A, 408 A3L18 1990 c. 2 PUBL National Medical Expenditure Survey Mental Health and Functional Status of Residents of Nursing and Personal Care Homes Research Findings 7 Agency for Health Care Policy and Research 3“" ,, Department of Health & Human Services C Public Health Service 0 “~- D:Posx'ronv OCT 3 1 i990 PUBLIC HEALTH LIBRARY /’srnuir i t | \\ LIBRARY 1 UNIVERSiTY Of cmrcanm/ The following is the recommended biblio- graphic citation for this publication: Lair, T. and D. Lefkowitz. (1990, Septem- ber). Mental health and functional status of residents of nursing and personal care homes (DHHS Publication No. (PHS) 90—3470). National Medical Expenditure Survey Research Findings 7, Agency for Health Car Policy and Research. Rock- ville, MD: Public Health Service. Abstract The Institutional Population Component ofthe 1987 National Medical Expenditure Survey includes a nationally representa— tive sample of current residents of and new admissions to nursing and personal care homes. This report uses the current resi- dent sample to describe the mental health and functional status of people living in nursing and personal care homes in the United States on January 1, 1987. Nation- al estimates of resident characteristics in— cluding age, sex, marital status, and race are presented. This report also provides estimates of cognitive impairment, mental disorders, emotional/behavioral prob- lems, psychiatric symptoms, and the ex- tent and nature of functional limitations in this population. Finally, selected meas- ures of functional and mental status of the resident population are considered as they relate to various facility characteristics such as certification, bed size, and facility type. September 1990 DHHS Publication No. (PHS) 90-3470 Background The 1987 National Medical Ex- penditure Survey (NMES) continues a series of national health care ex— penditure surveys, most recently the 1980 National Medical Care Utiliza— tion and Expenditure Survey and the 1977 National Medical Care Expendi- ture Survey. Like these earlier sur- veys, NMES uses a national probabili- ty sample of the civilian, noninstitu— tionalized population in a Household Survey. NMES includes as well the population resident in or admitted to ntirsing homes and facilities for the mentally retarded over the course of the survey year. To provide focused estimates of the provision of long-term care rendered in the community, the Household Sur- vey oversampled population seg- ments known to be at risk of needing or using services for chronic impair- ments or long-term illness, particular- ly the elderly and those with difficul- ties in performing activities of daily living. Because of continuing policy concern with populations having re- stricted access to the health care deliv- ery system, oversampling was di- rected also at poor and low-income families and the black and Hispanic minorities. In addition, the American Indian and Alaska Native population living on or near reservations and eli— gible for services from the Indian Health Service was included in NMES on the basis of a separate household sample. Together, the major components of the National Medical Expenditure Survey provide measures of health status and estimates of insurance cov- erage and the use of services, expendi— tures, and sources of payment for the period from January 1 to December 31, l987,forthe civilian population of the United States. The reports of health care expenditures and insur— ance coverage obtained in the house- hold surveys are being verified and supplemented by additional surveys. Most important among these are the Health Insurance Plans Survey of em- ployers and insurers of consenting Household Survey respondents and the Medical Provider Survey of physi- cians, including osteopathic physi— cians, and inpatient and outpatient fa- cilities reported as providing services to consenting members of the nonin- stitutionalized population sample. The Medicare Records Component will provide a record check on 1987 eligibility status and claims informa— tion of all sampled Medicare benefi- ciaries, including those in the institu— tional population. Household Survey The Household Survey was fielded over four rounds of personal and tele- phone interviews at 4-month inter— vals, with a short telephone interview constituting the final fifth round. The reference period was calendar year 1987. Baseline data on household composition and employment and in- surance characteristics were updated each quarter, and information on all use of and expenditures for health care services and sources of payment was obtained. A long—terrn care supple- ment permits the development of esti- mates of persons with functional dis— abilities and impairments and their use of formal home and community services, as well as the burden of pro- viding care as expressed by family and friends. Survey of American Indians and Alaska Natives Conducted with the same data col- lection instruments and interview pro- cedures over the same period, the Sur- vey of American Indians and Alaska Natives provides a basis for compar- ing the use of health services by the American Indian and Alaska Native population eligible for care through the Indian Health Service. Special at- tention was paid to measures of access to care, sources of payment for serv— ices other than those provided by the Indian Health Service, and the use of traditional medical care. Institutional Population Component The survey of persons resident in or admitted to long—term care facilities (nursing homes and facilities for the mentally retarded) at any time in 1987 was conducted to provide data on health care use and expenditures both within the institution and in the com- munity forthe survey year. This infor- mation and extensive data on health and functional status, demographic characteristics, and living arrange— ments and insurance coverage at the time of admission were obtained from two primary sources. The Survey in Institutions collected data from facil- ity administrators and designated staff; a Survey of Next of Kin col- lected data from the respondent’s next of kin or other knowledgeable person in the community to obtain additional personal history and related informa- tion. A11 survey components were de- signed to provide statistically un- biased national estimates that are rep- resentative of the civilian population of the United States in 1987. The Household Survey sample can be characterized as a stratified multi- stage area probability design with ato— tal sample of 36,400 individuals in roughly 15,000 households. Over- sampling of the population subgroups of interest was based on a separate screening interview conducted in the fall of 1986 with a sample of approxi- mately 35,000 addresses. The institu- tional population sample was based on a three—stage probability design. Facilities were selected in the first two stages. The final stage sampled resi- dents as of January 1, 1987, and ob— tained a sample of admissions be— tween January 1 and December 31, 1987. Sampling specifications re- quired the selection of a total of 1,500 facilities, 7,000 current residents, and 3,500 new admissions to either nurs— ing homes or facilities for the mental— ly retarded. The sample frame for fa- cilities in the Institutional Population Component was derived from the 1986 Inventory of Long-Term Care Places. Taken in conjunction, these surveys yield comprehensive, population- based estimates that will permit sepa- rate and comparative analyses of most population groups of policy interest, including those presently outside the scope of the various public and private financing mechanisms. In contrast to information limited to program or provider statistics, the National Medi- cal Expenditure Survey will permit comprehensive analyses of data on all public and private sources of coverage for health care services and on out-of- pocket payments by individuals and families in the US. population. The