) 1 ll | 1 A 1 Sond 1 Lod LA — TRA ~ BV IN JCLTTRIHS JCOND.|TA~ ~ [RA™ TOV TINJ. | TLITRTHS [COND. CA 7|RA™ DV JINJ. T|CLLTRIHS [COND|JUA ~ TRA ~| DV] INJ. TCCLTR|HS | COND. | | | | | | | | | | | | | | | | | | | | ls | | | | | | 1 1 | | 3 | | | | | | 1 | dk | | 1 | 1 1 LA ~ TRA IDV INJT [CLLTRIHS [COND.|LA ~ [RA [DV [INJ.” | CLLTRIHS [COND. CA™ TIRA DV [INJ. T|CLLTRIHS JCOND.LA~ TRA ~| DV] INJ. jCLLTR|HS | COND. | | I | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | HOSPITAL STAY 2 HOSPITAL STAY 3 HOSPITAL STAY 4 | 1. | PERSON NUMBER 1. |PERSON NUMBER 1. |PERSON NUMBER Month Date Year Month Date Year Month Date Year . 19 2. 19 2. 19 3. | 0000] None (Next HS) 3. | 0000] None (Next HS) 3. | 0000] None (Next HS) Nights Nights Nights 4. 1 [J Normal delivery 4. 1 [J Normal delivery 4. 1 [J Normal delivery 2] Normal at birth (5) 27] Normal at birth (5) 2 [J Normal at birth (5) 3[J No condition 3[J No condition 3[J No condition [J condition ¥ [J condition 3 [J condition ¥ J1 ad At least one night in 2-week reference period (Enter condition in C2, THEN 5) O No nights in 2-week reference period (5) [J At least one night in 2-week reference period (Enter condition in C2, THEN 5) Ono nights in 2-week reference period (5) [J At least one night in 2-week reference period (Enter condition in C2, THEN 5) Od No nights in 2-week reference period (5) 5a. 5a. 5a. 10 Yes 2[0No (6) 10 ves 2[0No (6) 10 ves 2[No (6) b.| im) b.| LY (2) (2) 2) (3) 3) 3) Cc. Cc. Cc. [J ves (Reask 5b and c) Ono [J Yes (Reask 5b and c) Ono [J Yes (Reask 5b and c) Ono 6. |Name 6. [Name 6. |Name Number and street Number and street Number and street City or County State City or County State City or County State FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 25 77 78 O od age . Do all of the persons you have named usually live here? A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | | "Yes," enter i= - names in columns Relati . elationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies orsmallichildren? ...............c0vvennennnnennrnnnennns 0 0 Date ot birth op Wear — any lodgers, boarders, or persons you employ who livehere? . ............... 8] | ; : — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK.DV travelingorina hospital? . ...............c0iteieennennennennnennens O Od boC IN OLIN — anyoneelsestayinghere? ................0tiitinneinrenernnraneanns O Oa C1 oO wa [10 ves ons Number 2Lws 200 No Number Probe if necessary: Does — — usually live somewhere else? [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) Ask for all persons beginning with column 2: What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) A REFERENCE PERIODS 2-WEEK PERIOD 1 13-MONTH HOSPITAL DATE A 2 ASK CONDITION LISTS 1,2, and 3. LA" ~ TRA “IDV TiNJ.TCLLTRI HSTCOND. | | | | 1 J 1 J 1 1 [LA= 7 TRA TIDV TINJ.TCLLTRI HSTCOND. | | Io 1 l 1 1 1 1 [LA TRA IDV 1INJ.TCLLTRI HSICOND. mmm mmm mmm mmm ——— | | | | 1 1 1 1 1 1 [LA ~ TRA IDV TINJ.TCLLTRI HSTCON DV |INJ.|CLLTRI HSICOND. ! | Il | | I CONDITION 1 | PERSON NO. ___ Ask 3g if there is an impairment (refer to Card CP2) or any of the following entries in 3b—f: 1. Name of condition Abscess Damage Palsy Ache (except head or ear) Growth Paralysis Mark *“2-wk. ref. pd.’’ box without asking if ‘DV’ or “HS"’ ( H 9 Pp in C2 as source. Blood clot Infection Sore(ness) 2. When did [— —/anyone] last see or talk to a doctor or assistant Boil Inflammation Stiff(ness) about — — (condition)? Cancer Neuralgia Tumor | 5 [] 2 yrs., less than 5 yrs. Cramps (except menstrual) Neuritis Ulcer > 2 neve me east 2) 6 [J] 5 yrs. or more Cyst Pain Varicose veins 2 [J Over 2 weeks, less than 6 mos. 20 Dr.seen, DK when | Wesiinasa) 3 0 6 mos., less than 1 yr. 8 O DK if Dr. seen } (3b) 40 1yr., less than 2 yrs. 9 [J Dr. never seen . What part of the body is affected? (Specify) 3a. (Earlier you told me about — — (condition)) Did the doctor or assistant Show the following detail: call the (condition) by a more technical or specific name? Ov 2 [IN sox HBB. «+i viv vin sinivvivminin siesine viesls ws vie win sv seis e sms skull, scalp, face BL Yes 2 Back/SpINe/Vertebras . . . . ............eeeeenennnn. upper, middle, lower Ask 3b if “Yes” in 3a, otherwise transcribe condition name from Bide «vss EVER SC AE RARE SEER Eee ee ee eee left or right item 1 without asking: rr i eg TE inner or outer; left, right, or both id h h nit? BYO ivnsissvrsnsrinnsin shams een messiness yee left, right, or both b. What did he or she call i eT Breve? em AAT shoulder, upper, elbow, lower or wrist; left, right, or both 0 b C se (3e) Pp ¥ 0 Ee NE SOE entire hand or fingers only; left, right, or both 1 0 Colgt Blindness NC) id il nl ROU rrr 2 ARTS eT Ten hip, upper, knee, lower, or ankle; left, right, or both 3 Normal pregnancy, . Norma isk sony Y } (5) «0 Other 130) FOO i nuniv ws rmmmanive vimminra ms bine entire foot, arch, or toes only; left, right, or both vasectomy c. What Was the cause of —— (condition in 3bR (Specify) 3 iia Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: Infection Sore Soreness a oT -——-——————=———-——-—-————————- h. What part of the (part of body in 3b—g) is affected by the [infection/ Mark box if accident or injury. o [J Accident/injury (5) sore/soreness] — the skin, muscle, bone, or some other part? d. Did the (condition in 3b) result from an accident or injury? 10 ves (5) 20 No mm mm mm mm (Specify) Ask 3e if the condition name in 3b includes any of the following words: Ask if there are any of the following entries in 3b—f: Ailment Cancer Disease Problem Tumor Cyst Growth Anemi Condi D Asthma Cyst Growth Trouble Is this [tumor/cyst/growth] malignant or benign? Attack Defect Measles Tumor Bad Ulcer 10 Malignant 2 [Benign 9 Dok pyr ri Foi a. When was — — (condition in 3b/3f) 1 [J 2-wk. ref. pd. e. What kind of (condition in 3b) is it? Cae) first noticed? 2 Ll Over desta months Sr te Ss ert SE Sm rte sp rey remand ES LET PEE edi ae 3 3 [J over 3 months to 1 year Ask 3f only if allergy or stroke in 3b—e: b. When did —— (name of injury in 3b)?] | [1 5yer 1 year to 5 years f. How does the [allergy/stroke] NOW affect — —? (Specify) z 5 [J Over 5 years For Stroke, fill remainder of this condition page for the first present effect. Enter in item C2 and complete a separate condition page for each additional present effect. Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) (Was it less than 1 year or more than 1 year ago?) (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 26 oid age CJoid age Cold age [Joid age 4. | First name Mid. init. FAge First name Mid. init. fage 1. [First name Mid. init. fAge First name Mid. init. fAge Last name ex Last name ex Last name ex Last name ex 10m 10m 10m 10m 200F 2[JF 200F 2[JF 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth Month |Date Year Month Date | Year Month |Date Year Month |Date | Year 1 1 1 | ! ! | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV | HOSP. WORK RD 2-WK. DV c1 00] None 1Owal10 ves 00] None Joo] None 10wa hO ves 00] None c1 00] None yOwa [10 Yes loo [J None oo [J None 1Dwa (100 Yes ool] Pore “omper | 20wWb|20No | omer | amper | 20Wb pOINo | omer ~omper | 20Wb 200No | omper | Tamper [2OWe |200No | amber c2 c2 [LA — TRA ~|DV |INJ. [CLLTR|HS [COND.|LA ~ [RA |DV |INJ.” |CLLTRIHS COND. CA™ 7|RA DV |INJ. |CLLTRIHS [COND.JUA™ TRA | DV|INJ. [CLLTR|HS | COND. | | Io | 1 | Io 1 [I | [I | | 1 1 I | | - 1 Leal | Lo 1 Looe 1 Lo 1 | [LA ~ TRA — BV INU [CLTTRjHS COND. |TA~ — [RA [OV [INJ. | CLITRIHS [COND. CA™ ~|RA~ DV [INJ. |CLLTRIHS [COND.JLA ~ THA | DV] NJ. TCCLTR|HS | COND. | [I | [| | [I | id | [I | [a | Io | [I 1 jt 1 et 1 bal. 1 ed 1 (I 1 coal 1 Ll l Lo [LA — TRA —|DV |INJ [CLULTR|HS COND. [TA ~ [RA DV [INJ.” | CLLTRIHS = CA™ T|RA™ DV |INJ. T|CLLTRIHS NL ~ TRA 7 DV] INJ. TCLLTR|HS 7] COND. | [I | Io I [| | L 4 | Io | 1 | [i 1 1 1 1 1 1 1 1 1 1 1 1 l 1 1 1 1 1 1 ~ 1 i 1 Jd. 1 1 » ~ TRA TOV INU JCLUTTRIHS T[COND. |[TA~ ~ [RA [DV TINJ.” | CLLTRIHS N CA™ T|RA™ DV [INJ. T|CLLTRIHS [COND.|UA ~ TRA ~| DV] INJ. TCCLTR{HS 7] COND. 1 a. | ro | [I | i | i | ro | A ' Io 1 Ll 1 [I | Ll 1 I | 1 1 | 1 i 1 Ll [LA = TRA “IDV IINJ. [CLLTRIHS TJCOND.|LA~ ~ [RA [DV [INJ.” | CLLTRIHS COND. LA™ “IRA DV [INJ. TICLLTRIHS [COND.JLA ~~ TRA | DV] INJ. TCLLTR|HS | COND. 1 od 1 | 1 [I | 1 i | [I | i 1 i | | | | | | | [I | 11 1 Vo) | ng 1 1 Refer to RD and C2. + pte K 1 100 "Yes" in *‘RD"* box AND more than 1 condition in C2 (6) 13. Is this (condition in 3b) the result of the same accident you already 0 told me about? 8 Ll Other (K2) 0 Yes (Re d diti by hy 6a. During the 2 weeks outlined in red on that calendar, did — — dich ile hy NC) (condition) cause — — to cut down on the things — — usually does? 0 Page No. Oves ONo x2) No b. During that period, how many days did — — cut down for more than half of the day? 14. Where did the accident happen? 10] At home (inside house) 00 JNone (k2) —— Days 2[0] At home (adjacent premises) 7. During those 2 weeks, how many days did — — stay in bed for 3[] street and highway (includes roadway and public sidewalk) more than half of the day because of this condition? a] Farm 0 s(J Industrial place (includes premises) 00L_INone — Days se] School (includes premises) Ask if “Wa/Wb’’ box marked in C1: 70] Place of recreation and sports, except at school 8. During those 2 weeks, how many days did — — miss more than 8] Other (Specify) ¥ half of the day from — — job or business because of this condition? 00[JNone Days Mark box if under 18. [J Under 18 (16) 5 5a. Was — — under 18 when the accident happened? Ask if age 5— 17: 3 : 100 Yes (16) Ono 9. During those 2 weeks, how many days did — — miss morethan {} — ~~ = —* | half of the day from school because of this condition? b. Was —— in the Armed Forces when the accident happened? Y, oo] None Days 20 (83116) A d No RE a, c. Was —— at work at — — job or business when the accident happened? [J Condition has “CL LTR" in C2 as source (10) 0 J 0 in K 2 . . . 3L] Yes 4L No [J condition does not have “CL LTR’ in C2 as source (K4) y 16a. Was a car, truck, bus, or other motor vehicle involved in the accident 10. About how many days since (12-month date) a year ago, has this inany way? condition kept — — in bed more than half of the day? (Include days Cy SIN iT while an overnight patient in a hospital.) ph SR DILL ms oh ome blir on Ett mein pm b. Was more than one vehicle involved? 000[JNone mmm BY 10 Yes 20 No 11. Was — — ever hospitalized for — — (condition in 3b)? c. Was [it/either one] moving at the time? 10 ves 20 No 10] ves 200No d0 Missing extremity or organ (K4) 17a. At the time of the accident what part of the body was hurt? K3 | dove 112 What kind of injury was it? . - . Anything else? 12a. Does — — still have this condition? . 100 ves (kd) Cine Part(s) of body Kind of injury b. Is this condition completely cured or is it under control? 2cured 8] other (specify) + | (K4 Se Ts ee en Be lit mf eet seis si } 3 under control { , ] : ira iE) Ask if box 3, 4, or 5 marked in Q.5: c. About how long did — — have this condition before it was cured? b. What part of the body is affected now? 0 How is — — (part of body) affected? h To E stl lass thane marlh Gh { 1 5 foe) s Is affected in any other way? oo Number 2-1 Years Part(s) of body * Present effects ** d. Was this condition present at any time during the past 12 months? 10ves 2[0No oN ident/i NC, : - 0 at acct ent ik eh * Enter part of body in same detail as for 3g. K4 1 First accident/injury for this person (14) ew 5 5 cL Other (13) If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-190) Page 27 79 O ow age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. |First name Mid. init. Age one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name °5 b. What are the names of all other persons living or staying here? Enter names in columns. | |f “yes, enter 0 3 names in columns 2. [Relat - . elationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies or smallchildren? . ..............c.c.c0viiiinniinnnennnnnnns 0 0 3. |Date of birth Ipate Iver — any lodgers, boarders, or persons you employ who livehere? ................ 0 0 ! — anyone who USUALLY lives here but is now away from home 0 0 HOSP. | WORK RD [2-WK.DV traveling orin a hospital? «cc... .coovrsssvmrrsssrnnnnssnvnnsssnssnnons 0 N oo IN — anyone else staying here? . . . ............uuennnnerennnnneenennneenns 0 0 IC 1 [relINone |, 14, | 7 ves 20H Nore “Number 20 Wb 20] NO | “rm d. Do all of the persons you have named usually live here? [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP R Probe if necessary: RULES. Delete nonhousehold members C 2 by an “‘X’’ from 1—C2 and enter reason.) | | _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Does — — usually live somewhere else? AAA Loy “UN. Vac) HisTeons 1 1 1 1 1 1 Ask for all persons beginning with column 2: 2. What is — — relationship to (reference person)? [LA — TE —I BV" TING. TCCLTRI HSTCONG. 3. What is — — date of birth? (Enter date and age and mark sex.) . J y y | ! REFERENCE PERIODS ~~ (~~ LA [RA “IDV TINJ. TCCLTRI HSTCOND. | | i Io 2-WEEK PERIOD i L bd bd 12-MONTH DATE [LA ~ TRA IDV 1INJ.TCLLTRI HSICOND. mm mm mr em mm mm em mm mm me ee 1 Ed hd CV iL) 13-MONTH HOSPITAL DATE LA IRA IDV TINJ.TCLLTRI HSTCOND. A2 (Ask CONDITION LISTS 1,2, and 3. bolo CONDITION 2 Ask 3g if there is an impairment (refer to Card CP2) or any of the following entries in 3b—f: 1. Name of condition Abscess Demags Palsy Ache (except head or ear) Growth Paralysis Mark **2-wk. ref. pd.’’ box without asking if ‘DV’ or ’HS"’ Bleed H ge Pp in C2 as source. Blood clot Infection Sore(ness) 2. When did [— —/anyone] last see or talk to a doctor or assistant Boil Inflammation Stiff(ness) about — — (condition)? Cancer Neuralgia Tumor 0 [J Interview week (Reask 2) 5 0 2yrs., less than 5 yrs. Side foxcept menstriai eg Ys oh v3 2.wi. vet. pd. ss yrs. or more ys ain ~ a veins 2 [J] Over 2 weeks, less than 6 mos. 0 Dr.seen.DKwhen =~ | kines) 30s mos., less than 1 yr. 8 [J DK if Dr. seen } (3b) 40 1yr., less than 2 yrs. 9 [J Dr. never seen §( g. What part of the body is affected? Specify 3a. (Earlier you told me about — — (condition)) Did the doctor or assistant Show the following detail: call the (condition) by a more technical or specific name? Hons in OB» ¢v a nbn REE EA Re ee A skull, scalp, face 100 ves 2 0 No ® ul PK Back/spine/vertebrae . .............cc00uinaannan upper, middle, lower Ask 3b if Yes’ in 3a, otherwise transcribe condition name from Billo cnn nnn sn nn so rin wy 4 0 Sa let csight item 1 without asking: Se Rn AR RO A EN EA re NN aE inner or Suber lors right, or both | YB Co eren nr RA AN ARS AER BLARNEY , right, or both b. What did he or she call it? To BI iqvvrvswn meters shoulder, upper, elbow, lower or wrist; left, right, or both HA «vere vivian inion wore enon oe age fi 3 y by VE] Color Blindness iC) 2 i ve) hand hip. Any A a. _ ; ve eee , . y ; loft, 5 aL] Normal ve 2 : Eyes a FOOT «vv vevnanananensn entire foot, arch, or toes only; left, right, or both vasectomy c. What was the cause of —— (condition in 3b (Specify) 3 A Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: Infection Sore Soreness A TT TT Sao oo. ~~ — ~~~) h. What part of the (part of body in 3b—g) is affected by the [infection/ Mark box if accident or injury. o [J Accident/injury (5) sore/soreness] — the skin, muscle, bone, or some other part? d. Did the (condition in 3b) result from an accident or injury? 100 ves (5) 20 No (Specify) Ask 3e if the condition name in 3b includes any of the following words: ] Ask if there are any of the following entries in 3b—f: Ailment Cancer Disease Problem Tumor Cyst Growth Anam c Disord Asthma Cyst Growth Trouble 4. Is this [ /cyst/growth] mali or benign? Attack Defect Measles Tumor Bad Ulcer 1d Malignant 2 [Benign 9 ok Simrad i a. When was — — (condition in 3b/3f, 1 [J 2-wk. ref. pd. e. What kind of (condition in 3b) is it? i —— 5 first noticed? 2 Ll Over ZweckstoB months —_————————————_—_—_————_—_—————————_——_——_——-—-—-—Y | TT TTF Tm mmm mmm ——— 3 [J Over 3 months to 1 year Ask 3f only if allergy or stroke in 3b—e: b. When did — — (name of injury in 3b)?) 