f } 1C -T I2 HAH Health Planning Bibliography Series Methods for Setting Priorities in Areawide Health Care Planning: An Annotated Bibliography U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Health Service Health Resources Administration The Division of Health Planning, Bureau of Health Planning and Resources Development, through the National Health Planning Information Center, is a primary resource for current information on a wide variety of topics of interest to health planners. To facilitate the dissemination of information to health planners, the Center issues selected publications in three series: & 1. Health Planning Methods and Technology This series focuses on the technical and administrative aspects of health planning. Included are such areas as methods and approaches to the various aspects of the health planning process, techniques for analyzing health planning information and problems, and approaches to the effective dissemination and utilization of technical information. 2. Health Planning Information This series presents trend data, data collection, and data analysis methods, including sources of data to support health planning activities. 3. Health Planning Bibliography Bibliographies related to specific subject areas in health planning are published in this series. "Methods for Setting Priorities in Areawide Health Care Planning: An Annotated Bibliography" is the eighth publication in the Health Planning Bibliography Series. Methods for Setting Priorities in Areawide Health Care Planning: An Annotated Bibliography Prepared by Arthur Young & Company Washington, D.C. Under Contract No. HRA 230-76-0066 April 1978 HRP-0300801 U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Health Service Health Resources Administration Bureau of Health Planning and Resources Development Division of Health Planning National Health Planning Information Center DHEW Publication No. (HRA) 78-14006 iemand. fee ad + FOREWORD "Methods for Setting Priorities in Areawide Health Care Planning: An Annotated Bibliography," is intended as an aid to health planners involved in developing areawide plans for health care. It was de- veloped as a companion document to "Methods for Setting Priorities in Areawide Health Care Planning" being published simultaneously by the Bureau of Health Planning and Resources Development. In preparation of this document, the authors analyzed the wide range of priority setting methods that would be useful to planners and included those considered to have the greatest relevance. Planners seeking guidance in determining a methodology for setting priorities are encouraged to review pertinent literature referenced in this publication to gain a broader perspective of the advantages and disadvantages of a variety of currently available priority setting approaches. Bureau of Health Planning and Resources Development Health Resources Administration iii Q1928 # kee oo M CONTENTS PAGE FOREWORD jai INTRODUCTION j CHAPTER I. ANALYTICAL APPROACHES 3 PRIMARY EMPHASIS ON _ CONCEPTS % Archibald, K.A.: Three Views of the Expert's Role in Policymaking: Systems Analysis, Incrementalism, and the Clinical Approach. Policy Sciences 3 2. Bell, D.: Twelve Modes of Prediction -- A Preliminary Sorting. Daedalus: The Proceedings of the American Academy of Arts and Sciences 4 3. Berki, S.E.: "Demand and Utilization," chapter of 6 of Hospital Economics 5 4 . Boulding, K.E.: The Concept of Need for Health Services. Milbank Memorial Fund quarterly 6 9. Dunlop, D.W.: The Development of An Output Concept For Analysis of Curative Health Services. Social Science and Medicine 8 6. Etzioni, Amitai.: "On the Process of Making Decisions." Science 10 7% Feldstein, P.J.: Research on the Demand for Health Services. Milbank Memorial Fund quarterly ° 11 8. Goldsmith, Seth B.: - "The Status: of Health Status Indicators." Health Services Report 12 9 . Grossman, M.: On the Concept of Health Capital and the Demand for Health. Journal of Political Economy 13 TO. Maass, Arthur A.: "Benefit-Cost Analysis -- Its relevancy to Public Investment Decisions." Quarterly Journal of Economics XXIX 14 11. 