data base will also enable assessment of the implications of recent and pro- jected changes in public and private health care benefits; in methods of fi- nancing both health care and insur— ance coverage; and in various public and private subsidies, income tax ex- emptions, and employee compensa— tion arrangements. The 1987 National Medical Expen- diture Survey was guided by the infor- mation requirements of the Center for General Health Services Intramural Research, Agency for Health Care Policy and Research. The Survey of American Indians and Alaska Natives was cosponsored by the Indian Health Service. The Health Care Financing Administration, the National Center for Health Statistics, the National In- stitute of Mental Health, and the Of- fice of the Assistant Secretary for Planning and Evaluation in the De- partment of Health and Human Serv- ices provided consultation and techni- cal assistance during the development of the survey instruments. Field work was conducted by Westat, Inc., Rock- ville, MD, as the primary contractor and by NORC, University of Chicago; the Council of Energy Resource Tribes, Denver, CO; and Stephen R. Braund and Associates, Anchorage, AK. Data processing during the analy— sis stage is being provided by Social and Scientific Systems, Inc., Bethes- da, MD. The data were collected and are be— ing edited and published in accord- ance with the confidentiality provi— sions of the Public Health Service Act and the Privacy Act. A series of public use tapes is being released to ensure timely access to these data by the re— search and policy community. Additional information on the Na- tional Medical Expenditure Survey is available from Daniel C. Walden, Di— rector, Division of Medical Expendi- ture Studies; Steven B. Cohen, Direc- tor, Division of Statistics and Re- search Methodology; Pamela Farley Short, Senior Research Manager; and Renate Wilson, Project Editor; Center for General Health Services Intramu- ral Research, Agency for Health Care Policy and Research; Room 18—A—55, Parklawn Building; 5600 Fishers Lane; Rockville, MD 20857 (301/443—4836). (SW 4/35 as” 1//< Ma 73.??? Mental Health and Functional Status of Residents of Nursing and Personal Care Homes Tamra J. Lair and Doris Cadigan Lefkowitz The decline in the birth rate and an increase in life ex- pectancy have changed the age composition of the United States, raising both the median age of the popula— tion and the percent of the population over age 65. In fact, the most rapidly growing population segment is over age 85. This age group, which accounted for 8.8 percent of the US. population in 1980, is expected to account for 14.6 percent of the population by the year 2000. Further, while only 5 percent of persons over age 65 reside in nursing homes, 22 percent of those over age 85 do so (Hing, 1987). In 1987, expenditures for nursing home care amounted to $40.6 billion, nearly evenly split between public and private sources (Letsch, Levit, and Waldo, 1988). The high and increasing costs of providing long—term institutional care, coupled with changes in the age composition of the population, make it important to understand the characteristics of persons who use this type of care. This report provides estimates of the number of per- sons in nursing and personal care homes on January 1, 1987, and information on their mental and functional health status using data from the Institutional Popula— tion Component (IPC) of the 1987 National Medical Ex- penditure Survey (NMES). The IPC is composed of representative national samples of nursing and personal care homes and facilities for the mentally retarded. The estimates in this report are based on findings from phase 1 of the nursing home sample of the IPC. (See Cunning- ham and Mueller, in press, for initial estimates relating to the population in facilities for the mentally retarded.) The NMES IPC differs in important ways from pre- vious national surveys of institutionalized populations, including the National Nursing Home Surveys of 1977 (Zappolo, 1981) and 1985 (Hing and others, 1989) and the 1976 Survey of Institutionalized Persons (Brown and Stoudt, 1976). It is the first national survey to ob- tain full—year information on use and expenditures for persons in nursing homes. The IPC also collected data on the use of and expenditures for long-term care serv- ices provided in the community to persons living in a nursing home for only part of the year. The nursing home component of the NMES IPC is composed of two surveys: one, of persons residing in nursing homes on January 1, 1987, and a second, of persons admitted to nursing homes during 1987. The capacity to derive separate estimates for current resi- dents, all-year residents, and new admissions is an im— portant analytic advantage. Samples of current resi- dents, while providing an accurate picture of a “day of care,” are more highly representative of long-stay resi- dents. New admission or discharge samples give a more accurate picture of those admitted to a nursing home, about a third of whom leave within 3 months (Keeler, Kane, and Solomon, 1981; Liu and Manton, 1983). The data used in this report are restricted to the current resident survey of the NMES IPC and therefore should be considered to be a description of residents of nursing homes on a given day—in this case, January 1, 1987. National estimates are presented for selected demo- graphic characteristics of persons who resided in nurs- ing and personal care homes in the United States on J an- uary 1, 1987. Their distribution by time since admission is shown. The extent of functional limitations, behav- ioral problems, and mental disorders is also described. In addition, the functional limitations and mental disor- ders of residents are presented with respect to a variety of facility characteristics. The measures of functional status and mental health attempt to capture the range of health problems experi- enced in this population. They are current and widely used measures (Ernst and Ernst, 1984; Mulken and Manderscheid, 1989) and facilitate comparisons with other national studies. For example, the nursing home population is characterized both in terms of psychiatric symptoms and mental disorders in order to capture the wide range of potential behaviors and diagnoses exhib- ited in nursing homes. Likewise, functional status is as- sessed in terms of activities of daily living (ADLs), both in sum and by individual activities, so as to provide a detailed picture of functional difficulties experienced by this institutional population. (See the technical appen- 43' dix for information on the derivation of all estimates. The technical appendix also describes sampling infor- mation and standard error estimates that must be consid- ered in assessing the confidence level of the national es- timates.) Future reports from the NMES will examine socioeco- nomic and financial characteristics of nursing home res— idents and their families and will present estimates of health care expenditures and utilization of services by both current residents of and new admissions to nursing homes. Demographic Characteristics In 1987, slightly more than 1.5 million people over the age of 18 lived in nursing homes or personal care homes (Table 1). Over 23 percent of these residents were 90 years or older and just under 10 percent were younger than 65. Over 73 percent of all nursing home residents were women, and this disproportion was even more pro- nounced in the oldest age groups. For example, 51 per- cent of women in nursing homes in 1987 were 85 years or older, while only 31.1 percent of the men were in this age range. Conversely, the proportion of men under age 65 (16.5 percent) was well over twice the proportion of resident women in this age category (7.2 percent). The absolute number of women far exceeds the number of men at all ages, and the ratio of women to men becomes larger for each successive age cohort. The vast majority