4 [J Over 1 year to 5 years f. How does the [allergy/stroke] NOW affect — —? (Specify) g 5 [J Over 5 years Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) For Stroke, fill remeinger of this Sandpion page for Be first present (Was it less than 1 year or more than 1 year ago?) effect. Enter in item and complete a separate condition page for s each additional present effect. (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 28 Cod age Oo age Oo age Cow age 4. | First name Mid. init. FAge First name Mid. init. fAge 1. [First name Mid. init. Age First name Mid. init. fAge Last name ex Last name ex Last name ex Last name ex 10m 10m 10m 10m 200F 20¢ 200 Clr 2. | Relationship Relationship 2. [Relationship Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth \ Month |Date Year Month Date ] Year Month |Date Year Month |Date Year 1 1 als 1 1 1 1 1 HOSP. | WORK | RD [2-WK.DV | HOSP. | WORK | RD [2-WK.DV HOSP. | WORK | RD |2-WK.DV| HOSP. | WORK | RD |2-WK.DV c1 00 None 10Owal10 ves 00] None oo] None 10wa h[O Yes 00 None c1 00 J None 10wa [10 Yes oo [J None [oo] None 10O0wa [10 Yes 00] None Number 20wb 20No Number Number 2Uwe 2CINo Number Number 200wb 2000 Number | Number [we [200 No Number ET —— ii] c2 C2 [LA ~ TRA ~|DV |INJ. [CLLTR|HS |COND.|TA~ ~ [RA |DV [INJ.” | CLLTR[HS COND. CA~ “JRA DV |INJ. |CLLTRIHS [COND.JLA™ TRA ~|DV|INJ. |CLLTR|HS | COND. | 1 | | I. | | | ol | ol | ol | Ei 1 1 | 1 1 | 1 1 1 1 1 H 1 1 1 1 1 1 1 1 1 1 1 4 [LA ~ TRA ~|BV IN [CLTTRHS COND. |TA~ — [RA OV TINJ. | CLITRHS [TOND. [A~ ~|RA DV [INJ. |CLLTRIHS [CONDJUA ~ THA ~| DV] INJ. [CCLTR|HS | COND. | 1 | [I | [I | Io | Io | Io | [I | - 1 1 1 1 1 1 1 1 1 1 | 1 1 1 1 1 1 1 1 | | l 1 1 [LA ~ TRA ~|OV |INJT [CLULTRHS COND. [TA ~— [RA [DV [INJ.” | CLLTRTHS COND. CA™ T|RA DV |INJ. T|CLITRIHS NE TRA 7 DV] INJ. TCCLTR|HS | COND. | Io | [I | [I I Io | Io | Io | Io | Io 1 1 ! 1 Ll. l 1 1 1 1 1 1 1 1 l l k 1 1 1 1 dod, »- ~ TRA T|BVINJT [CLTTRJHS TCOND.|TA™ — [RA TOV TINJ. T| CLITATHS [COND. [CA™ T|RA_ TDV [INJ. T|CLLTRTHS [CONDJLA ~ TRA ~| DV NJ. TCCLTR|HS | COND. | [I | [I | Io | [I | Io | Io | Io | Io | 1 1 ] Ll 1 | | Lo | ed | Lal 1 fel [LA ~ TRA IDV |INJ. [CLLTRIHS COND.|TA ~~ [RA [DV [INJ. | CLLTRIHS COND. LA™ TIRA” DV [INJ. TICLLTRIHS [CONDJUA~ TRA | DV] INJ. [CLLTR|HS | COND. | [I | [I 1 iH | [I I Io | 1 | [I | | | | | [| | Ef | i | [I | -_ | | | fool Refer to RD and C2. g aioe E o IK | Eves" in Ro" box AND more than 1 condition in C2 (6) 13. Is this (condition in 3b) the result of the same accident you already 0 told me about? 8 Ll Other (K2) O 6a. During the 2 weeks outlined in red on that calendar, did — — Yas (eqard ann Page umbel Whete oy NC) ition) cause — — to cut down on the things — — usually does? Page No. Oves Ono k2) Uno b. During that period, how many days did — — cut down for more than half of the day? 14. Where did the accident happen? 10 At home (inside house) 00 None (k2) 5 Days 2[J At home (adjacent premises) 7. During those 2 weeks, how many days did — — stay in bed for 3[] street and highway (includes roadway and public sidewalk) more than half of the day because of this condition? ad Farm 0 s(] Industrial place (includes premises) 00LNene Days sd School (includes premises) Ask if ““Wa/Wb'’ box marked in C1: 70] Place of recreation and sports, except at school 8. During those 2 weeks, how many days did — — miss more than 8[] Other (Specify) ¥ half of the day from — — job or business because of this condition? 00[JNone Days Mark box if under 18. CO under 18 (16) 15a. Was — — under 18 when the accident happened? Ask if age 5—17: 101 ves 1169 oN Pp 9. During those 2 weeks, how many days did — — miss more than mA a, half of the day from school because of this condition? b. Was — — in the Armed Forces when the accident happened? 00 4d None Days al] ves ¢ 16) EE CE I ESE d No ee mm So EE RT 50 c. Was — — at work at — — job or business when the accident happened? K2 [J Condition has “CL LTR" in C2 as source (10) 3dv a es 4 No [J Condition does not have “CL LTR" in C2 as source (K4) 6 = 16a. Was a car, truck r other le invol n th 10. About how many days since (12-month date) a year ago, has this ! in any way? » bus, or other motor vehic olved in the accident condition kept — — in bed more than half of the day? (Include days Ov Clsevz while an overnight patient in a hospital.) Ud XO mica BEANGII om mi rm e em immo pert si b. Was more than one vehicle involved? 000 JNone Days 1 ves 20 No 11. Was — — ever hospitalized for — — (condition in 3b)? c. Was [it/either one] moving at the time? | 100 ves 20 No 100 ves 200No [J Missing extremity or organ (K4) 17a. At the time of the accident what part of the body was hurt? K 3 OJ Other (12) What kind of injury was it? Anything else? 12a. Does — — still have this condition? ¥ihing 10ves (ka) ONo Part(s) of body * Kind of injury b. Is this condition completely cured or is it under control? 20cured 8 [J Other (Specify) 2 30 under control (k4) a 8 vu go hie Se eT I TTT fq Ask if box 3, 4, or 5 marked in Q.5: c. About how long did — — have this condition before it was cured? b. What part of the body is affected now? 0 How is — — (part of body) affected? 000 J Less than 1 month OR ——— { 0 Mists Is —— affect n any other way? Number ears Part(s) of body * Present effects ** d. Was this condition present at any time during the past 12 months? 1 Oves 2 0 No oJ Not an accident/injury (NC) 10 First accident/injury for this person (14) 8] Other (13) K4 ** If multiple present effects, enter in C2 each one that is not the * Enter part of body in same detail as for 3g. same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-90) Page 29 81 82 0 Old age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. Age one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | If “yes, enter = . names in columns 2. [Relationship 2 c. | have listed (read names). Have | missed: Yes | No REFERENCE PERSON — any babies or smallchildren? ..................ccciiuuunnn. Safin 4 3 05 O a Date ohoinh | re Iyear — any lodgers, boarders, or persons you employ who livehere? . ............... 0 Od ! y — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK.DV traveling Or INA NOBPIRAI? « . « « viviv vs ve Comrie e s suns ss sawn yes omkine sens Od ] 0 0 — AIYONOOIBB STAY INBIOP « ; . & + viv vx wiwiv es ss avin sass amaissesusesdesns 0 = IC 1 [pol None sD waa wes [02 Nore “Nomper |2LIW ell Ne | pe d. Do all of the persons you have named usually live here? [J Yes (2) ii YS [J No (APPLY HOUSEHOLD MEMBERSHIP jd Probe if necessary: RULES. Delete nonhousehold members |G 2 by an “’X’’ from 1—C2 and enter reason.) | | _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ Does — — usually live somewhere else? kA TA 1oy Inte ip) hgicohD 1 1 1 1 1 1 Ask for all persons beginning with column 2: 2. What is — — relationship to (reference person)? [iA — TRE —1BV" TIN. TCCLIRI HSTCONE 3. What is — — date of birth? (Enter date and age and mark sex.) ' ' y ! ! ' REFERENCE PERIODS [LA 7 TRA IDV TINJ.TCLLTRI HSTCOND | | [I Io 2-WEEK PERIOD L - bd bd |12-MONTH DATE [LA ~ IRA IDV 1INJ.TCLLTRI HSICOND. mt mms rr SG TT TT ot te ee or mS i | | | Io 1 1 1 : 1 1 13-MONTH HOSPITAL DATE A 2 [LA ~ TRA “IDV TINJ.TCLLTRI HSTCOND. ASK CONDITION LISTS 1,2, and 3. I 4 4 1 CONDITION 3 Name of condition . When did [— —/anyone] last see or talk to a doctor or assistant Mark “’2-wk. ref. pd.’’ box without asking if ’“DV’" or "HS"’ in C2 as source. about — — (condition)? 5 O 2 yrs., less than 5 yrs. 6 0 5 yrs. or more 7 [J Dr. seen, DK when 8 0 DK if Dr. seen } 9 [J Or. never seen 0 [J Interview week (Reask 2) 1 [J 2-wk. ref. pd. 2 [J Over 2 weeks, less than 6 mos. 3 [J 6 mos., less than 1 yr. a4 0 1 yr., less than 2 yrs. b. What did he or she call it? (Specify) 1 [J Color Blindness (NC) 2 [J cancer (3e) 3[J Normal pregnancy, 4 [J o1d age (NC) normal delivery, (5) 8 0 Other (3c) vasectomy c. What was the cause of —_— (condition i in 3b)? (Specify) 4 Mark box « if accident or injury. od Accident/injury (5) d. Did the (condition in 3b) result from an accident or injury? . (Earlier you told me about — — (condition)) Did the doctor or assistant Aiiment Cancer Disease Problem Anemia Condition Disorder Rupture Asthma Cyst Growth Trouble Attack Defect Measles Tumor Bad Ulcer . What kind of (condition in 3b) is it? (Specify) call the (condition) by a more technical or specific name? 100 Yes 2 [No 9 [J ok Ask 3b if “’Yes’’ in 3a, otherwise transcribe condition name from item 1 without asking: 100 ves (5) 20 No . What part of the (part of body in 3b—g) is affected by the [infection/ Ask 3g if there is an impairment (refer to Card CP2) or any of the following entries in 3b—f: Abscess Damage Palsy Ache (except head or ear) Growth Paralysis Blood clot Infection Sore(ness) Boil Inflammation Stiff(ness) Cancer Neuralgia Tumor Cramps (except menstrual) Neuritis Ulcer Cyst Pain Varicose veins Weak(ness) What part of the body is affected? (Specify) Show the following detail: HOB. « «vis «minions seins «aE EEE dE Cee skull, scalp, face Back/spine/vertebrae . . ..............cc00innnanann upper, middle, lower BMD « +See vialniare niniain mR ERIE Wr TREE 5 nein owes left or right BA 5 TES SRT ERs. $i Shel SAR inner or outer; left, right, or both BY® viviv.v altinks wi alninin hs arasaln wv sEwinln « SEWiaE & Sain Rie sal Ew, left, right, or both BU ooo vivinie vivimsnio shoulder, upper, elbow, lower or wrist; left, right, or both BBN. vine nc re ee entire hand or fingers only; left, right, or both RB oui0ins 4 waminnm wwe wich hip, upper, knee, lower, or ankle; left, right, or both POOR. « viviv sc vivivin vs viivinls viviiow entire foot, arch, or toes only; left, right, or both Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: Infection Sore Soreness sore/soreness] — the skin, muscle, bone, or some other part? (Specify) Ask 3e if the condition name in , 3b includes any of the following words: | Ask if there are any of the following entries in 3b—f: Tumor Cyst Growth Is this [tumor/cyst/growth] malignant or benign? 9 Cok 1 0 Malignant 2 [ClBenign Ask 3f only if allergy or stroke in 3b—e: How does the [allergy/stroke] NOW affect ——? (Specify) 4 For Stroke, fill remainder of this condition page for the first present effect. Enter in item C2 and complete a separate condition page for each additional present effect. 1 0 2-wk. ref. pd. 2 [J Over 2 weeks to 3 months 3 [J Over 3 months to 1 year a4 0 Over 1 year to 5 years 5 0 Over 5 years a. When was — — (condition in 3b/3f) first noticed? b. When did —— (name of injury in 3b 3b)? Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) (Was it less than 1 year or more than 1 year ago?) (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 30 oud age Ooid age oid age Cloud age 4. | First name Mid. init. FAge First name Mid. init. fAge 1. |First name Mid. init. fAge First name Mid. init. FAge Last name ex Last name ex Last name ex Last name ex 10m 10m 10m 10m 20F 200F 2[JF 200F 2. | Relationship Relationship 2. [Relationship Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth ‘ Month Date Year Month Date | Year Month Date [Year Month Date 1 1 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD | 2-WK. DV 1 1 HOSP. WORK RD [2-WK. DV| ‘HOSP. WORK c1 ool None| , wa 100 yes [000 None 00] None we hives 00] None c1 00] None JClwe ives oJ None [oo [J None ay "Number 20No “Number | Number 2U0we CINo "Number Number 20wb [2[INo Number | Number 2Owo EL rs mL T= c2 LA ~ JRA ~|DV |INJT [CLLTR|HS T|COND.|[LA ~ [RA [DV [INJ.” |CLLTR|HS [COND. CA~ T|RA DV |INJ. |CLLTRIHS [COND.JUA™ TRA | DV|INJ. Lo Pol I HE ET [LA TRA ~ DV |INJ. [CLTTRIHS | | | | | 1 Lal 1 1 | | | 1 1 1 MA — JRA —|OV INU" [CLTTA|HS JCOND.|TA™ — [RA™ [OV TINJ. "| CLLTRTHS COND. CA ~|RA~ TDV JINJ. CLITRTHS [CONGJLA ~ TRA —; DV INJ. JCCLTRAS | COND. I I | | | I | | | | | | | | | | I | | | | | | | 1 1 1 I 1 1 l 1 "i". 1 1 1 1 1 1 1 1 1 l J 1 l, 1 l TeONG. | TA ~ [RA™ TOV TINJ. | CLITATHS =| [A~ "|RA~ TDV [INJ. TjCLLTRTHS NL ~ TRA ~| DV] INJ. TCCLYR| HS | COND. [LTA — TRA — DV |INJT CUTTRIHS COND. |TA™ ~ [RA | | | | | 1 1 1 | 1 LA IRA IDV |INJ. |CLLTR|HS |COND.|LA IRA IDV |INJ. |CLLTRIHS |COND. | | | | | | | | | | | | | | | | | IRA [DV [INJ. |CLLTRIHS [COND|LA | | | | | | | | | | | | | | | | | | Refer to RD and C2. K 1 10] ""Yes’ in "RD" box AND more than 1 condition in C2 (6) 8 [J Other (k2) 6a. During the 2 weeks outlined in red on that calendar, did — — (condition) cause — — to cut down on the things — — usually does? 13. Is this (condition in 3b) the result of the same accident you already told me about? J ves (Record condition page number where i ions first leted.) = (NC) Page No. Ono 32:0 Oves Ono x2) b. During that period, how many days did — — cut down for more than half of the day? 14. 00 Od None (K2) — Days 7. During those 2 weeks, how many days did — — stay in bed for more than half of the day because of this condition? Where did the accident happen? 100 At home (inside house) 200 At home (adjacent premises) 3d Street and highway (includes roadway and public sidewalk) 4[J Farm s(J Industrial place (includes premises) K 2 [J condition has “CL LTR" in C2 as source (10) [J Condition does not have ““CL LTR" in C2 as source (K4) 00[INone mms D8Y 5 6] school (includes premises) Ask if “’"Wa/Wb'’ box marked in C1: 70 Place of recreation and sports, except at school 8. During those 2 weeks, how many days did — — miss more than 8] Other (Specify) ¥ half of the day from — — job or business because of this condition? 00[JNone Days Mark box if under 18. [J Under 18 (16) " —— 1 i Fok fogs 577: 15a War p Udo 8 when |e accident happened? 9. During those 2 weeks, how many days did — — miss more than Pole ea BD half of the day from school because of this condition? b. Was — — in the Armed Forces when the accident happened? Oo 2[J Yes (16) Ono 00L_ None aneesDBYE, ee oh i Sra enti Fin Feige = Emits =o) . Was — — at work at — — job or business when the accident happened? 3] ves 40No 10. About how many days since (12-month date) a year ago, has this condition kept — — in bed more than half of the day? (Include days 16a. Was a car, truck, bus, or other motor vehicle involved in the accident in any way? 100 Months 20 years d. Was this condition present at any time during the past 12 months? 10ves 20No 000[JLess than 1 month OR ——— { Number while an overnight patient in a hospital.) Oves 20N02 b. Was more than one vehicle involved? 000 J None Days 1[J Ves 2CINe 11. Was — — ever hospitalized for — — (condition in 3b)? c. Was [it/either onel moving at the time? 10ves 2[0No 100 ves 20No [J Missing extremity or organ (K4) 17a. At the time of the accident what part of the body was hurt? K3 [J other (12) What kind of injury was it? A : 12a. Does — — still have this condition? fything else? 10 ves (K4) Ono Part(s) of body * Kind of injury b. Is this condition completely cured or is it under control? 20 cured 8] Other (Specify) ¥ 3 Od Under control (K4) RAI Rr TG TT mre mm et lt ce a ce mt tS i c. About how long did — — have this condition before it was cured? b. Ask if box 3, 4, or 5 marked in Q.5: What part of the body is affected now? How is — — (part of body) affected? Is — — affected in any other way? Part(s) of body * Present effects ** oJ Not an accident/injury (NC) K4 100 First accident/injury for this person (14) 8] Other (13) * Enter part of body in same detail as for 3g. ** If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-90) Page 31 83 0 Old age A. HOUSEHOLD COMPOSITION PAGE 1 by an “’X’’ from 1—C2 and enter reason.) Does — — usually live somewhere else? Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. |First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | If "Yes," enter ) = y names in columns Relationship 2 c. I have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies or smallchildren? ...............ccovvvvvnnnennnnnnn..in..| O 0 Dustin ie Waar — any lodgers, boarders, or persons you employ who livehere? ................ 0 0 . : — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK.DV travelingorina hospital? . . ..............00tiurnnnennennennenneenns Od Od 0 0 — anyone else staying here? . ...............uuieininnnenrnennnnenanann i] O IC 1 poli Nore 2 Lda y [J ves Po Nore Tamper [20 Wo pL No | np d. Do all of the persons you have named usually live here? OJ Yes (2) AL ode MAL [J No (APPLY HOUSEHOLD MEMBERSHIP Probe if necessary: RULES. Delete nonhousehold members REFERENCE PERIODS 2-WEEK PERIOD ar, 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. LA TRA IDV TINJ.TCLLTRI HSTCOND. | | | | Io 1 1 1 1 [LA ~ TRA “IDV TINJ.TCLLTR| HSTCOND | | | | 1 1 1 1 1 l 1 [LA ~ TRA “IDV TINJ.TCLLTRI HSTCOND. | | | | Io | 1 1 1 1 4 [LA ~ IRA IDV 1INJ.ICLLTRI HSICOND. | I | | Io 1 ccs 1 1 1 | LA IRA IDV |INJ.|CLLTRI HSICOND. | | | | Io | | | | lo CONDITION 4 PERSON NO. Ask 3g if there is an impairment (refe following entries in 3b—f: 1. Name of condition r to Card CP2) or any of the f. How does the [allergy/stroke] NOW affect — —? (Specify) 32 Ask probes as necessary: or was it before that date?) effect. Enter in item C2 and complete a separate condition page for each additional present effect. Abscess Damage Palsy Ache (except head or ear) Growth Paralysis Mark “*2-wk. ref. pd.’’ box without asking if ’“DV’’ or “’HS"’ Bleeding (except menstrual) Hemorrhage Rupture in C2 as source. Blood clot Infection Sore(ness) 2. When did [— —/anyone] last see or talk to a doctor or assistant Boil Inflammation Stiffness) about — — (condition)? Cancer Neuralgia Tumor 0 [ Marview week fReask 2) 5 [J 2 yrs., less than 5 yrs. Cramps (except menstrual) Neuritis Ulcer 1 Cl Zk. ref pd 6 [] 5 yrs. or more Cyst Pain Varicose veins 2 [J Over 2 weeks, less than 6 mos. 0 Dr.seen, DKwhen ~~ | Weakness) 30s mos., less than 1 yr. 8 [J DK if Dr. seen } 135) a0 yr., less than 2 yrs. 9 [J Dr. never seen g. What part of the body is affected? (Specify) 3a. (Earlier you told me about — — (condition)) Did the doctor or assistant Show the following detail: call the (condition) by a more technical or specific name? 0 i 0 N g 0 OK BBB Lx + cainiv svininie wwii wa RAS RES SRE ei skull, scalp, face 1 6s 9 Back/spine/vertebrdn . . . «viv s s.5v's vive vniv es vives upper, middle, lower Ask B5IE "You" in 89, cllerwiseiransarie condition name Hom SIO + 5 vimiv 3 wT ES STE AWE CAE SS SR CE left or right item 1 without asking: BRP wiv ivjvin an aisinie b Samia Nii Se Re inner or outer; left, right, or both BYS iv «ivinilnrs win ny RTE I A TR eT a A left, right, or both b. What did he or she call it? (Specify) APN aviv imitans m minima shoulder, upper, elbow, lower or wrist; left, right, or both y p Y. AR. «arene iinnine iw a rmanes inn ae entire hand or fingers only; left, right, or both 10 Color Blindness (NC) 2 [J cancer (3e) O RBG «oven sivmnim iw emma hip, upper, knee, lower, or ankle; left, right, or both 30 Normal pregnancy, 4 LJ Old age (NC) F tire h ly: left, righ both normal delivery, (5) % 0 Other (3c) + SE entire foot, arch, or toes only; , right, or bot! vasectomy c. What Was the causs of a (condition in 3b (Specify) 3 TT Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: Infection Sore Soreness Te CTT TT 55 ——————-—=—==-4 h. What part of the (part of body in 3b—g) is affected by the [infection/ Mark box if accident or injury. o [J Accident/injury (5) sore/soreness] — the skin, muscle, bone, or some other part? d. Did the (condition in 3b) result from an accident or injury? 10 ves (5) 20 No (Specify) Ask 3e if the condition name in 3b includes any of the following words" Ask if there are any of the following entries in 3b—f: Ailment Cancer Disease Problem Tumor Cyst Growth Asthma Cyst Growth Trouble 4. Is this [tumor/cyst/growth] malignant or benign? Attack Defect Measles Tumor Bad Ulcer 1d Malignant 2 [Benign 9 Ook i ptr 2 : . —— | ition ii -wk. ref. pd. e.What kind of (conditionin3b)isit? ______ 2 y ha n ee a7 ‘condition in 3b/3f) 1 0 2-wk. ref. pa - (Specify) 5 rst notice 2 [J Over 2 weeks to 3 months Ask 3f only if allergy or stroke in 3b—e: b. When did —— (name of injury in 3b)? |, [7 over 1 year to 5 years (Was it on or since (first date of 2-week ref. period) (Was it less than 3 months or more than 3 months ago?) For Stroke, fill remainder of this condition page for the first present (Was it less than 1 year or more than 1 year ago?) (Was it less than 5 years or more than 5 years ago?) 3 [J over 3 months to 1 year 5 OJ Over 5 years FORM HIS-1 (Evaluation) (2-1-9C" Page 32 84 [oid age [oid age Joid age [Joid age 4. | First name Mid. init. FAge First name Mid. init. Age 1. [First name Mid. init. Age First name Mid. init. Age Last name eX Last name ex Last name ex Last name ex 10m 10m 10m 10m 20F 2JF 200F 20F 2. | Relationship Relationship 2. [Relationship |Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth Month |Date Year Month J Date | Year Month |Date Year Month |Date | Year | 1 — 1 1 | | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00] None 1Owal10 ves 00] None [oo] None 10wa © Yes 00] None 20wb|200No 20wb pONo ered Number Number Number Number JRA [DV INJT [CLTTR|HS COND. [RA” TOV TINJ. | CLLTRTHS COND. 1 | | | Io | (1 1 1 1 1 1 1 1 1 1 1 Lol, c1 00[J None 10wa [10 Yes loo [J None [oo [J None 10Owa 10 Yes 00] None 20ws [200No 200wb |20No Number Number Number Number | | | | | | | | | ILA — TRA —|OV INJT [CLLTR|HS COND. [TA ~ [RA [DV |INJ.” | CLLTRHS [TOND. | | | 1 1 i 1 1 1 1 1 1 1 1 1 CA~ ~|RA™ TDV jIND. JE ola T | | | 1 1 1 I~ d g MA ~ TRE BV INIT [CUTTRIHS No ~ RA” OV TINJ. 7| CLITRTHS N | | | | | | | | | | | | 1 Lilien kik 1 ft 1 1 | aA ia Eek. Sje.ons we i ~ TRA" TOV INU [CLTTRIHS COND. |TA™ ~— RA” TOV TINJ. | CLLTAIHS = | | | | | | | | | | | | 1 1 | 1 1 | | I 1 1 1 1 CA~ ~|RA TDV [INT. TLE eon OCA ~ THA ~ DV] INJT TCCLTR HS T COND. | 4 | I | 1 I | | Seine LA IRA IDV |INJ. [CLLTRIHS |COND.|LA IRA [DV |INJ. |CLLTRIHS |COND. | | | | | | | | | | | | | | | | | | | | | | | | IRA [DV |INJ. |CLLTRIHS |COND.|JLA | | | | | | | | | | | | | | | Refer to RD and C2. K 1 1 "Yes" in "RD" box AND more than 1 condition in C2 (6) 8 [J Other (k2) 6a. During the 2 weeks outlined in red on that calendar, did — — 13. Is this (condition in 3b) the result of the same accident you already told me about? 0 Yes (Record condition page number ier first (NC) O No Page No. cause — — to cut down on the things — — usually does? Yes OI No (k2) b. During that period, how many days did — — cut down for more than half of the day? 00 None (k2) ee DOYS 7. During those 2 weeks, how many days did — — stay in bed for more than half of the day because of this condition? 14. Where did the accident happen? 1 At home (inside house) 200 At home (adjacent premises) 3[J street and highway (includes roadway and public sidewalk) Farm s(J Industrial place (includes premises) J School (includes premises) 0 Place of recreation and sports, except at school 8] Other (Specify) Z Mark box if under 18. OJ Under 18 (16) 00 a None eee DEY S Ask if ““Wa/Wb'’ box marked in C1: 8. During those 2 weeks, how many days did — — miss more than half of the day from — — job or business because of this condition? 00 £5) None Days Ask ifage 5—17: 9. During those 2 weeks, how many days did — — miss more than half of the day from school because of this condition? oo] None — Days K 2 [J Condition has “CL LTR" in C2 as source (10) 15a. Was — — under 18 when the accident happened? 100 ves (16) Ono b. Was — — in the Armed Forces when the accident happened? 2[J ves (16) Ono c. Was — — at work at — — job or business when the accident happened? 3] ves ano [J condition does not have “CL LTR’ in C2 as source (K4) 10. About how many days since (12-month date) a year ago, has this condition kept — — in bed more than half of the day? (Include days while an overnight patient in a hospital.) 000 JNone Days 11. Was — — ever hospitalized for — — (condition in 3b)? 10ves 20 No 16a. Was a car, truck, bus, or other motor vehicle involved in the accident in any way? 20 No (17) c. Was [it/either one] moving at the time? 100 ves 20No K 3 Od Missing extremity or organ (K4) OJ other (12) 12a. Does — — still have this condition? 10ves (ka) Ono b. Is this condition completely cured or is it under control? 20cured 3[Junder control (k4) c. About how long did — — have this condition before it wa: 8 O Other (Specify) z 100 Months 20 years d. Was this condition present at any time during the past 12 months? 10Yes 20No 000 Less than 1 month OR ——— { Number 17a. At the time of the accident what part of the body was hurt? What kind of injury was it? Anything else? Part(s) of body * Kind of injury Ask if box 3, 4, or 5 marked in Q.5: b. What part of the body is affected now? How is — — (part of body) affected? Is — — affect: in any other way? Part(s) of body * Present effects ** oJ Not an accident/injury (NC) K4 10 First accident/injury for this person (14) 8] Other (13) * Enter part of body in same detail as for 3g. * If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-0) Page 33 85 J od age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name 5 b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes,” enter 0 y names in columns 2. [Relationship c. I have listed (read names). Have | missed: Yes | No REFERENCE PERSON — any babies or smallchildren? ....................000n.. rr mm 0 0 3. |Dateofbith Yaar — any lodgers, boarders, or persons you employ who livehere? . ............... Od Od ; : — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK. DV travelingorina hospital? . ...............c0uiirnnnennrnnennnnnnennns Od ad Cobin Wm — anyone else staying here? . .. ..............uouieeeeunuinneeennnneeennnn O 0 C1 ore wa [100 ves |° ong Number 20] No Number d. Do all of the persons you have named usually live here? [Yes (2) - [J No (APPLY HOUSEHOLD MEMBERSHIP REE Probe if necessary: RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) Does — — usually live somewhere else? Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? iA — TE IBV" TiN. TCCLTR| HSTCONG. 3. What is — — date of birth? (Enter date and age and mark sex.) ! rE ! ! . y REFERENCE PERIODS ~~ | ~~ LA IRA IDV TINJ.|CLLTR| HSTCOND. | | | 1 2-WEEK PERIOD L L 1 L be 12-MONTH DATE LA IRA IDV IINJ.ICLLTRI HSICON 13-MONTH HOSPITAL DATE 2 LA IRA IDV [INJ.|CLLTRI HSICOND. A2 |\sk CONDITION LISTS 1,2, and 3. iF + id CONDITION 5 | PERSON NO._ Ask 3g if there is an impairment (refer to Card CP2) or any of the following entries in 3b—f: 1. Name of condition Abscess Demags Palsy Ache (except head or ear) Growth Paralysis Mark **2-wk. ref. pd.’’ box without asking if ’DV’’ or ’HS"’ B H h Pp in C2 as source. Blood clot Infection Sore(ness) 2. When did [— —/anyone] last see or talk to a doctor or assistant Boll Inflammation Stiffness) about — — (condition)? Cancer Neuralgia Tumor 0 [J Interview week (Reask 2) 50] 2yrs.. less than 5 yrs. Crane {except menatruni) Noire ulcer oi 1 [0 2-wk. ref. pa. 6 [J 5 yrs. or more ys aly ease veins 2 [J Over 2 weeks, less than 6 mos. 70 Dr. seen, DK when | 8alciness) ase mos., less than 1 yr. 8 [J DK if Dr. seen } (3b) a1 yr., less than 2 yrs. 9 [J or. never seen g. What part of the body is affected? (Specify) 3a. (Earlier you told me about —— (condition) Did the doctor or assistant Show the following detail: call the (condition) by a more technical or specific name? "ai kod or VEE Be Sens Cee ee ER SEE See , scalp, face 10] ves 20No sok Back/spine/vertebrae . ..............c00nnnnneanan upper, middle, lower ef FAINT i eh Ee rE rT B07. vvi% +o THES GARE SIR A RT SAS SA Woes § left or right Ask 3b if “’Yes’’ in 3a, otherwise transcribe condition name from item 1 without asking: nd FAR Tp wR WIR SER STE or inner or outer; left, right, or both BEIVEE SAUTE £AWIE & SRI 3 Hisatals smmnsve stseanste = Sues + left, right, or both b. What did he or she call it? Ry: — ae yn is | Yo rhe, py Beaty} Ameren... shoulder, upper, 3 tat, ” 1 color Blindness (NC) 2 [J cancer (3e) and EE es ie, SAIS MAE St dons any: foi oy "oth sO n ; ; sOodageves | rem s , , ; loft, y ey, } (5) loner BOOT crss0.0 40005808 3140708 HEE entire foot, arch, or toes only; left, right, or both vasectomy c. What Was the cause of eae] (condition in 3b) (Specify) 5 TTT] Ci for oye, oars, of ipeerns) organs, ask 3h if there are any of the ollowing entries in 3b—f: Infection Sore Soreness I I y= ~~ ~~~ ———~——=~———-1 h. What part of the (part of body in 3b—g) is affected by the [infection/ Mark box if accident or injury. o [J Accident/injury (5) or — the skin, muscle, ly or some a part? d. Did the (condition in 3b) result from an accident or injury? 10 Yes (5) 20 No (Specify) Ask 3e if the condition name in 3b includes any of the following words: J Ask if there are any of the following entries in 3b—f: Ailment Cancer Disease Problem Tumor Cyst Growth Ansmi Conditl Op Asthma Cyst Growth Trouble 4. Is this[ /cyst/growth] malig or benign? Attack Defect Moeasl Tumor Bed ot see Ulcer v3 Malignant 2 [Benign 9 Ook tatu} a. When was — — (condition in 3b/3f) 1 [J 2-wk. ref. pd. e. What kind of (condition in 3b) is it? Specify) first noticed? A mt sont mt et st re rm ee SS Fe, Sl. St Sts eh — ET 3 [J over 3 months to 1 year Ask 3f only if allergy or stroke in 3b—e: b. When did —— (name of injury in 3b)?) 4 [J Over 1 year to 5 years f. How does the [allergy/stroke] NOW affect — —? (Specify) 2 5 [J Over 5 years Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) For Stroke, fill semaines of this condition page for he first present (Was it less than 1 year or more than 1 year ago?) effect. Enter in item and complete a separate condition page for each additional present effect. (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 34 Dow age Cloud age Cold age [Jord age 41. | First name Mid. init. FAge First name Mid. init. fAge 1. [First name Mid. init. fAge First name Mid. init. fAge Last name ex Last name ex Last name Sex Last name ex 10m 10m 10m 10m 200F 200F 200F 200F 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth Month | Date Year Month Date | Year Month |Date [Year Month Date | Year 1 1 1 1 = 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| ‘HOSP. WORK RD 2-WK. DV oon. c1 one Owe 100 yes [00] None [ool None {Owe hClves 00] None c1 00 J None JCwe lives oJ None [oo] None enLm 00] None Number 2Uwo|2lINo Number Number 2Odws RlINo Number Number 20wb |2[INo Number | Number 200we |200No Number LA |RA IDV |INJ. |CLLTR|HS |COND.|LA |RA™ [DV |INJ. | CLLTRIHS |COND. LA IRA [DV |INJ. |CLLTRIHS |COND.|LA | RA | DV] INJ. |CLLTRIHS | COND. | | | | I | I | 1 | I I | | | | | | 1 1 I 1 1 | 1 l 1 1 1 | 1 1 1 1 1 1 1 | ' : 1 1 MLA ~ TRE —|0V {INU [CLTTRHS COND. |TA™ — [RA™ OV TIND. | CLITRTHE [COND. —|RA™ TOV INT. = ops JCONG|A ~ THA BV INJ. TCCLTR HS | COND I I I | I | I | I I | I | | | I | | | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ) . ! MLA ~ TRE ~|OV INU. [TCUTTAIHS 2 TA — RA” OV TINJ.™| CLOTAHS ul CA~ ~|RA~ TOV JIND. non eons] A ~ TRA | DV] INJT TCCLTR AS | COND. 1 | I I I I I I I I I | I | I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 ! 1 l 1 : ; ’ : Le ~ TRA BV INIT [CUTTRIHS CONG. |TA™ — (RA™ TOV moe ~|RA~ TOV TINJ. Tes JCONG|UA ~ TRA —| DV] INJT TCCLTRIAS COND. I I I I I I I | I I | I | 1 I ! I | | I | 1 1 1 1 1 1 L 1 1 1 1 1 1 1 1 1 : L 1 i l 1 1 1 [LA ~ TRA TOV INU. [CLLTAIHS TCOND.|LA ~ (RA OV TINJ. | CLLTRTHS [CON CA™ T|RA™ TDV TINJ. |CLLTAIHS [CONDJUA ~ TRA ~| DV] iNJ. TCLLTRAS | COND | CLLTRIHS |COND. | | | 1 I | | | | | | | | | | | | | | | | | Refer to RD and C2. I 10 Yes’ in *“RD"* box AND more than 1 condition in C2 (6) K1 8] Other (k2) 6a. During the 2 weeks outlined in red on that calendar, did — — (s 13. Is this (condition in 3b) the r sult of the s same accident you already told me about? 0 Yes (Record condition page number Wise first (NC) Ono Page No. cause — — to cut down on the things — — usually does? ves OI No k2) b. During that period, how many days did — — cut down for more than half of the day? 00 JNone (k2) 7. During those 2 weeks, how many days did — — stay in bed for more than half of the day because of this condition? Days 00 None Ask if “Wa/Wb’’ box marked in C1: 8. During those 2 weeks, how many days did — — miss more than half of the day from — — job or business because of this condition? Days 14. Where did the accident happen? 