of nursing home residents were white (90.8 percent). Another 7.5 percent of residents were black. There were more nursing home residents in both the Midwest (32.2 percent) and the South (29.9 per- cent) than either the Nonheast (22.2 percent) or the West (15.5 percent). Two-thirds of nursing home residents (67.9 percent) were no longer married. Only 12.8 percent were cur- rently married; another 18.1 percent had never been married. Time Since Admission Approximately 67 percent of residents had been ad- mitted to the nursing home more than a year prior to Jan- uary l, 1987; 20.6 percent had been in the facility for 5 years or more (Table 2). Long-stay residents were gen- erally below age 75 and above age 90. Approximately 10.7 percent had been admitted 90 or fewer days prior to January 1. Men were more likely than women to have short nursing home stays — 13.6 percent of the men Table 1. Selected demographic characteristics of nursing home residents, United States, January 1, 1987 Population Total characteristic population Percent (in thousands) Age (in years) All residents 1,5183 100.0 18 to 54 73 4.8 55 to 64 74 4.9 65 to 74 195 12.8 75 to 79 204 13.4 80 to 84 279 18.4 85 to 89 339 22.4 90 and older 354 23.2 Men 406 100.0 18 to 54 34 8.3 55 to 64 33 8.2 65 to 74 76 18.8 75 to 79 59 14.4 80 to 84 77 19.1 85 to 89 74 18.1 90 and older 53 13.0 Women 1,111 100.0 18 to 54 39 3.5 55 to 64 41 3 .7 65 to 74 118 10.6 75 to 79 145 13.0 80 to 84 202 18.1 85 to 89 266 23.9 90 and older 301 27.1 Racial background White 1,379 90.8 Black 1 14 7.5 US. Census region Northeast 342 22.5 Midwest 488 32.2 South 453 29.9 West 235 15.5 Marital status Married 195 12.8 No longer married 1,031 67.9 Never married 275 18.1 3 Includes all other racial groups not shown separately and persons of unknown marital status. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey —Institutional Population Component, phase 1. Table 2. Time since nursing home admission by age and sex, United States, January 1, 1987 Age (in years) Total and sex population 1-90 91-365 >1—<3 3- <5 5 (in thousands) days days years years years or more Percent of residents Total 1,518 10.7 21.7 31.5 15.3 20.6 18 to 54 73 14.6 21.5 22.4 16.4 25.0 55 to 64 74 7.9 18.5 31.8 10.9 30.9 65 to 74 194 11.7 23.3 30.0 12.7 22.5 75 to 79 203 11.4 24.5 31.9 13.1 19.1 80 to 84 279 13.1 25.5 30.2 14.9 16.3 85 to 89 339 11.0 20.5 35.9 15.9 16.5 90 and older 354 7.4 18.3 30.8 18.5 24.8 Men 406 13.6 24.7 29.5 14.7 17.6 18m 54 34 *12.1 21.6 23.5 17.3 25.6 55 to 64 33 14.9 *11.4 22.4 17.1 34.3 65 to 74 76 12.7 25.2 24.5 16.7 20.9 75 to 79 59 11.4 22.7 33.2 14.1 18.6 80 to 84 78 17.6 29.4 29.7 11.6 11.7 85 to 89 74 14.3 26.9 35.7 13.8 9.3 90 and older 53 10.6 26.9 31.8 14.6 16.1 Women 1,111 9.7 20.6 32.3 15.6 21.7 18 to 54 39 16.7 21.3 21.7 15.7 24.6 55 to 64 41 *2.2 24.3 39.5 *5.9 28.2 65 to 74 118 11.0 22.0 33.3 10.1 23.5 75 to 79 145 11.4 25.2 31.4 12.7 19.3 80 to 84 202 11.4 24.0 30.4 16.2 18.1 85 to 89 266 10.1 18.7 36.0 16.5 18.4 90 and older 301 6.9 16.8 30.6 19.2 26.4 aTime since admission is defined from the most recent date of admission to the facility to January 1, 1987, disregarding possible interruptions of the stay for hospitalizations. *Relative standard error is equal to or greater than 30 percent. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. had been in the facility for 90 or fewer days, compared to 9.7 percent of the women. The fact that men tend to have shorter stays becomes even more evident at later ages. Over 37 percent of men aged 90 years and older had been in the facility for a year or less, compared to 23.7 percent of women in that age group. For 26.4 per— cent of women 90 years or older, admission to the facil- ity had been 5 years or more prior to the beginning of 1987, while only 16.1 percent of the oldest men had stays of comparable length. Functional Status Overall, 1 1.3 percent of residents had no limitations in activities of daily living, while more than half (58.8 per- cent) had four or more difficulties (Table 3). (A resident was considered to be “limited” in function if personal assistance was needed to perform the ADL task.) For all residents, the age group below 55 years had relatively high proportions of residents either without ADL diffi- .. «a Table 3. Percent of nursing home residents requiring personal assistance with activities of daily living (ADLs) by age and sex, United States, January 1, 1987 Total Age (in years) population Number of ADL limitationsa and sex (in thousands) 0 1 2-3 4-5 Percent of residents Total 1,518 11.3 11.3 18.6 58.8 18 to 54 73 27.0 8.0 13.8 50.9 55 to 64 74 23.9 13.7 24.7 37.7 65 to 74 194 16.8 1 1.5 20.7 50.9 75 to 79 203 12.5 11.9 18.9 56.7 80 to 84 279 10.7 11.1 19.7 58.5 85 to 89 339 8.2 11.6 16.0 64.1 90 and older 354 5.0 10.8 18.5 65.7 Men 406 15.1 11.4 20.8 52.7 18 to 54 34 27.1 *10.8 *12.1 50.0 55 to 64 33 31.3 16.5 24.9 27.4 65 to 74 76 22.4 9.9 21.0 46.8 75 to 79 59 9.2 14.6 23.9 52.4 80 to 84 78 10.0 9.1 23.3 57.7 85 to 89 7‘4 9.7 7.7 17.3 65.3 90 and older 53 *8.2 16.1 21.0 54.7 Women 1,111 10.0 11.2 17.8 61.0 18 to 54 39 27.0 *6.0 15.3 51.7 55 to 64 41 18.0 11.4 24.6 46.0 65 to 74 118 13.3 12.6 20.5 53.6 75 to 79 145 13.9 10.9 16.8 58.4 80 to 84 202 10.9 11.9 18.3 58.8 85 to 89 266 7.8 12.7 15.7 63.8 90 and older 301 4.5 9.8 18.1 67.6 a ADLs include bathing, dressing, transferring from a bed or chair, toileting, and feeding. *Relative standard error is equal to or greater than 30 percent. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. culty or with four or more difficulties. For example, 27 percent of residents between 18 and 54 years had no lim- itations in activities of daily living and an additional 50.9 percent had at least four difficulties. Only 21.8 per- cent of residents in this age group had between one and three problems. In the older age groups, the distribution of ADL limitations shifts toward higher levels of diffi- culty. In most age groups, relatively few residents had only one difficulty. When there were differences between men and women within the same age group, women were more likely than men to have at least four ADL limitations and fewer had no difficulty: 10 percent of women compared to 15 .1 percent of men did not have any ADL difficulty. and 61 percent of women had at least four ADL difficul- ties, compared to 52.7 percent of men. While the proportion of women with severe levels of disability (for example, four or five ADL limitations or toileting or feeding difficulties) increased linearly with age, the pattern differed for men (Tables 3 and 4). The proportion of men who experienced four or more ADL difficulties increased until the age of 90, at which point the proportion of severely disabled men declined to well below the level of the prior age interval. Estimates of limitations in specific activities of daily living and of difficulty walking unaided show that the Table 4. Percent of nursing home residents requiring personal assistance for specific tasks by age and sex, United States, January 1, 1987 ADL limitatiomsa Total Age (in years) population Bed/chair and sex (in thousands) Bathing Dressing Toileting transfer Feeding Walkingb Total 1,518 87.5 76.0 63.6 61.7 33.8 63.2 18 to 54 73 72.2 62.7 53.3 51.8 38.8 52.8 55 to 64 74 75.2 59.2 45.2 40.6 17.0 44.8 65 to 74 194 81.9 70.8 55.2 53.7 27.9 56.3 75 to 79 203 84.7 75.4 63.8 59.0 33.7 59.9 80 to 84 279 87.7 77.6 64.8 61.7 33.9 61.3 85 to 89 339 91.0 78.0 67.7 66.6 37.1 69.0 90 and older 354 94.3 82.5 69.1 69.5 36.3 70.9 Men 406 83.9 72.5 57.4 54.9 30.1 57.3 18 to 54 34 72.9 62.1 47.7 50.0 40.9 49.8 55 to 64 33 68.7 52.3 37.5 27.4 *9.4 35.8 65 to 74 76 77.6 65.7 49.8 46.8 29.4 50.4 75 to 79 59 86.7 79.7 61.1 54.1 33.5 59.6 80 to 84 78 90.0 80.5 63.4 60.9 32.1 57.7 85 to 89 74 89.4 80.5 68.1 69.3 34.4 71.3 90 and older 53 90.1 71.0 59.5 59.4 24.7 63.1 Women 1,111 88.8 77.3 65.8 64.2 35.1 65.4 18 to 54 39 71.5 63.2 58.1 53.4 37.0 55.4 55 to 64 41 80.4 64.8 51.5 51.3 23.2 52.1 65 to 74 118 84.7 74.0 58.6 58.2 26.8 60.0 75 to 79 145 83.8 73.7 64.9 61.0 33.8 60.1 80 to 84 202 86.9 76.4 65.4 62.0 34.6 62.6 85 to 89 266 91.4 77.3 67.7 65.8 37.9 68.4 90 and older 301 95.1 84.5 70.8 71.3 38.3 72.3 3 Limitation is defined as requiring personal assistance with the activity. bWalking is included although not considered an ADL. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. vast majority of nursing home residents (87.5 percent) have difficulty bathing. There were no differences be- tween men and women in this respect. The remaining activities—dressing, toileting, transfer, and feeding— indicate progressively more serious levels of disability. With few exceptions, decreasing proportions of resi- dents had these limitations. Seventy-six percent of resi- dents needed assistance with dressing, 63.6 percent needed assistance toileting, 61.7 percent had problems transferring, and 33.8 percent had difficulty feeding themselves. Differences by age and sex generally be- come apparent when the activities indicating more se- vere disability are considered. For example, 57.4 per— cent of men had difficulty using the toilet, while 65.8 percent of women had this problem. Sex differences with respect to toileting become even greater in the old- est age group—59.5 percent of male residents 90 and older had toileting difficulties compared to 70.8 percent of female residents in the same age category. Walking difficulties were experienced by 63.2 percent of nursing home residents. The proportion of residents requiring personal assistance in walking was very simi- lar to the proportional distribution of residents with toi— leting difficulty. Sex and age differences reflect similar patterns to those of the ADL limitations. Another indicator of functional status considered in this report is urinary and bowel incontinence (Table 5). Over half (53.4 percent) of all nursing home residents had difficulty controlling their bladder at least several times a week or needed personal assistance in caring for bladder or bowel control equipment. Bowel inconti- nence was a problem for over 43 percent of residents. The proportion of residents with urinary incontinence increased with age, with one exception: incontinence was least prevalent in the group aged 55 to 64 years. Bowel incontinence showed less age-related increase. A higher proportion of women (55.3 percent) experi- enced urinary incontinence than men (48.1 percent), but there were no significant sex differences with respect to bowel incontinence. Mental Disorders Estimates of the number of residents by selected cate— gories of mental disorder (Table 6) indicate that 40.9 percent of nursing home residents were without a mental disorder. Another 28.7 percent had dementia only, in- cluding chronic or organic brain syndrome, and 13.7 percent had dementia in combination with one or more other mental disorders. A similar proportion (15.5 per- cent) had a mental disorder or disorders but no demen- tla. Table 5. Percent of nursing home residents with incontinence problems, by age and sex, United States, January 1, 1987 Total Percent incontinental Age (in years) population and sex (in thousands) Urinary Bowel Total 1,518 53.4 43.1 18 to 54 73 41.5 37.4 55 to 64 74 32.4 24.4 65 to 74 194 45.2 38.7 75 to 79 203 52.6 44.1 80 to 84 279 54.5 43.8 85 to 89 339 57.8 45.7 90 and older 354 60.3 46.9 Men 406 48.1 41.2 18 to 55 34 39.9 38.8 55 to 64 33 17.5 13.9 65 to 74 76 41.8 38.9 75 to 79 59 51.1 47.7 80 to 84 78 57.2 46.1 85 to 89 74 57.2 46.3 90 and older 53 52.6 41.6 Women 1,111 55.3 43.9 18 to 54 39 42.9 36.2 55 to 64 41 44.5 32.9 65 to 74 118 47.3 38.5 75 to 79 145 53.1 42.7 80 to 84 202 53.3 42.9 85 to 89 266 58.2 45.5 90 and older 301 61.6 47.8 “ Requires personal assistance in the care of urinary or bowel control devices or has difficulty controlling bladder or bowel several times or more per week. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. Table 6. Percent of nursing home residents with mental disorders by age and sex, United States, January 1, 1987 Total Dementia and Age (in years) population No mental Dementiaa other mental Other mental and sex (in thousands) disorder only disorder disorder only Percent of residents Total 1,518 40.9 28.7 13.7 15.5 18 to 54 73 49.5 *4.8 *6.7 36.1 55 to 64 74 40.2 *5.5 16.2 36.4 65 to 74 194 39.1 15.0 17.8 26.5 75 to 79 203 34.3 30.1 14.9 19.1 80 to 84 279 37.9 31.3 15.1 14.9 85 to 89 339 41.8 34.8 14.1 8.9 90 and older 354 45.6 37.3 10.2 5.6 Men 406 44.0 23.6 13.3 17.5 18 to 54 34 40.9 *6.3 *8.5 39.4 55 to64 33 42.9 *4.9 *15.1 36.0 65 to 74 76 41.9 17.0 14.5 24.8 75 to 79 59 40.9 26.1 12.3 18.0 80 to 84 78 41.9 29.3 17.2 10.8 85 to 89 74 45.9 34.1 12.1 7.3 90 and older 53 53.7 30.1 10.6 *4.5 Women 1,111 39.8 30.5 13.9 14.8 18 to 54 39 56.8 *3.4 *5.2 33.2 55 to 64 41 38.0 *5.9 17.2 36.7 65 to 74 118 37.3 13.6 20.0 27.6 75 to 79 145 31.6 31.7 16.0 19.6 80 to 84 202 36.3 32.1 14.3 16.5 85 to 89 266 40.7 35.0 14.7 9.4 90 and older 301 44.2 38.5 10.1 5.8 ‘ Includes any chronic or organic brain syndrome. b Includes at least one of the following: depressive disorder, schizophrenia, other psychoses, anxiety disorder, personality or character disorder, and other mental disorders. *Relative standard error is equal to or greater than 30 percent. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey—Institutional Population Component, phase 1. This pattern varied considerably by age and less by sex. Few residents under age 65 had any mental disorder involving dementia, although 36.1 percent of residents aged 18 to 54 had some type of mental disorder other than dementia. The pattern is very similar for residents 55 to 64 years old. With increasing age, however, small- er proportions of residents experienced disorders not in- volving dementia. In fact, for residents aged 90 and older, the most prevalent classification was either no mental disorder (45.6 percent) or dementia only (37.3 percent). Patterns of mental disorder in this nursing home population differed little between men and wom— 10 en. Where there were statistically significant sex differ- ences, a higher proportion of women than men had a diagnosis of mental disorder. Behavioral Problems and Psychiatric Symptoms Slightly less than half of all nursing home residents exhibited problem behaviors such as wandering, yell- in g, and hurting themselves or others physically (Table 7). The most prevalent behavior problem was getting upset and/or yelling (31.1 percent), followed by wan- Table 7. Percent of nursing home residents with behavioral problems by age and sex, United States, January 1, 1987 Age (in years) poprtiltziitlion Number of problems Type of problem and sex (in thousands) Upset/ Physically 0 1-2 3—4 5 - 10 yelling Wandering hurting others Percent of residents T0131 1,518 52.5 33.0 10.3 4.1 31.1 11.1 10.9 18 to 54 73 47.2 34.0 10.4 8.4 36.6 *5.7 12.5 55 to 64 74 42.8 37.4 15.2 *4.7 43.9 *7.2 14.5 65 to 74 194 47.3 35.7 11.9 5.0 35.7 8.8 14.0 75 to 79 203 53.4 29.8 12.6 4.1 30.4 11.7 12.2 80 to 84 279 53.8 32.5 9.3 4.3 27.1 11.7 10.2 85 to 89 339 54.3 32.0 9.4 4.3 30.3 14.7 9.8 90 and older 354 55.4 33.7 8.7 2.3 29.0 10.2 9.0 Men 406 48.3 35.4 11.3 5.0 33.4 12.0 16.1 18 to 54 34 48.5 32.3 *11.3 *7.9 31.9 8.0 16.0 55 to 64 33 40.8 41.2 *10.2 *7.8 42.5 *10.2 17.5 65 to 74 76 43.3 39.4 13.3 *4.0 36.8 8.6 20.1 75 to 79 59 52.8 27.8 12.9 *6.4 35.7 13.8 20.2 80 to 84 78 47.4 37.9 8.9 *5.7 27.9 13.9 13.7 85 to 89 74 48.4 36.8 11.2 *3.5 36.1 14.7 13.0 90 and older 53 56.2 30.7 10.7 *2.4 25.0 1 1.7 12.6 Women 1,111 54.1 32.2 9.9 3.8 30.3 10.8 9.0 18 to 54 39 46.1 35.5 9.6 *8.7 40.7 *3.7 *9.4 55 to 64 41 44.4 34.2 19.2 *2.2 45.0 *4.8 * 12.1 65 to 74 118 49.9 33.4 11.1 5.7 34.9 8.9 10.1 75 to 79 145 53.7 30.7 12.4 *3.2 28.3 10.8 8.9 80 to 84 202 56.3 30.5 9.4 3.8 26.8 10.8 8.9 85 to 89 266 55.9 30.7 8.9 4.5 28.7 14.7 8.9 90 and older 301 55.2 34.2 8.3 2.3 29.7 10.0 8.4 aReflects the most frequent problems and not mutually exclusive. *Relative standard error is equal to or greater than 30 percent. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. dering (11.1 percent) and physically hurting others (10.9 percent). There was a modest trend toward fewer problem behaviors with advancing age. Among resi- dents aged 18 to 54, 52.8 percent exhibited some behav- ior problems, compared to 44.7 percent of the oldest res- idents. Nearly 52 percent of male residents had behavioral problems, compared to 45.9 percent of fe- male residents. In particular, men were nearly twice as likely as women to hurt others physically (16.1 percent and 9 percent, respectively). The proportional differ- ences between men and women in the rate of problem behaviors were generally consistent at all ages. 11 More than two-thirds (68.4 percent) of all nursing home residents had one or more psychiatric symptoms, including dullness, withdrawal, impatience, delusions, and hallucinations (Table 8). Almost two-thirds of resi- dents exhibited at least one symptom of depression (64.3 percent). Nearly 30 percent experienced psychot- ic symptoms. The majority of residents exhibited either multiple psychiatric symptoms (44.5 percent) or no symptoms (31.6 percent); only 23.9 percent had only one symptom. While the number and type of symptoms did not vary considerably by age, women were more likely than men to experience multiple symptoms at Table 8. Type and number of psychiatric symptoms of nursing home residents by age and sex, United States, January 1, 1 987 Total Age (in years) population Number of symptoms Type of symptomsa and sex (in thousands) O 1 2 - 8 Depressiveb Psychoticc Percent of residents Total 1,518 31.6 23.9 44.5 64.3 29.7 18 to 54 73 40.3 17.5 42.2 55.5 32.9 55 to 64 74 32.8 18.0 49.1 64.5 32.8 65 to 74 194 31.3 22.8 45.9 63.8 30.0 75 to 79 203 30.6 25.8 43.6 64.6 29.3 80 to 84 279 29.5 26.8 43.6 65.8 29.5 85 to 89 339 32.5 23.8 43.7 64.8 27.0 90 and older 354 30.9 23.6 45.4 64.5 31.0 Men 406 35.8 23.0 41.1 61.0 24.1 18 to 54 34 46.7 17.0 36.3 49.6 27.5 55 to 64 33 37.1 18.8 44.1 59.8 29.1 65 to 74 76 35.2 22.2 42.6 61.5 23.6 75 to 79 59 32.7 23.8 43.4 63.7 18.7 80 to 84 78 31.4 24.6 44.0 64.0 27.4 85 to 89 74 36.0 25.8 38.2 62.1 24.3 90 and older 53 38.8 23.7 37.4 59.2 20.4 Women 1,111 30.0 24.2 45.8 65.5 31.7 18 to 54 39 34.8 17.9 47.3 60.6 37.5 55 to 64 41 29.4 17.4 53.2 68.3 35.8 65 to 74 118 28.8 23.2 48.0 65.3 34.2 75 to 79 145 29.8 26.6 43.6 65.0 33.6 80 to 84 202 28.8 27.7 43.4 66.5 30.3 85 to 89 266 31.5 23.3 45.2 65.5 27.8 90 and older 301 29.6 23.6 46.8 65.4 32.9 3 Categories are not mutually exclusive. b Includes one or more of the following: worry, apprehension, drowsiness, sluggishness, dullness, unresponsiveness, withdrawal, impatience, annoyance, and suspiciousness. C Includes one or more of the following: delusions, hallucinations, and verbalizing worthlessness. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. nearly all ages (overall, 70 percent of women compared with 64.2 percent of men). These differences between men and women were found for both depressive and psychotic symptoms, but the disparity was most notable for psychotic symptoms. Resident Composition by Facility Characteristics The health status of residents of nursing and personal care homes varied across different types of facilities. Independent nonprofit facilities differed from both non- 12 profit chains and independent for-profit nursing homes in the proportion of residents with high levels of ADL limitations (62.6 percent versus 53.3 and 54.4 percent, respectively; Table 9). Independent for-profit facilities had a higher proportion of residents with mental disor- ders (62.7 percent) than nonprofit chains (51.9 percent). Nursing home size, described in terms of the number of beds, appears to be associated with several health characteristics of residents. For example, facilities with fewer than 50 beds generally had lower proportions of physically and mentally frail residents. Almost a third Table 9. Functional limitations and mental disorders of residents by facility characteristics of nursing and personal care homes, January 1,1987 Total Facility population Number of ADL limitationsa characteristic (in thousands) Any mental O 1 2—3 4 - 5 Dementiab disordersc Total‘1 1,518 10.9 1 1.4 19.6 57.6 42.4 57.9 Facility ownership For—profit independent 3 85 14.7 12.0 18.9 54.4 44.7 62.7 For—profit chain 659 10.2 10.4 18.9 60.5 41.5 57.9 Nonprofit independent 233 8.5 12.2 16.7 62.6 43.4 56.4 Nonprofit chain 1 10 14.7 15.0 17.0 53.3 37.7 51.9 Public 131 9.1 8.9 20.7 61.3 42.4 59.5 Bed size 3—49 156 30.1 15.1 20.3 34.5 34.8 54.5 50—99 392 12.0 13.3 18.7 56.0 42.1 61.3 100- 149 478 7.3 9.6 17.5 65.6 42.2 55.7 150 or more 492 8.6 10.1 19.0 62.2 45.3 60.5 Facility type Hospital-based 55 3.4 6.9 18.6 71 . 1 42.9 60.5 Retirement center 1 13 20.9 12.4 16.6 50.0 39.5 53.6 Freestanding facility 1,335 10.7 1 1.4 18.7 59.2 42.6 58.9 Certification status SNF and [CF 669 6.1 9.8 18.2 65.0 42.6 56.9 SNF only 314 3.7 7.5 15.5 73.3 49.3 64.7 ICF only 347 8.6 13.7 23.9 53.9 44.2 61.6 Noncertified 188 47.6 18.7 15.4 18.3 26.8 49.1 “ ADLs include bathing, dressing, transferring from a bed or chair, toileting, and feeding. b Includes any chronic or organic brain syndrome. C Includes one or more of the following: dementia, depressive disorder, schizophrenia, other psychoses, anxiety disorder, personality or character disorder, and other mental disorders. dIncludes all other health facility types not shown separately. Source: Agency for Health Care Policy and Research. National Medical Expenditure Survey — Institutional Population Component, phase 1. (30.1 percent) of residents in these small facilities had no ADL limitations, compared to only 8.6 percent of residents in facilities with more than 150 beds. Like- wise, the proportion of residents with dementia residing in facilities with fewer than 50 beds (34.8 percent) was lower than in the largest facilities (45.3 percent). These differences were not apparent, however, when all men- tal disorders were considered. Not surprisingly, hospital-based facilities had the highest proportion of physically frail residents—only 3.4 percent had no ADL limitations while 71.1 percent had four or more—than residents in retirement centers 13 or freestanding facilities. Residents in retirement facili- ties appeared to be least frail, with 20.9 percent report- ing no ADL limitations and just half having four or more limitations. By contrast, there was little difference by facility type with respect to the prevalence of mental dis- orders, including dementia. Homes certified either as skilled nursing facilities (SNFs) or intermediate care facilities (ICFs) had higher proportions of ADL-limited residents than noncertified facilities. Nearly half of residents in noncertified facili— ties (47.6 percent) had no ADL limitations. A far small- er proportion of residents of noncertified facilities had dementia (26.8 percent) than those in facilities with any type of program certification. Summary Consistent with other national studies of nursing and personal care home residents, such as the 1977 and 1985 National Nursing Home Surveys, the estimates from phase 1 of the 1987 NMES Institutional Population Component confirm that the majority of nursing home residents were female, white, no longer married, and over age 80. Overall, residents had high levels of func— tional difficulty: more than half needed help with four or more activities of daily living; a similar proportion were bladder incontinent, and more than 43 percent were bowel incontinent. Over 42 percent of residents had some form of dementia, either alone or in conjunction with other mental disorders. Given that this is an analy- sis of all current residents, it is not surprising that over 20 percent had been in the nursing home for more than 5 years and almost 70 percent for more than 1 year. However, the nursing home