10 At home (inside house) 20 At home (adjacent premises) 30] Street and highway (includes roadway and public sidewalk) 40 Farm sO] Industrial place (includes premises) 6] School (includes premises) 703 Place of recreation and sports, except at school aL] Other (Specify) 00 4d None Days Mark box if under 18. CO under 18 (16) Ask if age 5—17: 9. During those 2 weeks, how many days did — — miss more than half of the day from school because of this condition? 00JNone Days [J Condition has “CL LTR" in C2 as source (10) 15a. Was — — under 18 when the accident happened? 100 Yes (16) Ono b. Was — — in the Armed Forces when the accident happened? 200 ves (16) Ono c. Was —— at work at — — job or business when the accident happened? ad Yes 40No K 2 [J condition does not have “CL LTR" in C2 as source (K4) 10. About how many days since (12-month date) a year ago, has this condition kept — — in bed more than half of the day? (Include days while an overnight patient in a hospital.) ooo] None eens: DAYS 11. Was — — ever hospitalized for — — (condition in 3b)? 10 ves 20 No 16a. Was a car, truck, bus, or other motor vehicle involved in the accident in any way? 20No (17) c. Was [it/either one] moving at the time? 100 ves 20 No K 3 0 Missing extremity or organ (K4) [J other (12) 12a. Does — — still have this condition? 100 ves (ka) Ono b. Is this condition completely cured or is it under control? 20cured 8 [J other (specify) 3 0 Under control (K4) (K4) — have this condition before it was cured? 1 J Months 20 Years d. Was this condition present at any time during the past 12 months? 10 ves 20No cool Less than 1 month OR ——— { Number 17a. At the time of the accident what part of the body was hurt? What kind of injury was it? Anything else? Part(s) of body * Kind of injury Ask if box 3, 4, or 5 marked in Q.5: b. What part of the body is affected now? How is — — (part of body) affected Is — — affected in any other way? Part(s) of body * -~ Present effects ** oJ Not an accident/injury (NC) 10 First accident/injury for this person (14) 8] Other (13) K4 * Enter part of body in same detail as for 3g. ** If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-90) Page 35 87 88 a Old age d. Do all of the persons you have named usually live here? Probe if necessary: Does — — usually live somewhere else? [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. Age one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Ph b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes, enter 20 y names in columns Relationship c. I have listed (read names). Have | missed: Yes | No REFERENCE PERSON — any babies or small children? ........ RE THAR ER SANE ® ENTREE J 0 Date obbinth re Year — any lodgers, boarders, or persons you employ who livehere? ................ 0 0 y ! — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK.DV travelingorinahospital? . ................0iiinurrrnnnnnnnnnnnnennn 0 O Cc 1 bol Nore TO — anyone else stayinghere? . ........... Selene 4 ARENT 8 ATES AH EET, A O al 10wa [100 Yes Toma Wb (20 No | pe Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD A1 13-MONTH HOSPITAL DATE A 2 ASK CONDITION LISTS 1,2, and 3. RULES. Delete nonhousehold members C 2 by an “’X"’ fr —- Il Omorroason.)] 1° Joes symm, mmm aoe a me y om 1—~C2and enter 1eason.) LA JRA "IDV TiNJ.TCLLTRI HSTCOND. | | | 1 1 1 1 1 1 1 ILA" = TRA TIDV TINJ.TCLLTRI HSTCOND. | | [I | 1 L dad _ [LA ~ TRA “IDV TINJ.TCLLTRI HSTCOND. | | | Io 1 | 1 1 1 1 LA IRA IDV IINJ.1CLLTRI HSICOND. = ii eye mit Siri SR | | Io | | 1 I I [LA ~ TRA “IDV TINJ.TCLLTRI HSICOND. | | Lo [I | | | I CONDITION 6 PERSON NO.____ 1. Name of condition i Mark *’2-wk. ref. pd.’’ box without asking if ‘DV’ or *’HS"’ in C2 as source. 2. When did [— —/anyone] last see or talk to a doctor or assistant about — — (condition)? s[2 yrs., less than 5 yrs. 6 0 5 yrs. or more 70 or. seen, DK when 8 [J DK if Dr. seen } 9 [J Or. never seen 0 [J Interview week (Reask 2) 1 Od 2-wk. ref. pd. 2 [J Over 2 weeks, less than 6 mos. ad 6 mos., less than 1 yr. 401 yr., less than 2 yrs. 3a. (Earlier you told me about — — (condition)) Did the doctor or assistant call the (condition) by a more technical or specific name? 100 ves 2 Ono es [J ok Ask 3b if “’Yes’’ in 3a, otherwise transcribe condition name from item 1 without asking: b. What did he or she call it? (Specify) 1 Color Blindness (NC) 2 [J cancer (3e) 300 Normal pregnancy, 4 oid age (NC) normal delivery, (5) 8 [J Other (3c) vasectomy Mark box if accident or injury. o [J Accident/injury (5) d. Did the (condition in 3b) result from an accident or injury? 100 Yes (5) 20 No Abscess Damage Palsy Ache (except head or ear) Growth Paralysis H hag Blood clot Infection Sore(ness) Boil Inflammation Stiff(ness) Cancer Neuralgia Tumor Cramps (except menstrual) Neuritis Ulcer Cyst Pain Varicose veins Weak(ness) . What part of the body is affected? (Specify) Show the following detail: HOB, ivi via vo aie SO ERR RRR RR RR skull, scalp, face Back/spine/vertebrae . ..............c00 inna upper, middle, lower SHOE ain 2 mim mma a in A TR Te SR SR ST left or right BBE: iresein « snmere.& wraaie wish iurnos whe beim ea hesemie winks inner or outer; left, right, or both BYG: semis x oreieres winsare + an ine FS NRA AWE A left, right, or both AVI isn ween 4 ao ain sn Sow shoulder, upper, elbow, lower or wrist; left, right, or both BOR «viv cove wininin win we wm entire hand or fingers only; left, right, or both OG vv vimiminie vim seni niin hip, upper, knee, lower, or ankle; left, right, or both FOOL vivin viviiwiicnwivivimivie slwians entire foot, arch, or toes only; left, right, or both . What part of the (part of body in 3b—g) is affected by the [infection/ Ask 3g if there is an impairment (refer to Card cP2) or any of the following entries in 3b—f: Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: sore/soreness] — the skin, muscle, bone, or some other part? (Specify) Ask 3e if the condition name in 3b includes any of the following words: Aillment Cancer Disease Problem A 5 Condit Disord: Asthma Cyst Growth Trouble Attack Defect Measles Tumor Bad Ulcer Ask if there are any of the following entries in 3b—f: Tumor Cyst Growth Is this [tumor/cyst/growth] malignant or benign? 153 Malignant 2 [Osenign 9 Ook eo. What kind of (condition in 3b) is it? (Specify) Ask 3fonly if allergy or stroke in 3b—e: f. How does the [allergy/stroke] NOW affect ——? (Specify) 5 For Stroke, fill remainder of this condition page for the first present effect. Enter in item C2 and complete a separate condition page for each additional present effect. 1 0 2-wk. ref. pd. 2 [J Over 2 weeks to 3 months 3 [J over 3 months to 1 year 4 overt year to 5 years s (Jovers years a. When was — — (condition in 3b/3f) first noticed? b. When did — — (name of injury in 3b)?| Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) (Was it less than 1 year or more than 1 year ago?) (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 36 Cord age Cloud age Cod age Cloud age 1. | First name Mid. init. FAge First name Mid. init. fAge 1. |First name Mid. init. FAge First name Mid. init. fAge Last name ex Last name ex Last name ex Last name 8X 10m 10m 10m 10m 200F 200F 200F 20F 2. | Relationship Relationship 2. |Relationship Relationship 3. |[ Date of birth Date of birth 3. |Date of birth Date of birth Month ) Date Year Month ) Date I Year Month y Date Year Month Date | Year 1 1 sboarins | 1 1 | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| ‘HOSP. WORK RD 2-WK. DV c1 00] None 10Owel100 Yes 00] None [oo [J None 10wa hO ves 00] None 20wb [200No 20wb No Number Number Number Number | | | I | | | | | | | | 1 1 3. 1 1 | 1 1 1 1 1 | c1 00] None 10wa [100 Yes loo [J None [oo [J None 10wa [100 ves 00] None 20wb [200No 200wb [200 No Number | Number Number Number | | | | | 1 | | | | | | | | | | L 1 1 1 | 1 | MA ~ TRA —|BV INU" [CUTIRHS COND. |TA™ — [RA™ [OV TIND. | TLITRTHS [CONG. I I | | 1 | I | | | | | 1 1 1 1 1 1 1 1 1 1 1 1 CA~ ~|RA™ TDV JING. Rs CORO|LA | | | y 1 l 1 1 1 THA ~| DV INJT TCCLTR) HS T COND. I | I | | | 1 1 1 1 1 1 [LA = TRA 7 |OV INU [CLTTRIHS Ne ~ [RA” TOV TINJ.7| CLLTRTHS [COND. | 1 | Lo | [A | Ld 1 1 1 1 1 1 1 1 1 1 4 A™ iia Te Te Roh ol TRA 7 DV]| INJ. TCCLTR| HS | COND. | | | | | | | 1 1 1 1 1 * ~ TRA IBV INU [CUTTRIHS TCOND.|TA™ ~ {RA™ TOV TIND. | CLITATHS ~ K 1 1 [J "Yes" in “RD"’ box AND more than 1 condition in C2 (6) 8 [J Other (k2) 6a. During the 2 weeks outlined in red on that calendar, did — — ( cause — — to cut down on the things — — usually does? Yes Ono x2) b. During that period, how many days did — — cut down for more CA” 7 |RA™ DV INJ. T|CLITRIHS [CONO.JLA — THA ~| DV INJT TCLLTRHS 7 COND. i | Id 1 Io | Io | i | 1 1 | 1 [I 1 le lt} 1 cd 1 fad 1 ed 1 i 1 Lt 1 A) [LA ~ TRA ~|DV INJ. [CLLTRIHS TTCOND.|TA~ ~ [RA [DV TINJ.” | CLLTRIHS (COND. CA™ T|RA™ DV [INJ. ICLLTRIHS [COND.JUA™ TRA | DV] INJ. TCLLTRIHS | COND | | | [I 1 1 1 | | 1 1 1 | | _ | (I | [I | [— 1 [I | 1 1 _— J Refer to RD and C2. 13. Is this (condition in 3b) the result of the same accident you already told me about? Oa Yes (Record condition page number where til first ~~ (NC) Page No. Ono than half of the day? 00 None (k2) 7. During those 2 weeks, how many days did — — stay in bed for more than half of the day because of this condition? Days 00 JNone Ask if ““Wa/Wb'’ box marked in C1: 8. During those 2 weeks, how many days did — — miss more than half of the day from — — job or business because of this condition? Days 14. Where did the accident happen? At home (inside house) 2[J At home (adjacent premises) 30 Street and highway (includes roadway and public sidewalk) 4] Farm s{J Industrial place (includes premises) 6] School (includes premises) 0 Place of recreation and sports, except at school 8] Other (Specify) ® J Under 18 (16) d. Was this condition present at any time during the past 12 months? 1 Oves 2 O No 00[JNone Days Mark box if under 18. = 15a. Was — — under 18 when the accident happened? Ask if age 5—17: 100 ves (16) On 9. During those 2 weeks, how many days did — — missmorethan | ~~ — & °° = T° Eee a eer) half of the day from school because of this condition? b. Was —— in the Armed Forces when the accident happened? 2[0 ves (16) Ono 00 None Days = Rm ee me se em ey c. Was — — at work at — — job or business when the accident happened? K 2 [J Condition has *’CL LTR" in C2 as source (10) 3] Yes a No [J Condition does not have “CL LTR" in C2 as source (K4) - 10. About how many days since (12-month date) a year ago, has this 16a. Was, 2 Sal fuck: bus, or other motor vehicle involved in the accident condition kept — — in bed more than half of the day? (Include days 9 0 7 while an overnight patient in a hospital.) DAVY ime BRNOAT i ins we is i ii b. Was more than one vehicle involved? 000 J None Days 10 Yes 200No 11. Was —— ever hospitalized for — — (condition in 36)? 0. Was Elonor ons) moving SEtho timely ~ == = =m mm 10 ves 20 No 100 Yes 200No [J Missing extremity or organ (K4) 17a. At the time of the accident what part of the body was hurt? K3 OJ other (12) What kind of injury was it? Anythi Ise? 12a. Does — — still have this condition? tingieise 100ves as Cn Part(s) of body * Kind of injury b. Is this condition completely cured or is it under control? 20cured 8] other (Specify) ¥ SID rent] ARBEIT TT c. About how long did — — have this condition before it was cured? b. What part of the body is affected now? 0 How is — — (part of body) affected? ooo JLess than 1 month OR —— { ! 0 Manths Is —— affected in any other way? Number 2] Years Present effects ** Part(s) of body * oJ Not an accident/injury (NC) K4 10d First accident/injury for this person (14) 8] Other (13) * Enter part of body in same detail as for 3g. ** If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-90) Pag e 37 89 [J od age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | if “yes, enter 7 = y names in columns 2. [Fertonship 4 c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies or small children? .................... HE SR Od 0 3. [ome otoimn, Wear — any lodgers, boarders, or persons you employ who livehere? ................ Od 0 : y — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD |2-WK. DV traveling or ina hospital? . . .............0tiiriinenneenernnennennnnn Od 0 0 0 —anyoneelse staying here? . ...............couinueinennnennennennnns a Oa IC 1 jpoliNone 10 wa [100 ves [20 Nore = Number 20 wb 200 No Number d. Do all of the persons you have named usually live here? [J Yes (2) - [J No (APPLY HOUSEHOLD MEMBERSHIP - Probe if necessary: RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) | | _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Does — — usually live somewhere else? 1A" Toh =| Dy" TINS. [CLLR HEICOND. 1 1 1 1 1 1 Ask for all persons beginning with column 2: N\ 2. What is — — relati hip to (reference person)? bo — TE —1 BV" TING. TCCLTRI HSTCONG 3. What is —— date of birth? (Enter date and age and mark sex.) fdr Ly REFERENCE PERIODS = + LA TRA “IDV TINJ.TCLLTR| HSTCOND. | | | Il 2-WEEK PERIOD L L dd | |12-MONTH DATE =~ I EE i a i m——— [LA"™ TRA “1DV TINT CLLTRI HSICOND. Loy 13-MONTH HOSPITAL DATE A2 [LA" 7 TRA "IDV TINJ.TCLLTRI HSTCOND. ASK CONDITION LISTS 1,2, and 3. I CONDITION 7 [ PERSON NO. Ask 3g if there is an impairment (refer to Card CP2) or any of the following entries in 3b—f: 1. Name of condition Abscess Damage Palsy Ache (except head or ear) Growth Paralysis Mark **2-wk. ref. pd.’ box without asking if ‘DV’ or "’HS"’ H 9 Pp in C2 as source. Blood clot Infection Sore(ness) 2. When did [— —/anyone] last see or talk to a doctor or assistant Boil Inflammation Stiffiness) about — — (condition)? Cancer Neuralgia Tumor o J interview week (feask 2) 5 [] 2 yrs., less than 5 yrs. Cramps (except menstrual) Neuritis Ulcer vl; rah. pd, eds yrs. or more Cyst Pain Varicose veins 2 [J Over 2 weeks, less than 6 mos. 7 J Dr. seen, DK when W Sakiriasd) s0e mos., less than 1 yr. 8 [J DK if Dr. seen } (36) 401 yr., less than 2 yrs. 9 [J Dr. never seen g. What part of the body is affected? (Specify) 3a. (Earlier you told me about —— (condition) Did the doctor or assistant Show the following detail: call the (condition) by a more technical or specific name? Ov 2 CNG s[lox HBB. + winnie vivre Sunwin ani RE ER WE wie skull, scalp, face } 6s Back/spine/vertebrae . ...................0000nnn upper, middle, lower Jism Ser, tn J Te Re wi Ph Se en en 1 [TTT mmm m——— left or right Ask 3b if ‘Yes’ in 3a, otherwise transcribe condition name from item 1 without asking: Ear wise wmEee seis § ewes Hee Bele Lee inner or outer; wii Sole WO testa r tars arrears assr nasser narra ner left, , or b. What did he or she call it? 5 ity) Baw iia nies 5 aman shoulder, upper, elbow, lower or wrist; left, right, or both 0 Color Bind NC) 2 0 Cancer (Ja) pecity. BAN i vonivinns wiwininin wininsnse simi entire hand or fingers only; left, right, or both : Dior ICISSS 0 BG «vivia viv cuivn vinininim wi ninin hip, upper, knee, lower, or ankle; left, right, or both a0 Normal pregnancy, 4 L1 Old age (NC) x normal delivery, } (5) & CJ Other (3c) POOR «ovine ovvnrn vnwmn mms entire foot, arch, or toes only; left, right, or both vasectomy c. What was the cause of —— (condition in 3b)? (Specify) 7 TT] Except for eyes, ears, or internal organs, ask 3h if there are any of the following entries in 3b—f: Infection Sore Soreness Ee CT TT ToT TTT -= o_o. ~—=—————- h. What part of the (part of body in 3b—g) is affected by the [infection/ Mark box if accident or injury. o [J Accident/injury (5) sore/soreness] — the skin, muscle, bone, or some other part? d. Did the (condition in 3b) result from an accident or injury? 100 ves (5) 20 No (Specify) Ask 3e if the condition name in 3b includes any of the following words: Ask if there are any of the following entries in 3b—f: a : i DJ a L Tumor Cyst Growth Asthma Cyst Growth Trouble 4. Is this [tumor/cyst/growth] malignant or benign? Attack Defect Measles Tumor Bad Ulcer 10d Malignant 2 [Benign 9 [ok regi i a. When was — — (condition in 3b/3f) 1 [J 2-wk. ref. pd. eo. What kind of (condition in 3b) is it? Seay) 5 first noticed? 2 oer ee to Zeit rm mm em mm m= i 3 [J over 3 months to 1 year Ask 3f only if allergy or stroke in 3b—e: b. When did —— (name of injury in 3bl?] , [7 gyer 1 year to 5 years f. How does the [allergy/stroke] NOW affect — —? (Specify) 5 [J Over 5 years Ask probes as necessary: (Was it on or since (first date of 2-week ref. period) or was it before that date?) (Was it less than 3 months or more than 3 months ago?) For Stroke, fill remainder of this condition page for the first present (Was it less than 1 year or more than 1 year ago?) effect. Enter in item C2 and complete a separate condition page for each additional present effect. (Was it less than 5 years or more than 5 years ago?) FORM HIS-1 (Evaluation) (2-1-90) Page 38 Ooi age Cloud age Cord age Cloud age 1. | First name Mid. init. FAge First name Mid. init. §Age 1. |First name Mid. init. §Age First name Mid. init. fAge Last name ex Last name ex Last name ex Last name ex 10m 10m 10m 10m 20k 20JF 2[JF 20F 2. | Relationship Relationship 2. |Relationship Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth Month Date Year Month Date Year Month Date vear Month Date : Year 1 1 1 1 1 1 | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00] None 10Owal10 Yes 00] None Joo] None 10wa hd ves 00] None c1 00 J None 10wa [10 Yes oo [J None [oo [J None 1COwa [1 Yes 00] None Number 200wo|20INe Number Number 20we pOno Number Number 200wb |2[INo Number | Number 200we |200Ne Number LA IRA IDV |INJ. |CLLTR|HS |COND.|LA IRA | | | | | | | | | | | | 1 1 1 1 1 1 1 1 1 1 1 | RA [DV [INJ. |CLLTRIHS |COND.|JLA | | | | | | | | 1 1 1 1 1 3 1 | | | | | | | | | | | | | 1 1 1 1 1 | 1 1 1 1 1 | MA ~ TRA —|OV {INU [CUTTRHS JCOND.|TA™ — [RA™ [OV IND. | TLOTAHE (COND. CA™ ~|RA~ TDV JINJ. CLITA{HS JCONDUA ~ THA DV] INJ. TCCLTR| AS | COND. | | | | | I | | | | | 1 1 1 1 1 1 1 1 1 | | | 1 1 1 1 1 1 1 1 1 1 | 1 LA ~ TRA ~ DV INJT [CLTTRIHS IY TA™ 7 [RA” [DV TINJ.7 | CLLTRIHS ~ CA” T|RA™ DV |INJ. T|CLLTRIHS NL | | | | 1 1 1 | | | | | | | | | | | | | | | | | | | | | | | | Lc ~ TRA ~|BV {INU [CUTTRIHS JCONG. |TA~ — [RA TOV TINJ. | CLITATHS 5 CA~ JRA TDV [INJ. |CLLTRTHS [COND|LA ~ TRA | DV] INJ. TCCLTR|HS | COND. | 1 1 ro 1 ro 1 [I | ro | ot | Io | iA 1 IL 1 | 1 el 1 Lod, 1 LL 1 | | 1 iJ 1 tq [LA ~ TRA ~ IDV INJ. [CLLTRIHS TTCOND.|TA~ ~ {RA [DV TINJ.” | CLLTRIHS COND. LA™ TIRA IDV TINJ. T|CLLTRIHS [CONDJUA~ TRA ~| DV iNJ. TCLLTRIHS | COND | | | | | | | | | | | | | | | | | | | Refer to RD and C2. 10 "Yes" in “RD’* box AND more than 1 condition in C2 (6) 8 [J Other (k2) K1 13. Is this (condition in 3b) the result of the same accident you already told me about? [J ves (Record condition page number where > : z | — 2[0No 10 ves K4 oJ Notan accident/injury (NC) 1 ad First accident/injury for this person (14) 8[J other (13) Part(s) of body * Present effects ** 6a. During the 2 weeks outlined in red on that calendar, did — — (NC) t ition) cause — — to cut down on the things — — usually does? Page No. Yes ONo (k2) Ono b. During that period, how many days did — — cut down for more than half of the day? 14. Where did the accident happen? 1 At home (inside house) 00 INone (k2) Days 20 At home (adjacent premises) 7. During those 2 weeks, how many days did — — stay in bed for 3[J street and highway (includes roadway and public sidewalk) more than half of the day because of this condition? a[J Farm 0 5] Industrial place (includes premises) 00L_INone Days sd School (includes premises) Ask if “"Wa/Wb’’ box marked in C1: 70] Place of recreation and sports, except at school 8. During those 2 weeks, how many days did — — miss more than 8] Other (Specify) Z half of the day from — — job or business because of this condition? 00 JNone Days Mark box if under 18. Under 18 (16) Askifage 5—17- 15a. Was — — under 18 when the accident happened? / 9. During those 2 weeks, how many days did — — miss more than 10] ves ¢ we 0 TI half of the day from school because of this condition? b. Was — — in the Armed Forces when the accident happened? 0 ves (16) Ono 00[JNone Days 2 i al ei a eS de pi a, re St c. Was — — at work at — — job or business when the accident happened? K 2 [J Condition has “CL LTR" in C2 as source (10) 3] Yes i Ono oe [J condition does not have “CL LTR" in C2 as source (K4) - - - 10. About how many days since (12-month date) a year ago, has this 152 Wasa S81 fruck, bus, orather motor velicle involved in the accident condition kept — — in bed more than half of the day? (Include days 0 0 while an overnight patient in a hospital.) WAYS, ee cme TWD oe oe mmm mim mimi or b. Was more than one vehicle involved? 000 INone Days 10 Yes 20 No 11. Was — — ever hospitalized for — — (condition in 3b)? c. Was [it/either one] moving atthe ime? | 10 ves 2[0No 100 ves 200No K3 [J Missing extremity or organ (K4) 17a. At the time of the accident what part of the body was hurt? OJ other (12) Whatidnd ut ury was it? 12a. Does — — still have this condition? nything eles 100 Yes (ka) Ono Part(s) of body * Kind of injury b. Is this condition completely cured or is it under control? 20cured 8 [J other (specify) Od Under control (K4) RE RT a Te 7%. F dm ee = - ya a AE lea mse x 4 Ask if box 3, 4, or 5 marked in Q.5: c. About how long did — — have this condition before it was cured? b. What part of the body is affected now? How is — — (part of body) affected? 000[] Less than 1 month OR — { ya pone Is — — affected in any other way? Number rs * Enter part of body in same detail as for 3g. ** If multiple present effects, enter in C2 each one that is not the same as 3b or C2 and complete a separate condition page for it. FORM HIS-1 (Evaluation) (2-1-0) Page 39 91 92 Od Old age A. HOUSEHOLD COMPOSITION PAGE 1 [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) d. Do all of the persons you have named usually live here? Probe if necessary: Does — — usually live somewhere else? Ask for all persons beginning with column 2: 2. Whatis — — relati hip to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) REFERENCE PERIODS 2-WEEK PERIOD EE 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. L. DEMOGRAPHIC BACKGROUND PAGE 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. Age one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Soh b. What are the names of all other persons living or staying here? Enter names in columns. | if “Yes, enter 0 . names in cok « |Relationship c. | have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies or smallchildren? .................cc0iuiiiininirnnnrnnennns 0 Doms Shin | te Wear — any lodgers, boarders, or persons you employ who livehere? ................ 0 0 ! ! — anyone who USUALLY lives here but is now away from home O 0 HOSP. | WORK RD [2-WK.DV Sraveling or INA NOBPISBIT «o.oo vou nmiivssnnnssvwamesys seuss sees 0 N N — anyone else stayinghere? . ............... IR wT Ta PTE 0 0 IC 1 poliNore wa [10 ves palInone Number 20 wh 20] No Number LA JRA IDV TINJ.TCLLTRI HSTCONG I | | | 1 | 1 1 1 1 [LA = TRA "IDV TINJ.TCLLTRI HSTCOND. | | | Io 1 1 1 1 1 [LA TRA "IDV TINJ.TCCLTRI HSTCOND. | | | Io ! 1 1 1 1 1 [LA TRA IDV TINJ.TCLLTRI HSICOND | | | | 1 1 1 1 1 1 [LA ~ TRA IDV TINJ.TCLLTRI HSTCOND L1 [J Under 5 (NP) L1 | Refer to age. Os—-17 (2) [J 18 and over (1) 1a. Did — — EVER serve on active duty in the Armed Forces of the United States? 1a 100 ves 2[0No (2) b. When did — — serve? Vietnam Era (Aug. ‘64 to April ‘75) ........ VN | b. 10vn spun Korean War (June ‘50to Jan. ‘'565) . ........ KW 20kw sos Mark box in descending order of priority. World War Il (Sept. ‘40 to July 47) von vo wwii sOwwa oClok Thus, if person served in Vietnam and in Korea World War | (April “17 to Nov. 18) ........ Wwi «0 mark VN. Post Vietnam (May ‘75 to present) . ....... PVN wwi Other Service (all other periods) ........... 0s c.Was — — EVER an active member of a National Guard or military reserve unit? | ET re Pe ne ites] Oves 20No2 700k 2) d. Was ALL of — — active duty service related to National Guard or military reserve training? | I 10ves 300No sok 2a. What is the highest grade or year of regular school — — has ever attended? 2a. | (0 [] Never attended or kindergarten (NP) Elem: 1 2 3 4 5 6 7 8 High: 9 10 11 12 I College: 1 2 3 4 5 6 + b. Did —— finish the (number in 2a) [grade/yearl? 777] “Bel re ore nis number in 2a) [grade/year] 9s 2 Cine Hand Card R. Ask first alternative for first person; ask second alternative for other persons. = Bas What is the number of the group or groups which represents — — race? 3a. 12 3 4 54 What is — — race? Circle all that apply 1 — Aleut, Eskimo, or American Indian 4 — White 2 — Asian or Pacific Islander 5 — Another group not listed — Specify 3 — Black (Specify) ‘Ask if multiple entries: ~~ TT TT TTT TTT TTT TTT TTT B.] 1 2 3 a 5 b. Which of those groups; that is, (entries in 3a) would you say BEST represents — — race? ¥ (Specify) c. Mark observed race of respondent(s) only. TTT TTT TTT CN 1Ow 20s 30o Hand Card O. 10 Yes 4a. Are any of those groups — — national origin or ancestry? (Where did — — ancestors come from?) 200 No (NP) b. Please give me the number of the group. b. Circle all that apply. 1 — Puerto Rican 5 — Chicano 2 — Cuban 6 — Other Latin American 1 2 2 4 5 6 7 3 — Mexican/Mexicano 7 — Other Spanish 4 — Mexican American FORM HIS-1 (Evaluation) (2-1-90) Page 40 Cod age Clow age Cloud age Cloud age 4. | First name Mid. init. FAge First name Mid. init. fage 1. |First name Mid. init. FAge First name Mid. init. FAge Last name ex Last name ex Last name ex Last name ex Om 10m 10m 10m 20¢ 200F 200F 20°F 2. | Relationship Relationship 2. [Relationship Relationship 3. | Date of birth Date of birth 3. [Date of birth Date of birth \ Month ; Date Year Month J Date | Year Month ' Date Year Month ) Date | Year 1 1 1 1 1 1 1 HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV | ‘HOSP. WORK RD 2-WK. DV c1 00 None 1Owal10 Yes 00] None [oo] None 10wa hives 00 None c1 oo] None 10wa [10 Yes bo None [oo [J None 10Owa [10 Yes ool] None Number 2[dwb 2000 Number Number 20we pONo Number Number [we [200No Number | Number 200we |200No Number C2 LA TRA DV [INJC TCYLTRINS COND. | TA [RA IDV 7INJ.” | CLLTR[HS |COND. CA~ “|RA [DV |INJ. |CLLTRIHS |COND.JCA™ TRA ~| DV INJ. JCLLTR/HS | COND. 1 [| ro | Io | [I | ol | ro 1 | | | 1 Ld LL | Ld I Le | Ll 1 I 1 Ly 1 cca) [LA ~ TRA — BV INU [CUTTRIHS JCOND.|TA~ — jRA™ [OV JINJ. | CLLTR]HE [COND. CA™ ~|RA~ TDV [INJ. jCLLTRHS JCOND|LA ~ TRA —| DV INJT TCCLTR AS | COND. | [I 1 od | [I | 1 | [I 1 [I 1 yl | 1 1 - 1 _ 1 i 1 11 1 Yoel 1 | it Lol 1 1 LA ~ TRA ~|6V INU. [CLTTAHS NL ~ RA” TOV TINJ. | CLITRTHS 5) CA™ TRA TDV TINT. |CLLTRTHS NA ~ TRA 7 DV] NJ. TCCLYR|HS COND. I. | i 3 | ro | [I | ro | ii | [I J ) : ! 1 1 I | 1 l 1 1 ! 1 1 1 1 1 l 1 1 1 1 1 hr ~ TRA TOV INU [CLTTRIHS T[COND. |[TA~ ~ (RA [DV IND.” | CLITRTHS ~ CA™ TRA DV JINJ. |CLLTRIHS [COND.JUA ~ TRA ~| DV INJ. TCCLTR|HS COND. | od 1 ro | Io 1 [I | ol | 1 | i 1 od | Ll 1 | 1 1 1 | 1 bed. 1 feel 1 Ll 1 [I TRA IDV IINJC D.|TA~ ~ (RA [DV TINJ.” | CLLTRIHS COND. CA~ “IRA DV [INJ. |CLLTRIHS [CONDJUA~ T&A ~| DV] INJ. TCLLTR|HS | COND 1 i 1 IE | Nd 1 [I | iA | i | J | [I | I | [| | (I | jd | iJ | [- L1 [J under 5 (vp) [OJ under 5 (NP) L1 [J under 5 (NP) [J under 5 (vp) Os-172 Os-17 2) Os-172 Os-172 [J 18 and over (1) [J 18 and over (1) [J 18 and over (1) [J 18 and over (1) 1a. 10 ves 10 ves 1a. 10 ves 10 Yes 200No 12) 2[0No (2) 200No (2) 200No (2) b. Own sCJ PVN 10wn sCJ PVN b.| Own sO PN 1Oww sO PVN 20kw s(Jos 20kw sJ os 20kw sJ os 20kw s[J os 30 ww 90 Dk s0Owwa 90 bk a0 ww [J bk sd ww oJ DK «Owwi «Owwi sOwwi «Owwi c. c. Oves 20No(2 700k 2 Oves 200No2 700k (2) Oves 2002 700k (2) Oves 200Nor2 700k (2) soi in ct srt fra tn er ten pt oi] inn sm ms rs se [4 tes Sen momen een, mi : 10ves 30No od Dk 10ves 30nNo 90k | 10ves 30nNo ook 10ves 300No o[dbk 2a. 2a. 00] Never attended or 00] Never attended or 00] Never attended or 00] Never attended or kindergarten (NP) kindergarten (NP) kindergarten (NP) kindergarten (NP) Elem: 1 2 3 4 5 6 7 8 Elem: 1 2 3 4 5 6 7 8 Elem: 1 2 3 4 5 6 7 8 Elem: 1 2 3 4 56 7 8 High: 9 10 11 12 High: 9 10 11 12 High: 9 10 11 12 High: 9 10 11 12 College:1 2 3 4 5 6 + College:1 2 3 4 5 6 + College:1 2 3 4 5 6 + College:1 2 3 4 5 6 + b. b. 10ves 200No 10ves 200No 10ves 200No 10ves 20No . 4 1 . 3a 1 2 3 5 2 2 3 4 5 7 3a 1 2 3 4 5 7 1 2 3 4 5 7 (Specify) (Specify) (Specify) (Specify) b. 1 2 3 4 5 z 1 2 3 4 5 2 b. 1 2 3 4 5 rd 1 2 3 4 5 2 (Specify) (Specify) (Specify) (Specify) Tel TT TTT TTT TTT TTT TTT TTT TB ee em FT TT Ow 20s s0o Ow 20s s0o Ow 208 s0o Ow 20s s0o 4a. 10 ves 10 ves 4a. 10ves 10 ves 20 No (NP) 2[0No (NP) 2[0No (NP) 20 No NP) b. b. 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 FORM HIS-1 (Evaluation) (2-1-90) Page 41 a Old age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. fAge one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | If “Yes,” enter i y names in columns 2. [Relationship c. | have listed (read names). Have | missed: Yes | No REFERENCE PERSON — any babies or small children? . . . . .... oven ere e eee a | 0 3. [ppagtoin, West — any lodgers, boarders, or persons you employ who live here? . ............ vam: J 0 ; b — anyone who USUALLY lives here but is now away from home HOSP. | WORK RD [2-WK.DV travelingorinahospital? ......................000nnn We 4 3 BRleiv 0 8 PE J Od bol here ol Nee — anyone else stayinghere? . ......... RR ¥ 3 WI ¥ x ered § 8 EE 0 a C1 10wa [100 Yes Nomper [28 Wb 200 No | Norse d. Do all of the persons you have named usually live here? [J Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP Probe if necessary: RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) Does — — usually live somewhere else? Ask for all persons beginning with column 2: 2. What is — — relati hip to (reference person)? [LA = aE —\ BV" TiND. TCCLIRI FETCONE 3. What is — — date of birth? (Enter date and age and mark sex.) y d ) 1 ! y REFERENCE PERIODS = ( { _ LA TRA IBV TINJ. TCCLTRI HSTCOND. | | | Io 2-WEEK PERIOD : 1 L A bed 12-MONTH DATE LA ~ IRA IDV IINJ.TCLLTRI HSICOND. mm etme sm ems mS ee ef ene emp te mt SS, LE | | | | | 1 1 1 i 1 13-MONTH HOSPITAL DATE [LA = TRA IDV TINJ.TCLLTRI HSTCOND. A2 | Ask CONDITION LISTS 1,2, and 3. bb L. DEMOGRAPHIC BACKGROUND PAGE, Continued oo oJ under 18 (NP) L2 | Referto “Age” and “Wa/Wb’’ boxes in C1. L2| 1D] Webox marked ita 2[J Wb box marked (5a) 3[J Neither box marked (5b) 5a. Earlier you said that — — has a job or business but did not work last week or the week before. Sa. Was — — looking for work or on layoff from a job during those 2 weeks? 100 ves (5c) 200No (6b) b. Earlier you said that — — didn’t have a job or business last week or the week before. | b.| O oes 0 oo Was — — looking for work or on layoff from a job during those 2 weeks? 1] Yes 2LINo (NP) c.Which, looking for work or on layoff fromajob? 77° c.| 100 Looking 6c) 3[JBoth (66) | 2[J Layoff (6b) 6a. Earlier you said that — — worked last week or the week before. Ask 6b. : b. For whom did — — work? Enter name of company, business, organization, or other employer. | 6b. [Employer Clnev pr and mm me nt en em mm Oar (6e) c.For whom did — — work at — — last full-time job or business lasting 2 consecutive weeks or more? 8. Enter name of company, business, organization, or other employer, or mark ‘NEV’ or *’AF’’ box in person’s column. d. What kind of business or industry is this? For example, TV and radio manufacturing, d, [industry retail shoe store, State Labor Department, farm. IAF" in 6b/c, mark “AF” box in person's column without asking. |” o. [occupation ~~" ee eo. What kind of work was — — doing? For example, electrical engineer, stock clerk, typist, farmer. AF (NP) f.What were — — most important activities or duties at that job? For example, types, | g. [putes TT] keeps account books, files, sells cars, operates printing press, finishes concrete. ‘Complete from entries in 6b—F. If not clear, ask: 7 17" Jcmssotworker ~~ TT TTT] 9. Was — — g-| Oe si An employee of a PRIVATE company, business or Self-employed in OWN business, professional individual for wages, salary, or commission ......... P practice, or farm? 20€ ese A FEDERAL government employee? . .............. F Ask: Is the business incorporated? as 70we A STATE governmentemployee? . ................ Ss Yes oo... | aL s[INEV A LOCAL government employee? ................. L No ....oviiiii SE Working WITHOUT PAY in family business OF SAIN? oovvvrrrvrrsenvnnmnnnnsnsinss WP — NEVER WORKED or never worked at a full-time job lasting 2weeksormore . .............. NEV FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 42 IRA IDV |INJ. |CLLTR|HS |COND. | | | | | | 1 1 1 1 1 | | | | | | | 1 li 1 I | oid age [Joid age [Joid age [oid age First name Mid. init. fAge First name Mid. init. §Age . [First name Mid. init. FAge First name Mid. init. fAge Last name ex Last name ex Last name ex Last name ox 10m 10m 10m 10m 2JF 2[JF 2[0F 200F 2. | Relationship Relationship Relationship Relationship 3. | Date of birth Date of birth Date of birth Date of birth Month | Date Year Month J Date | Year onth |Date Year Month |Date | Year 1 1 1 1 1 1 l HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV C1 00 None 10wa 10 ves 00] None oo [J None 10wa [10 Yes 00] None 00] None 10wa [10 Yes 0] None [oo [J None 10wa [10 Yes 00] None Ramber | 2LJWP Romper | amber | 20We 2ONo Number Romper | 20Wb [200No | omer | amber 2lINo | gmper [LA ~ TRA — BV INU. [CLTTR{HS TCOND.|TA~ ~ [RA OV [INJ. | CLITRIHS COND. CA™ ~|RA~ DV TINT. JTLLTATHS [CONG|UA ~ JRA | DV INJ. TCCLTR| HS | COND | 1 | ro | [I | [I | 1 1 1 | Io | Ey 1 1 1 1 1 1 1 1 1 L 1 1 1 1 1 1 fla | — 1 1 | 1 "LA" ~ TRA ~|DV jINJ. [CUTTR|HS COND. |TA ~ [RA [OV JINJ. | CLOTATHS oy CAT 7|RA™ DV JINJ. T|CLLTRIHS NE TRA 7 DV] INJ. TCCLTR HS 7 COND. 1 [| | [I | 1 | ro | IO 1 [I 1 [I | [I 1 l 1 1 1 1 — 1 i 1 1 1 1 1 1 1 1 L 1 1 1 ! 1 1 X f= ~ TRA TOV |INJ. [CLTTR|HS JCOND.|TA~ — {RA [DV TINJ. | CLLTRTHS N CA TRA DV JINJ. |CLLTRTHS [CONDJUA ~ THA —| DV] INJ. TCLLTR|HS | COND 1 ol | Lo 1 ol 1 Lo | od | (I 1 od 1 bl 1 Ll 1 I 1 Ln 1 Lcd 1 feed { | 1 Ll 1 = [LA ~ TRA DV INJ. [CLLTRIHS TCOND.|TA~ ~ [RA [DV TINJ.” | CLLTRIHS [COND. LA™ “IRA DV [INJ. T|CLLTRIHS [COND.JUA~ TRA ~| DV] INJ. TCLLTRIHS | COND 1 1 ou | | 1 1 | | | [I | I | Ld | 1 1 | [I A - ee oe rm a. TT Tee — oJ under 18 (vp) 1 [J Wa box marked (6a) 2 [J Wb box marked (5a) 3 [J Neither box marked (5b) oJ under 18 (vp) 1 J Wa box marked (6a) 2] Wb box marked (5a) 3] Neither box marked (5b) oJ under 18 (NP) 1 LJ Wa box marked (6a) 2 [J Wb box marked (5a) 3] Neither box marked (5b) oJ under 18 (NP) 1 J Wa box marked (6a) 2] Wb box marked (5a) 3] Neither box marked (5b) 5a. 10 Yes (5c) 20 No (6b) 10 Yes (5c) 2 No (6b) 10 Yes (5c) 20 No (6b) 10 ves (5c) 20 No (6b) b.| oT rT TTT TTT | TT TT TT rrr 10ves 200No vp) 10ves 20 No (NP) 10ves 2 No (vp) 10 ves 20 No (vp) c. 100 Looking (6c) 30Both (6b) | 100 Looking (6c) 3[J Both (6b) 100 Looking (6c) 3 Both (6b) 10 Looking (6c) 3 Both (6b) 20 Layoff (6b) 27 Layoff (6b) 2[J Layoff (6b) 2 Layoff (6b) [Employer ~~ ~~ TT = 77 [employer ~~ _ ~~ Employer BG: oo and OINev (6g) OJ Nev (6g) CInev (6g) CInev (6g) Sc. OJ AF (6e) OJ AF (6e) CJ AF (6e) CJ AF (6e) d.[imdusty = TT TT TT TT TTT Tlindusty CC CTT TTT industy 7 7" “industry ~~ ~~ TT TTT 77 o.|Occupaton ~~ _ ~~ [occupation Occupation | Occupation e. CO AF (vp) 0 AF (ve) 0 AF (vp) Oar (ve) §. [Duties ~~ TT TTT TTT TTT [Duties ~~ ~~ ~~ TT TT TTT BE [ Duties [Classof worker ~~ [Classof worker ~~ “|Classof worker ~~ ~~ ~~ TT Class of worker g. 10p si 1Oe 51 10e sli 1Oe si 200F e[JsE 200F ese 20F s[JsE 20F ese 30s 70we 30s 70we 30s 70wep a0s 70we «dL s[INeEv «OL s[INEv «OL s[CINEV «OL s[INev FOOTNOTES FORM HIS-1 (Evaluation) (2-1-90) Page 43 96 J oud age A. HOUSEHOLD COMPOSITION PAGE 1 b. . Do all of the persons you have named usually live here? 1a. What are the names of all persons living or staying here? Start with the name of the person or one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. What are the names of all other persons living or staying here? Enter names in columns. | If “yes, enter names in columns I have listed (read names). Have | missed: Yes No — any babies or smallchildren? . .................c000uneunn. $0.4 5 § TERT 3 O 0 — any lodgers, boarders, or persons you employ who live here? . . . . . Fond ¢ ® a5 Ww © O Od — anyone who USUALLY lives here but is now away from home travelingorinahospital? . .................0.0iiirinenenrnrnnnnannnnn O | — anyone else stayinghere? . ......... ie ® SR % © SIRE W SENATE eee | Od O Yes (2) [J No (APPLY HOUSEHOLD MEMBERSHIP RULES. Delete nonhousehold members Probe if necessary: by an “’X’’ from 1—C2 and enter reason.) Does — — usually live somewhere else? First name Mid. init. JAge Last name ex 10m 2[JF Relationship REFERENCE PERSON Date of birth Month y Date Year nds C1 2. Ask for all persons beginning with column 2: What is — — relationship to (reference person)? 3. What is — — date of birth? (Enter date and age and mark sex.) A1 REFERENCE PERIODS 2-WEEK PERIOD 13-MONTH HOSPITAL DATE A2 ASK CONDITION LISTS 1,2, and 3. 7. 8a. Was the total combined FAMILY income during the past 12 months — that is, yours, (r L.DEMOGRAPHIC BACKGROUND PAGE, Continued Mark box if under 14. If “’“Married’’ refer to household composition and mark accordingly. Is — — now married, widowed, divorced, separated, or has — — never been married? 1 WORK | RD oo] None TOwa [J ves 00] None 20wb [200 No HOSP. 2-WK. DV Number Number 1 J: 1 1 1 1 1INJ. I CLLTRI HSI COND. | | Io 0 [J under 14 1d Married — spouse in HH 2 [J Married — spouse not in HH 3 [J widowed 4 [J Divorced s [J Separated 6 [J Never married L4 Enter person number of spouse or mark box. a. Is - - currently a member of GHA? b. At any time since October 1988, has - - been a member of GHA? i ; Armed Forces members living at home) more or less than $20,000? Include money from jobs, social security, 1 [J $20,000 or more (Hand Card I) retirement income, unemployment payments, public assistance, and so forth. Also include income from 2 [J Less than $20,000 (Hand Card J) interest, dividends, net income from business, farm, or rent, and any other money income received. Read if necessary: Income is important in analyzing the health information we collect. For example, this information helps us to learn whether persons in one income group use certain types of medical care services or have certain conditions more or less often than those in another group. Read parenthetical phrase if Armed Forces member living at home or if necessary. b.| coda 100k 200 u onde nL 22a0v b. Of those income groups, which letter best represents the total combined FAMILY income 0200c 120m 220w during the past 12 months (that is, yours, (read names, including Armed Forces members cade DIN 2s) % living at home)? Include wages, salaries, and other items we just talked about. ving at home ade 140o 240v Read if necessary: Income is important in analyzing the health information we collect. For example, osJF sp 22 this information helps us to learn whether persons in one income group use certain types of ese Da 260 zz medical care services or have certain conditions more or less often than those in another group. i edi 180s os[Ju 1e0T Ra.| ©L under 17 Mark first iate bi 1 [J Present for all questions a. Mark first appropriate box. 2 [J Present for some questions R 3 Not present b. Enter person number of respondent. b. Person number(s) of respondent(s) L3 L3 Enter person number of first parent listed or mark box. Person number of parent 00 [J None in household L4 Person number of spouse 00 [J None in household 1 [J Yes (NP) 2] No (b) FORM HIS-1 (Evaluation) (2-1-90) Page 44 7. Cod 2 First name Mid. init. Relationship Date of birth Month Date Year WORK RD 10wa|10 ves 200wb No HOSP. 2-WK. DV None Number Number |RA | ~ TRE ~|OV INU [CLTTAHS | I | I I ~ TRE ~|OV INST [CLTTRHS | | | I I ~ TRA TOV INU [CLTTRIHS | I IRA | 0 LJ Under 14 1 [J Married — spouse in HH 2 [J Married — spouse not in HH 3 OJ widowed 4 Divorced sO Separated 6 [J Never married 0 0 under 17 1 Od Present for all questions 2 [J Present for some questions 3 Not present Ooi 3 Mid. init. Last name Relationship Date of birth Month J Date | Year HOSP. WORK RD None 10Owa 0 00 20wb No Number Number |RA | | | | | | — RA” OV TINI. "| CLOTRHS I I | I I I ~ RA” TJDV TINJ.T| CLITAIHS | | | | 1 | ~ [RA™ TOV IND.” CLITATHE I IRA IDV [INJ. |CLLTRIHS | | I | | | 0 LI Under 14 1 [J Married — spouse in HH 2 [J Married — spouse not in HH 3 0 widowed 4 [J pivorced s [J Separated 6 [J Never married 0 under 17 1d Present for all questions 2 [J present for some questions 3 Not present Person number(s) of respondent(s) Person number(s) of respondent(s) Person number of parent 00 [J None in household Person number of parent 00 [I None in household Person number of spouse 00 [J None in household 1] Yes (NP) 2] No (6) FORM HIS-1 (Evaluation) (2-1-90) Person number of spouse 00 [J None in household 1 [J Yes (NP) 2 [J No () Page 45 2-WK. DV None c1 00 b. Cod 4 Mid. init. 1 M 200F birth Date Year HOSP. WORK RD DV None 10wa [10 ves 200wb [200 No None Number Number | RA | | | | | | ~ —/RA~ DV JINJ. |CLITAHS | | | | | | | ~ TJRA~ TOV JINJ. T|CLLTRTHS | | | | | | ~ TJRA” TDV |INJ. TiCLLTRIHS | | | | | ! IRA [DV [INJ. | | 0 J under 14 1 [J Married — spouse in HH 2 [J Married — spouse not in HH 3 [J widowed 4 [J Divorced 5 Od Separated 6 [J Never married 0 under 17 1 [J Present for all questions 20 Present for some questions 3 [Not present Cow Mid. init. birth Date | Year "HOSP. RD DV None Owa [10 ves ool] wb [200 No Number Number | RA | | | | | | ~ TRA ~ DV] INJT TCCURAS | | I | I I I ~ TRA 7 DV] INJT TCLLTR{HS | COND. | | | | | | ~ TRA 7 DV INJT TCLLTR|HS | COND. | | | | | | |'RA | DV] INJ. |CLLTRIHS | COND. | | | | | 0 J under 14 1 [J Married — spouse in HH 2 0 Married — spouse not in HH 3 [J widowed 4 [J pivorced s [J] Separated 6 [J Never married 0 under 17 1 [J Present for all questions 2 0 Present for some questions 3 0 Not present Person number(s) of respondent(s) Person number(s) of respondent(s) Person number of parent 00 [J None in household Person number of parent 00 [J None in household Person number of spouse 00 [J None in household 1 [J Yes (NP) 2 [J No (b) 1[J Yes Person number of spouse 00 [J None in household 1 [J Yes (NP) 2 [J No (b) 97 OJ ow age A. HOUSEHOLD COMPOSITION PAGE 1 1a. What are the names of all persons living or staying here? Start with the name of the person or 1. [First name Mid. init. Age one of the persons who owns or rents this home. Enter name in REFERENCE PERSON column. Last name Sex b. What are the names of all other persons living or staying here? Enter names in columns. | i ““Yes," enter ! 5 ¥ names in columns 2. [Relationship 2d c. l have listed (read names). Have | missed: Yes No REFERENCE PERSON — any babies or small children? . . .................... ess pian sam) OF J OF 3. [Qootoim, year — any lodgers, boarders, or persons you employ who live Hove? , 23 ewan re] 1 PJ ; — anyone who USUALLY lives here but is now away from home HOSP. -WK. travelingorinahospital? ...................... » RS, sesesemy LJ 0 i» A. a ne — anyone else stayinghere? . ............... B § SEE § % SWE € ¥ $A £ sam] IJ 0 IC 1 polINone 10 wa [113 Ves [OL None ————2[Jwb 20 No | d. Do all of the persons you have named usually live here? ~~ [] Yes (2) Nome: Li [J No (APPLY HOUSEHOLD MEMBERSHIP ash Probe if necessary: RULES. Delete nonhousehold members by an “’X’’ from 1—C2 and enter reason.) | | _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Does — — usually live somewhere else? LA JRA “IDV TINJ.TCLLTRI HSTCOND. | | | 1 1 1 1 Ask for all persons beginning with column 2: ' + 2. Whatis —— hip to (reference person 0 LL ______ at is relati p to (reference person)? fia — TE ~1BV" TiN. TCCLI HSTEONS. 3. What is — — date of birth? (Enter date and age and mark sex.) i I ! : : ) REFERENCE PERIODS 2-WEEK PERIOD 1 1 : 1 : L A1 13-MONTH HOSPITAL DATE A 2 ASK CONDITION LISTS 1,2, and 3. L. DEMOGRAPHIC BACKGROUND PAGE, Continued Refer to age. Complete a separate column for each nondeleted person aged 18 and over. L5 rto/8g8. Comp p parsonieg L5| PERSON NUMBER Read to respondent(s): In order to determine how health practices and conditions are related to how long people live, we would like to refer to statistical records maintained by the National Center for Health Statistics. Date of birth 5 . i Month Date Year L6 Enter date of birth from question 3 on Household Composition page. L6 2-13 9a. In what State or country was — — born? 9a. 990 okme Print the full name of the State or mark the appropriate box if the State person was not born in the United States. 01 [J puerto Rico ~~ 05 [Jcuba 020 Virgin Islands 08 Mexico 03 [J Guam 98 Clan other 04 [J canada Counties TT TT = TTT TT emer If born in U.S., ask 9b; if born in foreign country, ask 9c. 1 OD iessthan tyr 4 [J 10yrs., less than 15 b. Altogether, how many years has — — lived in (State of present residence)? b. 2 [J 1yr., lessthan 6 5 [J 15 yrs. or more me eer ea ion SAE ro oe re ame ep EE EE HE |_| _ 30 sys.lessthanto ook 15 c. Altogether, how many years has lived in the United States? vl Eaten a (Trove, 1s TRE Cc. 2 0 1yr., lessthan 5 5 Od 15 yrs. or more a0 5yrs., less than 10 9 Cok LL ORM HIS-1 (Evaluation) (2-1-90) Paget oid age [Joid age [Jold age [Joid age 2 3 4 5 1. | First name Mid. init. fAge First name Mid. init. §Age 1. [First name Mid. init. fAge First name Mid. init. fAge Last name ex Last name ex Last name ex Last name Sex 10m 10m 1am 10m 200F 200F 200F 20 F 2. | Relationship Relationship 2. |Relationship |Relationship 3. | Date of birth Date of birth 3. |Date of birth Date of birth Month |Date Year Month Date | Year Month Date Year Month Date | Year | 1 ink 1 | | | HOSP. WORK RD 2-WK. DV HOSP. WORK RD 2-WK. DV HOSP. WORK RD |2-WK. DV| HOSP. WORK RD 2-WK. DV c1 00] None 1Owal[10 ves 00] None [oo [J None 10wa 0 Yes 00] None c1 00] None 10wa [10 Yes [J None [oo (J None 10wa [10 Yes 00] None Number 200wb(2lINo Number Number Uwe LINo Number Number 200we [Ono Number | Number 2[0we |200Ne Number [LA ~ TRA ~|DV |INJT [CLLTR|HS |COND.|TA ~ [RA [DV [INJ.” | CLLTR|HS [COND. CA |RA~ [DV [INJ. |CLLTRIHS |COND.JUA ~ TRA ~ | DV] INJ. JCLLTR|HS | COND. | 1 | [| | [| | [I 1 [I | [I | [I 1 [| 1 1 1 1 1 I} 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 [LA"~ TRA —|OV INJ [CLTTRHS COND. [TA — [RA [OV [INJ.” | CLLTRIHS COND. CA™ T|RA DV |INJ. T|CLLTRIHS Na ~ THAT DV INJT JCCLTR|HS | COND. | ia 1 [I | 1 | 1 | [I 1 [I | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 A 1 1 1 1 1 1 1 1 1 1 1 l [LA ~ TRA —|DV INU [CLTTAHS = TA ~ [RA DV 7INJ.” | CLLTRIHS COND. CA™ T|RA DV JINJ. T|CLLTRTHS CONO.JLA — THA —| DV INJ. TCCLTR|HS 7] COND. 1 ro | [I | 1 | 1 1 od | [I | ( 1 [I 1 1 1 1 1 1 1 msi 1 1 ! 1 1 1 1 1 1 1 l 1 l l 1 1 - ~ TRA T|DVTINJT JTUTTR|HS TCOND. |TA~ ~ (RA [DV TINJ.” | CLLTRATHS N CA” 7)RA™ DV JINJ. T|CLLTRTHS JCONO.JUA — TRA | DV INJ. TCLLTR|HS 7] COND. | od 1 1 1 | 1 i 1 od | V2 | HE | i 1 te] 1 i=l 1 t=4 1 fl 1 ct 1 Led 1 1 1 fd [LA ~ TRA “IDV |INJ. JCLLTR{HS TTCOND.|LA ~ [RA [DV TINJ.” | CLLTRTHS [COND. CA” TIRA™ DV TINJ. TICLLTRIHS [COND.JUA ~ TRA ~ | DV] INJ. [CLLTR|HS | COND. | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | LS LS PERSON NUMBER PERSON NUMBER LS PERSON NUMBER Date of birth Date of birth Date of birth Month Date Year Month Date Year Month Date Year L6 L6 L6 12-13 12-13 es 9a.| 9 Jokmp = 99 [J bk (NP) [i2=79) 9a.| 99 dok mp i232 ee ———————— TS 818] —————————————— SUES State 01 [J Puerto Rico os (cuba 01 [J Puerto Rico 0s [Jcuba 01 [J Puerto Rico 0s [Jcuba 02 Virgin Islands 06 Mexico 02 Virgin Islands 06 Mexico 02 Virgin Islands 06 [Mexico 03 [J Guam 98 (Jan other 03 [J Guam 98 [J All other 03 [J Guam 98 [J All other 04 [J canada countries 04 [J canada Counties. 