population was not homo- geneous. There were important differences between younger and older residents, and between males and fe- males. In comparison to the rest of the nursing home popula- tion, the female-to—male ratio was considerably lower for residents younger than age 65. These younger nurs- ing home residents tended to have either very severe ADL difficulties or none. While few of the younger residents had dementia, they were more likely than older residents to have other mental disorders in the absence of dementia. They were also more likely than those aged 65 to 84 to have been in the nursing home for 3 or more years. These trends probably reflect the fact that younger persons enter nursing homes for different rea- sons than persons over 65, such as accidents resulting in severe long-term disability, degenerative neurologic disorder, or chronic mental disorders. By contrast, resi— dents over age 65 tend to be physically frail, to be suffer- ing from dementia, or to have outlived informal caregiv- ers. The ratio of female to male residents increased with age, and female nursing home residents differed from their male counterparts in important ways. They were older than male residents, and generally more likely to have four or more ADL limitations and dementia. This supports the notion proposed by several gerontologists (for example, Katz and others, 1983; Shanas and Mad- dox, 1985) that the duration of dependence—the length of time from the onset of functional limitations to 14 death—is considerably longer for women than for men. In addition to problems that result from largely physical conditions, women in nursing homes experienced more psychiatric symptoms while men exhibited more behav- ioral problems; this may reflect sex-related patterns of disability or differences in the way men and women ex- press emotional distress. Levels of disability among residents also varied by type of facility. Persons residing in small or noncertified facilities or retirement centers tended to be less disabled than those in larger or certified facilities; hospital—based nursing facilities and particularly those with SNF certi- fication had the most highly disabled residents. Further research is needed to examine the relationships between facility and resident characteristics suggested by these findings. In sum, this analysis confirms prior reports that the majority of nursing home residents are frail and that a significant proportion are without immediate family. Although the majority of residents in 1987 were female and over 65, male and younger residents differed in im- portant ways from the majority population in nursing homes. Given the demographic changes now occurring in the United States, the availability of family supports for physically and mentally frail elders may decrease as the size of this population continues to grow. These converging factors make investigations into institution- al care, of the residents in institutional care, and of those in the community but at risk for institutionalization high priorities for future research. Technical Appendix Data Sources and Methods of Estimation The data in this report are from the Institutional Popu- lation Component (IPC) of the 1987 National Medical Expenditure Survey (NMES). The IPC was designed to provide unbiased national and regional estimates for the population in nursing and personal care homes and facilities for the mentally retarded as well as to yield es- timates for the institutional facilities and a range of their characteristics. Data were collected in three phases with four interviews. The eligible sample consisted of 851 nursing and personal care homes, of which 95.2 percent responded; 99.4 percent of the responding facilities per- mitted sampling. In these facilities, resident data were obtained for 98.8 percent. Facility data were obtained from the facility question- naire. Respondents were facility administrators or des- ignated staff. The reference period was the date of the interview (February-May 1987). Data on resident char- acteristics were obtained from a baseline questionnaire for each resident. Respondents were staff responsible for direct patient/client care (caregivers) or other desig- nated staff. The reference period was January 1, 1987. For additional details on data collection methods see Ed- wards and Edwards (1989). The estimates in this report are based on 800 nursing and personal care homes and 3,337 residents 18 years of age and older residing in these facilities as of January 1, 1987, and for whom data were collected in the first phase of the IPC (February-May, 1987). These esti- mates are preliminary and subject to revision as more information from other parts of the survey becomes available. Definition of a Nursing Home A facility is considered a nursing or personal care home if the facility is certified by Medicare or Medicaid as a skilled nursing facility (SNF) or intermediate care facility (ICF) or is a separate place or unit of another institution that is licensed or officially recognized by a State, has three or more beds for patients where the pa- tients reside, and provides patients with personal care assistance. Excluded as a nursing or personal care home are noncertified places licensed only as hospitals and residential facilities that limit care exclusively to per- sons with specific physical, mental, and emotional con- ditions. Personal care assistance is defined as nursing or medical care, supervision of medications, or help with bathing, dressing, walking, eating, correspondence, or shopping. By the above definition, all SNF— or ICF—cer— tified units of licensed hospitals were considered nurs- ing homes. In this report, the term “nursing home” is used to refer to nursing and personal care homes. Derivation of Data on Residents Racial background. Classification by racial back- ground was based on information reported for each resi- dent. Respondents were asked if the racial background of the resident was best described as American Indian or Alaska Native; Asian or Pacific Islander; black, white, or other. Estimates of race were collapsed into three categories—white, black, and other; the “other” race categories are not shown separately due to the small sample size. Marital status. Facility respondents were asked if, on January 1, 1987, the sampled person was married, wid- owed, divorced, separated, or never married. These re- sponses were validated with next-of—kin survey data 15 (see Edwards and Edwards, 1989, for details on this component of the survey). In cases where the facility and the next of kin reported marital status inconsistent- ly, the next-of—kin data were favored, provided that time could be ruled out as a factor for the difference in re- ported status. For cases without available facility data, next-of—kin data were used. For approximately 3 per- cent of the cases, marital status was imputed logically. Marital status was collapsed into three categories for the purposes of this report: married, never married, and no longer married (a combination of widowed, divorced, and separated). Census region. Classification by region was deter- mined by the location of the sampled facility. Region was defined in terms of the categories developed by the US. Bureau of the Census. Time since admission. Time since admission was cal- culated by subtracting January 1, 1987, from the key ad- mission date (KAD) to yield a standardized time since admission for current residents. KAD is defined as the most recent date of admission to the nursing home, ig- noring discharges and admissions entirely related to the hospital care of patients previously in the facility (Short, Cunningham, and Mueller, in press). January 1, 1987, was used as the end date in the calculation rather than the interview date in order to further standardize the metric for this time interval. Activities of daily living (ADLs). Respondents were asked to indicate whether the resident had difficulty per- forming personal care activities commonly known as activities of daily living (Katz and Akpom, 1976) and widely used in both institutional and community set- tings to characterize the individual’s ability to function independently and without physical limitation resulting typically from a complex of physical conditions. Five activities were included in the summary ADL measure: bathing, dressing, transferring from a bed or chair, toi~ leting (getting to and from the toilet and physically using the toilet), and feeding oneself. Residents were deter- mined to have difficulty with a particular activity if they received personal assistance or supervision in perform- ing the task or could not perform the activity at all. Diffi- culty walking, while not considered an ADL and not in- cluded in the ADL totals in Table 3, is presented as a separate activity. Defined in this way, ADL difficulty was a constructed variable, where persons with missing data were assumed to have no difficulty with the activ- ity. In all cases, missing data never exceeded 1 .5 percent of the total. Incontinence. Respondents were asked about bowel and bladder control. For bowel continence, questions included whether the resident used a colostomy bag or other bowel control device, whether help from another person in caring for this device was necessary, whether the resident had any difficulty controlling his or her bowels, and the frequency of this problem. Separate questions for bladder control were asked and followed the same format except those asked about urinary cathe- ters and similar devices. Bowel incontinence was de— fined as either needing personal assistance in caring for bowel control equipment or having accidents at least several times a week. Similarly, urinary incontinence was defined as either needing personal assistance in car— ing for urinary catheters or other devices, or having ac- cidents at least several times a week. Persons with miss— ing data were assumed to have no problems with incontinence; in all cases, missing data never exceeded 1 percent of the total. Mental disorders. Respondents were asked to indi- cate whether according to the resident’s medical record, the resident had any of a series of mental disorders (the medical record was consulted in 97 percent of the inter- views). The questions were in list format and included the following: senile dementia/chronic or organic brain syndrome; depressive disorders; schizophrenia; other psychoses; anxiety disorders; personality/character dis- orders; and any other mental disorders. Multiple condi- tions were recorded and subsequently grouped into mu— tually exclusive categories as follows: 1) no mental disorders—no positive responses to any conditions in the checklist; 2) dementia only—positive response only to the question regarding senile dementia/chronic or or— ganic brain syndrome; 3) dementia and other mental disorders—positive response to the question regarding senile dementia/chronic or organic brain syndrome and at least one other condition in the checklist; 4) other mental disorders only—at least one positive response to the questions regarding other conditions in the checklist and a negative response to the senile dementia question. Persons with missing data were considered to have no disorders for the purposes of constructing these catego- ries. The magnitude of missing data on any item was consistently less than 1.5 percent of the total. Behavioral problems. The Uniform Client Data In- strument (Mulken and Manderscheid, 1989) was used to assess behavioral problems. Respondents were asked whether the resident sometimes disturbed the caregiver or others by engaging in any of 10 problem behaviors. These problems included: getting upset, physically l6 hurting others, physically hurting oneself, dressing in- appropriately, crying for long periods, hoarding, wan- dering, not avoiding dangerous things, stealing, and in- appropriate sexual behavior. Across all items there was consistently less than 0.5 percent of missing data and the item nonresponse was imputed as “no” for the purposes of this report. Psychiatric symptoms. The Uniform Client Data In- strument was also used to assess psychiatric symptoms. Conceptually, items group into “depressive” and “psy- chotic” symptoms. Endorsement of any of these items does not necessarily indicate that the resident had adiag- nosis of depression or any psychoses. Respondents were asked eight questions regarding frequently exhib- ited psychiatric symptoms. For depressive symptoms, these were worry/apprehension, drowsiness, withdraw- al, impatience, and suspicion. In addition, respondents were asked if the resident ever exhibited the following psychotic behaviors: delusions, hallucinations, and ver- balizations of worthlessness. Across all items there were consistently less than 0.5 percent of missing data and the item nonresponse was imputed as “no” for the purposes of this report. Facility Characteristics Ownership. Respondents reported which ownership type best described their facility: for-profit (individual, partnership, or corporation), private nonprofit (religious group, nonprofit corporation, etc.), or one of four types of public ownership—city/county government, State government, Veterans’ Administration (now the Department of Veterans Affairs), or other Federal agency. Respondents also reported whether their facil- ity was part of a chain or group of facilities operating under one general authority or ownership. Facilities with a negative response were classified as independ- ently owned. Facility size. Facility size was determined by the number of beds in the facility regularly maintained for residents. Beds used by staff or for day care patients were excluded. In the case of a unit in a larger facility, only the unit beds were included. Facility type. This variable was derived from data from the facility questionnaire and defined the facility’s organizational structure: hospitals; retirement facilities, including continuing care facilities; freestanding facili- ties; and other types of facilities. The order of priority for coding facility type followed the sequence listed above. Any facility that was not considered as one of the first three types was categorized as other. Certification status. Respondents were asked wheth- er their facility had any beds certified by Medicare as SNF beds; any beds certified by Medicaid as SNF beds; and any beds certified by Medicaid as either ICF beds or ICF-MR (intermediate care facility for the mentally re- tarded) beds. For the purposes of this report, facilities were assigned to mutually exclusive categories based on the responses to those questions. Facilities responding affirmatively to both the SNF and the ICF questions were classified as SNF— and ICF-certified facilities. A very small number of nursing homes responded posi- tively to the ICF—MR question and negatively to SNF Medicare, SN F Medicaid, and ICF Medicaid questions; these were classified as noncertified. Sample Design and Standard Error Estimates The IPC target population included all persons who in 1987 spent at least one night in a nursing or personal care home or a facility for mentally retarded persons. The sample was designed to yield unbiased national and re— gional estimates of health care parameters at the facility and resident level. At both levels, the information is classified as to the type of institution, i.e., nursing and personal care homes and facilities for the mentally re- tarded. The sample was selected according to a strati- fied three-stage probability design, with facility selec- tion in the first two stages. The Inventory of Long-Tenn Care Places served as the facility sampling frame. The final stage of selection consisted of a sample of residents as of January 1, 1987, and a “rolling” sample of persons admitted during the year. Thus, persons residing in sampled facilities on January 1 were included in the