04 [J canada Coiiries 14 14 1 a h b 1 0 Less than 1 yr. 4 O1oyrs., less than 15 b 1 Od Less than 1 yr. 4 O1oyrs., less than 15 b 1 a Less than 1 yr. 4 O 10 yrs., less than 15 : 201 yr., less than 5 s[1s yrs. or more : 2001 yr., less than 5 s [1s yrs. or more : 201 yr., less than 5 s [1s yrs. or more 3 sys. lessthan10 9 [(JoK 30 sys. lessthan10 9 (lok 30 syrs., lessthan 10 9 (ok [emp rm ae. me ot ees, Ste aime hci a ae Se en on eo es et Ye se a Ye ES ii rs smi simi 16 16 16 Cc. 1 [J Less than 1 yr. +O 10yrs., lessthan 15 | ©. 1 [J Less than 1 yr. +0 10yrs., lessthan15 | g. 1 [J Less than 1 yr. +010 yrs., less than 15 2001 yr., less than 5 s [1s yrs. or more 2001 yr., less than 5 s[J1s yrs. or more 2001 yr., less than 5 sis yrs. or more 30 5 yrs., less than 10 9 Ook ad 5yrs., lessthan10 9 Ook ad 5 yrs., less than 10 9 ok ORM HIS-1 (Evaluation) (2-1-90) Page 47 seyio Ov pesnjed [J € ployasnoH je ya ‘pauresqO 1oN [1] 2 paubis [J 1 pasinbay 1oN [J] 0 JByio pasnjay pIoyssnoH je ya -paureiqQ oN paubig palinbay J0N ldd syio (J v pasnjey [J € PloyasnoH je ya ‘paurerqo oN [J 2 paubis [J 1 paiinbay 1oN [J 0 J8yi0 pasnjay PlOYasnoH je ya7 ‘pautelqQ oN paubig paiinbay 10N seo [Jv pasnjed [J € PIOYyasnoH ie ya ‘paureiqo toN [J 2 paubis [J 1 paiinbay 1oN [] 0 tdd uosiad yoes 10) WIoJ uojssiuiad Jo smejs 18jug ldd S NOSH3d ¥ NOSH3d € NOSH3d ¢ NOSY3d 1 NOSH3d ‘||om se paubis way} eAey 0} 8buelie pue ‘way) Joj SWIof uoISSIwIad Ino [fi ‘pjoyasnoy ul SIequiel HL [eUORIPPE J Juspuodsal 0} wo) uoissiuwied pueH "UOI}08||00 Blep ay} Jaye pakolisap aq (im Ajiley JNoA jo siequiaw JO NoA Ayuspt pinom Jey uonewuojul Aue jeys nok puiwsel | uoissiwiad usHiM INOA paau am ‘sii op 0] sployssnoy ul ajdoad Buimaiaiaiul WO LUBY) Jeyiel SPIOJ8) [BOIPBW WO spew alam Asay} Jl Juslayip aq PINOM SONISIEIS Ueay [BUOIBU UIeLI®d MOY 88s 0} si Apnis siyl jo sesodind ayj jo 8UQ "YHY Fe SpI0dal [edIpall JNOA WO) UOHBWIOUI [RUOIHPPE 8WOS UIeIqo 0} 8XI| PINOM 8M ‘UODB||00 Blep ay} jo Wed sy “Apnis sil Uo Jeisap yum Buijiom si HD ‘Jsiiies pauonusw | SY :SIWHOH4 NOISSIWHId 3DNAOHLNI OL LdIHOS 31S3IDONS 100 Appendix lI Health Interview Evaluation Survey abstracting procedures HEALTH INTERVIEW EVALUATION SURVEY MEDICAL RECORD CODING GUIDELINES General Coding Rules Medical coding for the GHA medical records will utilize the Ninth Revision of the International Classification of Diseases and the Modifications and Special Instructions used for the Health Interview Survey in conjunction with guidelines provided by NCHS. GHA medical records to be coded have been copied in their entirety from October 1988 through the interview date. These records include GHA clinic visits, telephone encounters, referrals to GHA and non-GHA specialists, pathology reports, special procedure reports and hospitalization records. A Medical Record Coding Face Sheet (Exhibit 1) has been prepared for each respondent reporting a medical condition and is attached to the medical record. Before coding the record, verify that the name and ID numbers on the face sheet match those on the medical record. Coding of the record will be done in red pencil on the Medical Record Coding Sheet (Exhibit 2). Enter the batch number from the batch sheet. Enter the Westat ID number, medical record number and the GHA subscriber + family number as they appear on the Medical Record Coding Face Sheet. One encounter section of the form should be completed for each GHA and non-GHA encounter prior to the interview date, including hospital stays. Alternate Coding Method This refers to coding encounters within the two-month reference period only, i.e., two months prior to the interview date, and hospital stays occurring within 19 months of the interview date. 101 102 Exhibit 1 HEALTH INTERVIEW EVALUATION SURVEY MEDICAL RECORD CODING FACE SHEET March 7, 1991 Patient Name: Jane Doe Westat ID+Column Medical record # 430265-01 694723 Date of Interview: 06/12/90 2-month reference period: 04/13/90 19-month reference period: 11/14/88 Number of pages in the medical record: Inventory of attached coding sheets: Condition lists Subscriber+Family 864250-10 Encounter form coding sheets Exhibit 2 943731 Cod-Shifrm April 10, 1991 BATCH | MEDICAL RECORD CODING SHEET HEALTH INTERVIEW EVALUATION SURVEY 0 WESTAT ID SUBSCRIBER + FAMILY MEDICAL RECORD # lI] I_l EACH ENCOUNTER CODED: TOTAL ENCOUNTERS: HA: |_| FORM TYPE: |_|_| REASON: | ENCOUNTER: | HOSPITAL DISCH DATE: |_|_|-| bom) [-19] ENCOUNTER/ADM DATE: | HA: |_| HOSPITAL DISCH DATE: |__|_ |-| FORM TYPE: | I_1-19] REASON: | ENCOUNTER/ADM DATE: | ENCOUNTER: | I-19] _| dd HA: |_| REASON: |_| FORM TYPE: |_|_| ENCOUNTER: |__|_ |_| HOSPITAL DISCH DATE: |_| _|-|__|_|-19]_|_]| 30d ENCOUNTER/ADM DATE: |__|_|-|_|__|-19|_|_| 2 _d_l_l_l PROVIDER ID N | MD Contracept) Last Pap RY a me e DISPOSITION: QReturn in | wy wks, L yrs./or PRN for Q1sor LL Imins. Orel. to/from Patient S) Referral to: Dox 7 OTHER PROBLEMS SEEN: 5 / 0CC. ho7/297% 2x rob. # CLINICAL NOTE: ’ : Prob. # AUUiioiin - ASH Nedo f, [omy fog Cle err A p ~ J (ve, Catt 7D af — Clpaa. aL © Tv x2, DC p= Visco - 1988 ¢ —- 1995 Olesht - (198 2 (lant 40m 1952) 5 [ory on = Chronis hluphcridan) = Fe@ontitle (476 = pep [ Oposcopy (579 =r ONbe (98/7 =Su0 gustacls pn (Schl) s Pr ~ Aero. Hoa fits GS ple - No A 4 °] AS Gen Font 3/ wr stele Ca Be, 4, AY Sor nod SL. a Yorns) ” Csrcty ~ pe pleedls os occ “fr rl ] A fondly I WI rs fom fad lie [eons 105 Group Health Association, Inc., Washington, D.C. 20037 Form No. 2-040024R 106 Encounter This is a serial number identifying the specific encounter. The first section will be 001, the second 002, etc. The number from the last encounter section will be entered in TOTAL ENCOUNTERS when the medical record is completed. Reason This refers to the specific reason for coding the encounter. 1 = Encounter within two months of interview date. Alternate method If the alternate method of coding is used, enter "1" for each encounter, excluding hospital stays, occurring within the two-month reference period. Reference dates are on the Medical Record Coding Face Sheet. 2 = Most recent encounter if none within the two-month reference period. Alternate method. If the alternate method of coding is used and there are no encounters within the two-month reference period, excluding hospital stays, enter "2" for the encounter with the most recent date. For example, if the interview date is 06/07/90, the two-month reference period is 04/08/90 and there are Adult Medicine encounters dated 06/15/90 and 02/21/90, the visit of 02/21/90 would be entered. 3 = Health Assessment. It is important to identify all Health Assessments found on the Adult Medicine form only. This might be identified by "HA" in the Prob. Title/Dx and/or checked under Procedures and Services as "Initial HA new to MD" or as "Health Assessment." Health Assessments of other specialties are not included in this code, e.g., OB/GYN HA. Enter "3" if the record indicates it is an Adult Medicine HA. 4 = Hospital stay within 19 months of interview date. Hospital stay is defined as an overnight stay in a hospital. This must be documented by a discharge summary or other hospital records. If the only reference to a hospital stay is found on the Hosp. Adm/ER/In and Out Surgery form or other GHA encounter forms, code the GHA form only, not a hospital stay. Enter "4" if an overnight hospital stay is present. 5 = Possible overnight hospital stay within 19 months of interview date. This code should only be used if there is documentation of a hospital stay but no discharge date is available and the Hosp. Adm/ER/In and Out Surgery form does not confirm an overnight stay. Enter "5" if a possible overnight hospital stay is present. 6 = No eligible encounter form. Enter "6" if there are no eligible encounter forms in the record, i.e., prior to the interview date, stop coding and enter "001" in "Total Encounter" boxes. 7 = Other. Each encounter coded. This code will identify encounters, excluding Health Assessments, hospital stays and possible hospital stays, when the alternate coding method is not used. Enter "7" for all other encounters prior to the interview date ignoring the two- month reference period. Form e ADULT MEDICINE... iiiiieeeeeeeiereereeeeseeseessessesseseessessessessessessessesssessessessessensessons 01 ADVICE /PRESCRIPTION ......onmmmmmisssnsesmismvmissrtimstasssesiomtirmss sme sissies sssoeinn 02 ALLERGY /IMMUNOLOGY .....cvviririeinienriniensasisssisessssisesessssssesssssssssssssssssssssssesssssans 03 ANTICOAGULATION THERAPY ......ooteeeeeeteeeereereeteeseesessessessessessessessessssssessessesens 04 CONSULTATION/REFERRAL GHA (IN-HOUSE).........ccoscssnsnsmsmssnsmsssessussssasssesens 05 CONTACT LENS FORM... mara smasesemreriaersrenss 06 CONTINUING CARE... teeeeteeteetesteetestertestesesaessesse ess essessess ess essessassesssssessesens 07 DERMATOLOGY iciersonimmmiiommessseoiiamasresisminismnsissssnsisssssasssseisss srnseisi ere suse sh Enis sn sis baanes 08 EAR, NOSE AND THROAT .......eeeeeceeeeteeeereeeeseeseessetessessesseeseessensensessensessensenes 09 BYE CARE... rie rerrarmnenronivemenminsesserspes chem ieee teem mie eit iii hams nme nme seansrimer nesses ss datins 10 HOSPITAL ADM/ER/IN AND OUT SURGERY .......ccceoeuniininirinrnienicrnecrcnians il MINOR INJURY UNIT {MIU ).conisnrsmmmmmmsnmmumsmmsmmmsnmsmmmamssmamn msm 12 NEUROLOGY .o....onminsiiinsiinsivimmmimisimsmimisoess isso sesame anes sis sisson sosnoenessvuveiisosi 5553s 13 NUTRITION......cicummmmomimmmirsiomsrsmomarssicsmsoemsmurensarssrssssnnsosonsssesseesss eestsscs swarms essns 14 108 OBSTETRICAL/GYNBOUOLOGY ...cuvmrmmimmmmummmsanminssmmmmsmn momen 15 A EL RS ———— 16 PATIENT REFERRAL TO CONSULTING SPECIALIST ...cccuminmmmusnismue 17 (OUTSIDE, NON-GHA) PEDLATRIOS ....ovin vevuinssusns isms sermnsismosssnsosmsmases ns sams sagms sass iss i dasa sg sass st aps sass aay 18 PHYSICAL THERAPY ...cccsnmemmmmismimismimmumsmsssssmss ss sess sis sm sss san ses 19 PODIATRY ..ccrcrrisususneimermnesesasmeenineversssons revert rosssns oman is ne sof is soot a a i sash iss sha is basa sasisa sass 20 PRIMARY PREVENTION PROGRAM ........ccumumnmumnnmmmsmemuion sso siirsaistasssors 21 BADIOLOIY vv vovensnsesusnsnsiannssansnmasssisismsusnss seas sas ss sms ass ass ipa sh bei vm aa ts sais sens me assess stoamsos ss 22 SOCIAL SERVICE ...oroumsurirmsmsnmismesssmmsmsnsnmsus sesssnsadus ss sssasasss sata te tsa mass sass ossaveons 23 SURGERY /couiusuucrorernsmeorssusususnorerssns oo sosasihs uss has isa seins sea a sams a3 3 a 5 55 $0 RR RASTS 24 SURGICAL POSTING PATIENT PROFILE ....cunmnnnunmmannmunmmommimmnimm 25 TELE ENCOUNTER/ADVICE/RX REFILL - EYE... 26 TIROLOGY crrvnrren cxnerersnsnsmvsesnsnsstsgsssissssasssams sss ss mssna os os oss SE as Hs ERAS 8 27 OTHER NONAGTIA 1.oiinimmm vs sme snsinsnsnsis sans sonata naman i ss a5 p 648500 55st sss mas sss sess 28 This item refers to the GHA form in the medical record documenting the reason for the visit/encounter. Forms are labeled on the side or at the top. Typed GHA clinical notes should be matched by date to the appropriate form for coding purposes. Clinical notes are not entered as a separate encounter. Enter the correct code from the above list. Zero fill the lead box if needed. Reports from outside specialists should be coded "other non-GHA," "28." If other GHA forms not listed here are encountered, complete a problem sheet so the next available number can be assigned. Forms to be excluded are laboratory/pathology reports, consent forms, return-to- work forms, and encounter forms marked "NS" (no show). Health Assessment (HA) This item is important if the alternate coding method is used and serves to identify a Health Assessment when it is an encounter within the two-month reference period or is the most recent encounter prior to the two-month reference period. This item must be completed for all encounters regardless of the coding method used. Enter "0" (No) if the encounter is not a Health Assessment. Enter "1" (Yes) if the encounter is a Health Assessment. Encounter/Admission Date Enter the month, day and year of the date of the encounter or hospital admission. This date cannot be later than the interview date. If the alternate method is used, this date must be within the two-month reference period unless it is a Health Assessment, a hospital stay or most recent encounter if none within the two-month reference period. For dates which are missing or illegible, assume the records are in chronologic order and use the preceding and subsequent forms in an effort to establish a date. If this is not successful, enter "99" for the missing parts of the date. Fill the leading box with a zero as needed. Hospital Discharge Date This item is completed when records of an overnight hospital stay are available. Enter "99" for missing parts of the date. Zero fill the leading box as needed. If the encounter is not a hospital stay, leave the item blank. Provider ID Number Enter the provider ID numbers in the order of appearance on the GHA form. For example, if Providers 1 and 2 are blank on the form and 3 is 775, enter 0775 in the third set of coding boxes. The first two sets of coding boxes will be blank. Zero fill lead boxes as needed. If there is no Provider number on the GHA form, enter "9999" in the first set of boxes. For non- GHA encounters, leave the boxes blank. Diagnosis The diagnosis will usually be found on the GHA encounter form in the Prob. Title/Dx section. However, it will be necessary to skim the clinical notes for clarification of a diagnosis or to capture additional diagnoses, entering the primary reason(s) for the visit first. Enter the 109 110 diagnostic verbiage in the boxes using only one line for each diagnosis to be coded. Use abbreviations to conserve space and time. For operations occurring within one year of the interview date, a diagnosis or condition should be entered. For hospitalizations, the diagnoses should be on the discharge summary. The contents of the summary should be reviewed for additional diagnoses. Some encounters will not have a diagnosis or condition mentioned. Refer to the section on Code for recording a diagnosis for these encounters. Code Select the appropriate code for the diagnosis by consulting the special instructions used for the Health Interview Survey as well as Vol. 1 and 2 of the Ninth Revision of ICD. Enter the four-digit code in the boxes provided. For diagnoses not requiring a fourth digit, enter "+" in the last box. There should be no blank boxes. Some encounters will not have a diagnosis, e.g., routine examination on a healthy person or a telephone call requesting a prescription refill. For these encounters, use one of the following codes: NCO.1 = General checkup or examination NCO.2 = Tests only NCO.3 = Immunization only NCO.4 = Other (specify the reason) Enter the verbiage in the diagnosis boxes. Problem Sheets The medical coder should complete a Problem Sheet (Exhibit 4) when there is a question regarding the medical record forms or diagnostic codes. 111 112 History (Hx) The intent of this item is to capture significant medical conditions which were present at some time in the past but have been treated and may not be present at the time of the current encounter, e.g., a respondent has a history of prostatectomy due to cancer of the prostate. HIS rules do not permit the use of history, "V," codes so the diagnostic code will be flagged to indicate a "history of" condition. For operations more than one year prior to the interview date and the cause is stated, enter the diagnosis, code and indicate this is a history of the condition. 