sample, as were persons admitted at any time during 1987, up to and including December 31. This report restricts the sample to those residing in nursing or personal care homes as of January 1, 1987. For a detailed description of the survey design and of sampling, estimation, and adjustment methods, includ- ing weighting for nonresponse and poststratification, see Cohen, Flyer, and Potter (1987). Reliability of Estimates Since the statistics presented in this report are based on a sample, they may differ somewhat from the figures that would have been obtained if a complete census had been taken. This potential difference between sample 17 results and a complete count is the sampling error of the estimate. The chance that an estimate from the sample would differ from a complete census by less than one standard error is about 68 out of 100. The chance that the differ— ence between the sample estimate and a complete cen— sus would be less than twice the standard error is about 95 out of 100. Tests of statistical significance were used to determine whether differences between facility estimates exist at specified levels of confidence or whether they simply occurred by chance. Differences were tested using Z-scores having asymptotic normal properties, based on the rounded figures at the 0.05 level of significance. Unless otherwise noted, only statistically significant differences between estimates are discussed in the text. Rounding Estimates presented in the tables have been rounded to the nearest 0.1 percent or to the nearest thousand. The rounded estimates, including those underlying the standard errors, will not always add to 100 or the full total. Standard Errors Standard errors for the statistics in this report were ap- proximated, by interpolation where necessary, using a curve smoothing procedure developed by Cohen (1979). The statistical tests in this report, however, are based on direct estimates of standard errors using the Taylor Series linearization method. Relative standard errors of totals. Where the statis- tics of interest are total estimates (T) of the population, an estimate of the standard error, SE, can be obtained by multiplying the relative standard error, RSE(T), ex— pressed as a percent of the respective total (T), by T and then dividing by 100. Thus, T(RSE(7)) SE = — (T) 100 For estimated population totals for individuals, the ap- proximate relative standard errors expressed as percents are as shown in Table I. - Example—A rounded estimate of 1,111,000 fe- male nursing home residents aged 18 and older (Table 1) has a relative standard error of between 4.0 and 3.6 (Table I) or, by interpolation, a relative standard error of about 3.8 percent. The standard error of this estimate, then, is 1,111,000(3.8) 55(1) = 100 = 42,218 Direct standard error estimates. When the statistic of interest is expressed as a percent of the number of persons, direct estimates of standard errors have been total of 1,518,000 (Table 1). This estimate, by inter- polation, has a standard error of approximately 0.6 percent (Table II). Table I. Approximate relative standard errors Estimated population totals for individuals (in thousands) Relative standard error (percent) derived for ease of calculation. For the estimated per— 10 22.9 cent of nursing home residents by selected characteris- 25 14.7 tics, approximate standard errors expressed as a percent 138 1(7): are as shown in Table II. 250 5.5 o Example—The estimate of 90.8 percent of the U.S. 500 4.4 nursing home population aged 18 and older with a 750 40 white racial background is based on a population 1,518 3-6 Table II. Approximate direct standard errors Estimated percent Persons in the base of the percent 2 or 5 or 10 or 20 or 30 or 40 or (in thousands) 98 95 90 80 70 6O 50 25 2.0 3.1 4.3 5.7 6.6 7.0 7.2 50 1.4 2.2 3.0 4.1 4.7 5.0 5.1 100 1.0 1.6 2.2 2.9 3.3 3.5 3.6 250 0.6 1.0 1.4 1.8 2.1 2.2 2.3 500 0.4 0.7 1.0 1.3 1.5 1.6 1.6 750 0.4 0.6 0.8 1.0 1.2 1.3 1.3 1,518 0.3 0.4 0.6 0.7 0.8 0.9 0.9 References Institutional Population Component. Paper presented at Brown, R. and D. Stoudt .(1978). 1976 Survey oflnsti- tutionalized Persons: A study of persons receiving long-term care. Current Population Reports Special Study Series P—23, No.69. Washington, DC: U.S. Gov- ernment Printing Office. Cohen, SB. (1979). An assessment of the curve smoothing strategies which yield variance estimates from complex survey data. In American Statistical As- sociation: Proceedings of the Section on Survey Re- search. Alexandria, VA: American Statistical Associ- ation. Cohen,S.B.,P.Flyer, and D.E.B. Potter. (1987). Sam- ple design of the National Medical Expenditure Survey 18 the annual meeting of the American Public Health Asso- ciation, New Orleans, LA. Cunningham, P. and C. Mueller. (in press). C haracter- istics of residents in facilities for the mentally retarded. National Medical Expenditure Survey Research Find- ings 6, Agency for Health Care Policy and Research. Rockville, MD: Public Health Service. Edwards, W.S. and B. Edwards. (1989). Question- naires and data collection methods for the Institutional Population Component (DHHS Publication No. (PHS) 89—3440). National Medical Expenditure Survey Meth- ods 1, National Center for Health Services Research and Health Care Technology Assessment. Rockville, MD: Public Health Service. Ernst, M. and NS. Emst.(1984). Functional capacity. In DJ. Mangen and WA. Peterson (Eds.). Research In- struments in Social Gerontolo, y 3. Health, Program Evaluation, and Demography (pp. 9—84). Minneapolis, MN: University of Minnesota Press. Hing, E. (1987). Use of nursing homes by the elderly: Preliminary data from the 1985 National Nursing Home Survey. Advance Data from Vital and Health Statistics No. 135 (DHHS Publication No. (PHS)87-1250). Hy- attsville, MD: Public Health Service. Hing, E., E. Sekscenski, and G. Strahan. (1989). The National Nursing Home Survey: 1985 summary for the United States. Vital and Health Statistics Series 13, No. 97 (DHHS Publication No. (PHS)89—1758). Washing- ton, DC: US Government Printing Office. Keeler, E.B., R.L. Kane, and DH. Solomon. (1981). Short- and long-term residents of nursing homes. Medical Care 14(3), pp. 363-369. Katz, S. and CA. Akpom. (1976). A measure of pri— mary sociobiologic functions. International Journal of Health Services 6, pp. 493-508. Katz, 8., LG. Branch, M.H. Branson, J.A. Papsidero, J.C. Beck, and D.S.Greer. (1983). Active life expect- ancy. New England Journal of Medicine 309, pp. 1218-1224. 19 PUBLIC HEALTH LIBRARY Letsch, S.W., K.R. Levit, and DR. Waldo. (1988). Health care financing trends: National health expendi— tures, 1987. Health Care Financing Review 10 (2), pp. 109—122. Liu, K. and KO. Manton. (1983). The characteristics and utilization pattern of an admission cohort of nursing home patients. The Gerontologist 23(1), pp. 92-98. Mulken, V.M. and R. Manderscheid. (1989). Charac- teristics of community support program clients in 1980 and 1984. Hospital and Community Psychiatry 40(2), pp. 165-172. Shanas, E. and G.L. Maddox. (1985). Health, health resources, and the utilization of care. In R.H. Binstock and E. Shanas (Eds.). Handbook onging and the Social Sciences. New York: Van Nostrand Reinhold. Shah, B.V. (1981). SESUDAAN: Standard Errors for Computing of Standardized Rates from Sample Survey Data. Research Triangle Park, NC: Research Triangle Institute. Short, P., P. Cunningham, and C. Mueller. (in press). Standardizing nursing home admission dates for short-term hospital stays. Medical Care. Zappolo, A. (1981). Discharges from nursing homes. Vita] and Health Statistics Series 13, No. 54 (DHHS Publication No. (PHS)81- 1715). Hyattsville, MD: Na- tional Center for Health Statistics. I U.C. BERKELEY LIBRARIES lllllllllllllllll/l (0362147332 DEPARTMENT OF HEALTH & HUMAN SERVICES Public Health Service Agency for Health Care Policy and Research Parklawn Building, Room 18-12 Rockville, MD 20857 DHHS Publication No. (PHS) 90-3470