0 = No. Enter "0" if the diagnosis is still present or is an operation within one year of the interview date. 1 = Yes. Enter "1" if the diagnosis is stated as a history of the condition and is no longer present. Record Overflow If more than one coding sheet is required, continue coding on as many sheets as necessary. For continuation sheets, remember to complete the Batch, Westat ID, Medical Record and Subscriber + Family numbers. Enter "+ + +" in the boxes for Total Encounters and "+" in the box for Each Encounter Coded. For encounters having more than five diagnoses, enter the overflow in the next encounter secton. The total number of diagnoses is entered in the original section. It is not necessary to repeat any of the encounter identification information. Draw a line through the blank boxes from "ENCOUNTER" through "NO. DX." Overflow diagnoses for the Alternate Coding Method will be entered on a supplemental coding form. When the entire medical record has been coded, visually edit your work, making sure encounter numbers are sequenced correctly and all boxes requiring an entry have been completed. Enter the total number of encounters in the first section of page one and staple the forms in the upper left-hand corner. Exhibit 4 HEALTH INTERVIEW EVALUATION SURVEY 943731 GHA MEDICAL RECORD EFFORT PROBLEM SHEET WESTAT ID#: |__ |_ |__|__|__I1__I-1__I_1I DIAGNOSIS/CONDITION #: SENT FROM: MEDICAL RECORD #: |__|_ |__| __|__ |_| ENCOUNTER/CONDITION DATE: DATE: PROBLEM: SOLUTION: DECISION BY: DATE: 113 Appendix lll Loose match recommendations It was decided to exclude conditions on the medical record for which the “History” indicator was flagged, except those conditions on the “Ever” list (condition list 2). On this list are: Hardening of the arteries or arterio- sclerosis; congenital heart disease; coronary heart disease; hypertension/high blood pressure; angina pectoris; myocar- dial infarction; and any other heart attack. That is, for these conditions on the medical record, those for which a history was indicated will be kept. Arthritis Add: 274.0 Gouty arthropathy 274.1 Gouty nephropathy 274.8 Gout with other manifestations 710.2 Sjogrens Disease 717.7 Chondromalacia, knee 720.9 Unspecified inflammatory spondylopathy 720.2 Sacroiliitis, not elsewhere classified 722.9 Disc disorder,*Recode C 105 722.4 Lumbosacral/cervical degeneration, *Recode C 105 722.5 SAME 723.4 Cervical radiculopathy 724.3 Sciatica, *Recode C 104 724.4 Neuritis/radiculitis Rheumatism This category was not considered because of the low incidence (N = 1). Dermatitis Add: 039.0 Actinomycotic infections, cutaneous 110.4 Dermatophytosis of foot, athlete’s foot 110.0 Of scalp and beard 110.1 Of nail 110.2 Of hand 110.3 Of groin and perianal area 110.5 Of the body 110.8 Of other sites 110.9 Of unspecified site 114 111.0 Pityriasis versicolor (tinea) 111.9 Dermatomycosis, unspecified (BARN DOOR?) 111.8 Dermatomycosis, other (BARN DOOR?) 373.0 Blepharitis 373.3 Noninfectious dermatoses of eyelid 373.1 Hordeolum and other deep inflammation of eyelid 373.2 Chalazion 373.9 Unspecified inflammation of eyelid 682.9 Cellulitis and abscess, unspecified site 682.0 Face 682.2 Trunk 682.3 Upper arm and forearm 682.4 Hand, except fingers 682.5 Buttock 682.6 Legs, except foot 682.7 Foot, except toes 686.9 Unspecified local infection of skin and subcutaneous tissue 686.1 Pyogenic granuloma 686.8 Other local infections of skin and subcutaneous tissue 696.1 Other psoriasis, *Recode C 112 696.3 Pityriasis rosea 696.5 Other and unspecified pityriasis 707.9 Chronic ulcer of skin, unspecified site 707.0 Decubitus ulcer 707.1 Ulcer of lower limbs 707.8 Chronic ulcer of other specified sites 782.1 Rash and other nonspecific skin eruption 782.2 Localized superficial swelling, mass or lump 782.7 Spontaneous ecchymoses 782.8 Changes in skin texture Impairments There is no loose match for impairments, because the loose match is essentially a critique of Recode C, and matching conditions to impairments does not make sense in that context. Tinnitus There are no recommendations for a loose match. Cataracts There are no recommendations for a loose match. Constipation There are no recommendations for a loose match. Diabetes There are no recommendations for a loose match. Migraine No recommendations for a loose match were made. Heart conditions (ischemic, tachycardia, heart murmurs, other and unspecified rhythm disorders, congenital heart disease, other selected diseases of heart) There is no loose match, but heart conditions are aggre- gated as they appear in the NHIS prevalence reports. Hardening of the arteries 413 Angina pectoris 414 Other forms of chronic ischemic heart disease 437.0 Cerebral atherosclerosis 443.9 Other peripheral vascular disease, unspecified (usually claudication) Varicose veins of lower extremities There are no recommendations for a loose match. Hemorrhoids There are no recommendations for a loose match. Hypertension There are no recommendations for a loose match. Chronic bronchitis Chronic obstructive pulmonary disease, 496, is added to the 601 group. Asthma There are no recommendations for a loose match. Allergic rhinitis and chronic sinusitis There is a general upper respiratory category that includes: Recode C 603, Allergic rhinitis Recode C 605, Chronic sinusitis 472.0, Chronic rhinitis 472.2, Chronic Nasopharyngitis 115 Appendix IV Definitions of terms used in this report AB design — Study design for survey validity check in which population survey is conducted, then records are checked for characteristics elicited from survey. AC design — Study design for survey validity check in which cases containing characteristics of interest are se- lected from medical records, then interviews are con- ducted with those people and data compared; also called a “reverse record check.” Accuracy — Tendency of test measurement to center around the true value. Bias — Persistent or systematic error. Condition-level prevalence! —The number of different conditions within a National Health Interview Survey (NHIS) recode group per 1,000 population, as reported in a survey. More than one condition in the NHIS recode group may be counted per survey participant. Criterion validity—Measure of correctness of survey responses compared with true values. False negative — Failure of the survey to report a con- dition mentioned in the medical record, assuming the medical record to be true. False positive — A survey report not confirmed by the medical record, assuming the medical record to be true. Field bias — Systematic error arising from the differ- ence between the information derived from survey respon- dents and that from verification sources. Full design — Study design for survey validity check in which population is sampled independently of character- istic of interest, and survey and record information are obtained and compared for each sampled element. Household member '—A person living in the same household as a list-sample person, for whom data were collected in the Health Interview Evaluation Survey (HIES) interview and from Group Health Association (GHA) medical records. Kappa statistic — A statistic measuring agreement be- tween two sources of classification of the same phenome- non; the Kappa statistic is superior to “percent agreement” because the former takes into account the likelihood of chance agreement. List-sample person! —A person selected from GHA records to participate in the HIES. Net overreport ' —The net difference between preva- lence derived from two sources; specifically, the rate derived from the HIES interview minus the rate derived from GHA records. NHIS recode group — Groups of chronic conditions aggregated from codes assigned according to the NHIS 116 modifications to the International Classification of Dis- eases, Ninth Revision, Clinical Modification; Recode C is the aggregation used for producing prevalence estimates of chronic conditions from the NHIS. Nonresponse — The failure of a unit or units to respond to a survey entirely (unit nonresponse) or to particular items on a survey (item nonresponse). Nonsampling error — Difference between a survey esti- mate and the true value not due to sample design; in- cludes response, processing, and interpretation errors. Percent overreport ' — The relative difference between prevalence derived from two sources; specifically, the rate derived from the HIES interview divided by the rate derived from GHA records. Person-level prevalence ' — The number of persons per 1,000 population having one or more conditions in a particular NHIS recode group, as reported in a survey. Reliability — Tendency of repeated measurements on the same sample to yield the same result, providing consistent answers in comparable situations and without random errors. Response error— Errors, not due to sampling, intro- duced during the course of data collection because of such things as interviewing, enumerating, and counting or mea- suring problems. Sensitivity — True positive rate or proportion of cases known to be positive (confirmed by medical record), for which a positive household response is obtained. Specificity— True negative rate or proportion of cases known to be negative (absent from medical record), for which negative household responses are obtained. Type A match '—Match of positive response by house- hold interview and medical record, a “positive match.” Type B mismatch '— Mismatch caused by positive house- hold response on a specific item and negative or no medical record notation for the same item, an apparent “false positive.” Type C mismatch '—Mismatch caused by negative or no household response to a specific condition and a positive medical record notation for the same condition, an apparent “false negative.” Type D match ' — Match of negative response by house- hold interview for specific item and no medical record notation for that item, a “negative match.” Validity — Tendency of responses to a survey question to correspond to what the question is intended to measure. Term defined specifically for this study. “U.S. Government Printing Office: 1994 — 301-019/80026 Vital and Health Statistics series descriptions SERIES 1. SERIES 2. SERIES 3. SERIES 4. SERIES 5. SERIES 6. SERIES 10. SERIES 11. SERIES 12. SERIES 13. Programs and Collection Procedures — These reports describe the data collection programs of the National Center for Health Statistics. They include descriptions of the methods used to collect and process the data, definitions, and other material necessary for understanding the data. Data Evaluation and Methods Research — These reports are studies of new statistical methods and include analytical techniques, objective evaluations of reliability of collected data, and contributions to statistical theory. These studies also include experimental tests of new survey methods and comparisons of U.S. methodology with those of other countries. Analytical and Epidemiological Studies — These reports present analytical or interpretive studies based on vital and health statistics. These reports carry the analyses further than the expository types of reports in the other series. Documents and Committee Reports — These are final reports of major committees concerned with vital and health statistics and documents such as recommended model vital registration laws and revised birth and death certificates. International Vital and Health Statistics Reports — These reports are analytical or descriptive reports that compare U.S. vital and health statistics with those of other countries or present other international data of relevance to the health statistics system of the United States. Cognition and Survey Measurement — These reports are from the National Laboratory for Collaborative Research in Cognition and Survey Measurement. They use methods of cognitive science to design, evaluate, and test survey instruments. Data From the National Health Interview Survey — These reports contain statistics on illness; unintentional injuries; disability; use of hospital, medical, and other health services; and a wide range of special current health topics covering many aspects of health behaviors, health status, and health care utilization. They are based on data collected in a continuing national household interview survey. Data From the National Health Examination Survey, the National Health and Nutrition Examination Surveys, and the Hispanic Health and Nutrition Examination Survey — Data from direct examination, testing, and measurement on representative samples of the civilian noninstitutionalized population provide the basis for (1) medically defined total prevalence of specific diseases or conditions in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics, and (2) analyses of trends and relationships among various measurements and between survey periods. Data From the Institutionalized Population Surveys — Discontinued in 1975. Reports from these surveys are included in Series 13. Data From the National Health Care Survey — These reports contain statistics on health resources and the public's use of health care resources including ambulatory, hospital, and long-term care services based on data collected directly from health care providers and provider records. SERIES 14. SERIES 15. SERIES 16. SERIES 20. SERIES 21. SERIES 22. SERIES 23. SERIES 24. Data on Health Resources: Manpower and Facilities — Discontinued in 1990. Reports on the numbers, geographic distribution, and characteristics of health resources are now included in Series 13. Data From Special Surveys —These reports contain statistics on health and health-related topics collected in special surveys that are not part of the continuing data systems of the National Center for Health Statistics. Compilations of Advance Data From Vital and Health Statistics — Advance Data Reports provide early release of information from the National Center for Health Statistics’ health and demographic surveys. They are compiled in the order in which they are published. Some of these releases may be followed by detailed reports in Series 10-13. Data on Mortality — These reports contain statistics on mortality that are not included in regular, annual, or monthly reports. Special analyses by cause of death, age, other demographic variables, and geographic and trend analyses are included. Data on Natality, Marriage, and Divorce — These reports contain statistics on natality, marriage, and divorce that are not included in regular, annual, or monthly reports. Special analyses by health and demographic variables and geographic and trend analyses are included. Data From the National Mortality and Natality Surveys — Discontinued in 1975. Reports from these sample surveys, based on vital records, are now published in Series 20 or 21. Data From the National Survey of Family Growth —These reports contain statistics on factors that affect birth rates, including contraception, infertility, cohabitation, marriage, divorce, and remarriage; adoption; use of medical care for family planning and infertility; and related maternal and infant health topics. These statistics are based on national surveys of childbearing age. Compilations of Data on Natality, Mortality, Marriage, Divorce, and Induced Terminations of Pregnancy — These include advance reports of births, deaths, marriages, and divorces based on final data from the National Vital Statistics System that were published as supplements to the Monthly Vital Statistics Report (MVSR). These reports provide highlights and summaries of detailed data subsequently published in Vital Statistics of the United States. Other supplements to the MVSR published here provide selected findings based on final data from the National Vital Statistics System and may be followed by detailed reports in Series 20 or 21. For answers to questions about this report or for a list of reports pubiished in these series, contact: Data Dissemination Branch National Center for Health Statistics Centers for Disease Control and Prevention Public Health Service 6525 Belcrest Road, Room 1064 Hyattsville, MD 20782 (301) 436-8500 DEPARTMENT OF PRESORTED HEALTH & HUMAN SERVICES SPECIAL FOURTH-CLASS RATE POSTAGE & FEES PAID Public Health Service PHS/NCHS Centers for Disease Control and Prevention PERMIT NO. G-281 National Center for Health Statistics 6525 Belcrest Road Hyattsville, Maryland 20782 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 DHHS Publication No. (PHS) 94-1394, Series 2, No. 120 u. C. BERKELEY LIBRARIES Co4994kL7a%9 mera ie Ea WHEY oo . X SER