12. 13. 14. 1§. 16. Newhouse, J.P., C.E.: On Having Your Cake and Eating It Too: Econometric Problems In Estimating The Demand For Health Services. The Rand Corporation Plessas, Demetrius J., Fein, Ricca.: "An Evaluation of Social Indicators." AIP Journal Prest, A.R., and R. Turvey.: "Cost-Benefit Analysis: A Survey." The Economic Journal Quade, E.S.: On the Limitations of Quantitative Analysis. The Rand Corporation Reiner and Davidoff.: "Choice Theory of Planning." JAIP. Wiseman, Jack: "Cost-Benefit Analysis and Health Service Policy." The Scottish Journal of Political Economy PRIMARY EMPHASIS ON METHODS 17. 18. 19; 20. 21. 22. Abernathy, W.J. and Hershey, J.C.: A Spatial- Allocation Model for Regional Health-Services Planning. Operations Research Alonso, W.: Predicting Best with Imperfect Data. Journal of the American Institute of Planners Anderson, J.G.: Demographic Factors Affect- ing Health Services Utilization: A Causal Model. Medical Care Anderson, J.G. Bartkus, D.E.: Choice of Medical Care. A Behavioral Model of Health and Illness Behavior Anderson, R. and Benham, L.: Factors Affect- ing the Relationship Between Family Income and Medical Care Consumption, Empirical Studies in Health Economics Anderson, R., Smebdy, B., Eklund, G.: Automatic Interaction Detector Program for Analyzing Health Survey Data. Health Services Research vi PAGE 15 17 19 20 21 23 24 295 27 28 29 30 23. 24. 25. 26. 27. 28. 29. 30. 31. 32% 33. PAGE Ault, D.ER.-and Johnson, B., Probabllistic Models Applied to Hospital Plannlng. Growth and Change 32 Cooper., J.K., and Corcoron, Ti M/: ~ Estimat- ing Bed Needs by Means of Queueing Theory. New England Journal of Medicine 33 Cordle, F., and Tyroler, H. A.: The Use of Hospital Medical Records for Epidemiologic Research. I. Differences in Hospital Utiliza- tion and In-Hospital Mortality by Age-Race-Sex- Place of Residence and Socioeconomic status in in a Defined Community Population. Medical Care 34 Davis, K. and Russell, L.B.: The Substitu- tion of Hospital Outpatient Care for Inpatient Care. The Review of Economics and Statistics 35 Feldstein, M.S.: Hospital Cost Inflation: A Study of Nonprofit Price Dynamics. The American Economic Review 36 Feldstein, P.J., and German, J.J.: Predict, ing Hospital Utilization: An Evaluation of Three Approaches. Inquiry 37 Gabrielson, I.W., Siken, E., Sohler, K.B., and Stockwell, E.G.: Relating Health and Census Information for Health Planning. American Journal of Public Health 38 Grimes, R. M., Allen, C.L., and Sparling, T.R.: Use of Decision Theory in Regional Planning. Health Services Research 39 Hill, Morris.: "A Goal-Achievement Matrix for Evaluating Alternative Plans." JAIP 40 Kalimo, E., Sievers, K.: The Need for Medi- cal Care: Estimation On the Basis of Inter- view Data. Medical Care 41 Klee, A.J.: "Let DARE Make Your Solid Waste Decision." American city 42 vii 34. 35. 36. 37. 38. 39. 40. 410 42. 43. 44. 45. Kong-Kyun Ro: Interactions Among Variables Affecting Hospital Utilization. Health Services Research Leibenstein, H.: Allocative Efficience Vs. X-Efficiency. American Economic Review Litchfield, Nathaniel: "Cost-Benefit Analysis In Plan Evaluation." The Town Planning Review Meadows, D.H., Meadows, D.L., Randers, J., et al: The Limits To Growth: A Report For The Club of Rome's Project on the Predica- ment of Mankind Navarro, V., Parker, R., and White, K.L.: A Stochastic and Deterministic Model of Medical Care Utilization. Health Services Research Nomile, F.R., and Ziel, H.A. Jr.: Too Many OB Beds? Hospitals Phelps, C.E. and Newhouse, J.P.: Coinsurance, the Price of Time, and the Demand for Medical Services. The Review of Economics and Statis- tics Phillip, Joseph P.: Some Considerations Involved in Determining The Optimum Size of Specialized Hospital Facilities. Inquiry Pollack, J. and Taf, M.: - "Urban Planning; Ripe for Systems Analysis." Journal of Systems Management Quade, E.S.: Cost Effectiveness: Some Trends in Analysis Reinke, W.A.: Analysis of Multiple Sources of Variation: Comparison of Three Techniques. Health Services Research Rosenthal, G.D.: The Demand for General Hospital Facilities viii PAGE 4 4 45 4 6 47 48 49 50 52 53 54 56 PAGE 46. Rosenthal, G.: Price Elasticity of Demand for Short-Term General Hospital Services, Empirical Studies in Health Economics 56 47 . Schneider, J.B.: Measuring the Locational Efficiency of the Urban Hospital, Health Services Research 59 48. Schonick, W. and Jackson, J.R.: An Improved Stochastic Model for Occupancy-Related Random Variables in General-Acute Hospitals. Operations Research 60 49. Wirick, G.: A Multiple Equation Model of Demand for Health Care. Health Services Research SEA 61 50. Yett, D.E., Drabek, L., Intriligater, M.D.: A Macroeconometric Model for Regional Health Planning. Business and Economic Bulletin 62 S1. Vietorisz, Thomas.: "Quantized Preferences and Planning By Priorities." American Economic Review 63 CHAPTER II. GOAL-HARMONIZING APPROACHES 64 PRIMARY EMPHASIS ON CONCEPTS 1. Bender, Douglas A.: "Delphic Study Examines Developments in Medicine." Futures 64 2. Donabedian, A.V.: "Assessment of Need", Chapter 3 of Aspects of Medical Care Administration: _ Specifying Requirements for Health Care 65 3% Jeffers, J.R., Borgnanno, M.F., and Bartlett, J.C.: On the Demand Versus Need for Medical Services and the Concept of 'Shortage'. American Journal of Public Health 66 4. Jeffers, J. R., Borgnanno, M.F., and Bartlett, J.C.: Demand Versus and the Concept of Shortage - Rejoinder. American Journal of Public Health 66 ix 10. L1. 12. Kahn, H. and Wiener, A.J.: The Year 2000: A Framework for Speculation Kaufman, H.: "The Politics of Health Planning." AJPH Mason, R.0O.: A Dialectical Approach To Strategic Planning, Management Science Mucllor, H.F., Uphoff, W.H., and Zocliner, H.: Medical Services Demand Versus Need - A Comment. American Journal of Public Health Newhouse, J.P., Phelps, C.E., and Schwartz, wW.B.: Policy Options and the Impact of National Health Insurance. New England Journal of Medicine Schlesinger, J.R.: Systems Analysis And The Political Process. Simon, Herbert.: "On the Concept of an Organizational Goal." Administrative Science Quarterlz Woods, D.H.: Improving Estimates that Involve Uncertainty. Harvard Business Review PRIMARY EMPHASIS ON METHODS 13. 14. 15. 16. J 7. Delbecq, A. and Van De Van, H.: "A Group Process Model for Problem Identification and Program Planning." The Journal of Applied Behavioral Science Lamanna, Richard A.: "Value Consensus Among Urban Residents." JAIP Lawson, J.S.: How Many Beds? Problems in Estimating Requirements for Hospital and Nursing Home Beds. Medical Journal of Australia Moore, W.S. and Bock, H.B.: Estimating the Demand for Medical Care. Inquiry Navarro, V.: A Systems Approach to Health Planning. Health Services Research PAGE 68 70 73 66 7 2 72 74 75 76 7 78 T9 80 18. 19. 20. 21. 22» 23. 24. 29. 26. 27. 28. Pill, J.: The Delphi Method: Substance, Context, a Critique and An Annotated Biblio- graphy. Socio-Economic Planning Sciences Putnam, S.H.: Intra-urban Employment Fore- casting Models: A Review and Suggested Model: Journal of the American Institute of Planners Richardson, A.H. and H.E. Freeman: Evalua- tion of Medical Care Utilization by Interview Surveys. Medical Care Richardson, J.D. and Scutchfield, F.D.: Priorities in Health Care: The Consumer's Viewpoint in an Appalachian Community. American Journal of Public Health Sackman, Harold: Delphi Critique: Expert Opinion, Forecasting and Group Process Sears, D.W.: Elderly Housing: A Need Assessment Technique. The Gerontologist Shillif, I.E., and Smith, R.D.: A Fore- casting Method for Setting Short-Range Research Objectives. Research Management Swain, R.W.: "A General Model for Hospital Census Prediction and Control." Abstract of a Paper Presented at the Operations Research Society Convention Van de Van, A., and A.L. Delbecq: "Nominal vs, Interacting Group Processes for Committee Decision-Making." Academy of Management Journal Van de Van, A.H. and Delbecq, A.: "The Nominal Group As A Research Instrument For Laboratory Health Studies." AJPH Yett, D.E. Drabek, L., Intriligator, M.D., et al.: "A Microsimulation Model of the Health Care System in the U.S." CHAPTER III. ADDITIONAL REFERENCES REVIEWED xi PAGE 81 8 2 33 8 4 85 8 7 88 90 IL 92 93 94 INTRODUCTION To judge the appropriateness of priority setting methods for HSA purposes, one needs two sets of criteria. The first assists the evaluation of analytical methods used primarily by agency staff. The second helps evaluate methods for goal and priority establishment used primarily by agency boards. The set of analytical methods would be used by an HSA staff in analyzing projects and/or in appraising goals and objectives being considered for the areawide health plan. The objectives of these methods are evaluative in nature. Applied to a proposal under consideration, they seek to ascertain its major impact (s), to de- fine the planning issues which it raises, to clarity its implica- tions, to trace out its consequences and, generally, to provide HSA decision-makers with an "inhouse" appraisal of the proposal's merits as compared with other possibilities. The second set of methods would be used by HSA governing boards in achieving a concensus on the relative importance of various proposals before them. The objectives of this set of methods are consentaneous in nature. Applied in a situation where there are a number of competing proposals under consideration, they seek to assure a fair hearing for all serious points of view, to help keep group discussion centered on productive topics, to assure all points of view a fair weighting and assure that the group's decisions express an appropriate compromise among the divergent values and objectives of its members. There is likely to be overlap between these two types of methods. Most problems in social planning do not decompose neatly into mutually exclusive "analytical" and "goal-oriented" components. As a result, there is substantial overlap among criteria for appraising the two types of methods. Keeping this in mind, we pre- sent three interdependent sets of criteria. The first set applies to both sets of methods, the second primarily to analytical methods, and the third mostly to goal-oriented methods. CHAPTER I--ANALYTICAL APPROACHES Primary Emphasis on Concepts Archibald, K.A.: Three Views of the Expert's Role in Policy- making: Systems Analysis, Incrementalism, and the Clinical Approach. Policy Sciences. (1) 1970, pp. 73-86. This paper compares three approaches to improved policy making and planning: systems analysis, incrementalism, and clinical. By describing characteristics of each approach, the author attempts to combine the potential contributions of each into an effective synthesis in order to strengthen policy analysis. Archibald gives detailed attention to the three approaches for conceptualization of organizational structure and expert role playing. She focuses primarily on the comparison between system analysis and the clinical approach, suggesting that incrementalism is basically compatible with systems analysis. The author further discusses the implications of the differences between the two approaches, the potential con- tributions of the clinical approach for improving policy analysis and a likelihood of a synthesis between the two based on their compatible elements. Bell, D.: Twelve Modes of Prediction -- A Preliminary Sorting. Daedalus: _The Proceedings of the American Academy of Arts and Sciences, 93:845-73, Summer 1964. The objective of this paper is to categorize and evaluate the strengths and weaknesses of various approaches to predicting the future in a social science context. It is not specifically concerned with the prediction of health service needs or demands. However, a number of the approaches discussed could be adopted to the health service planning context. The twelve modes of prediction, culled from the literature in a variety of social sciences, include the following: (1) - the search for "laws of social physics", (2) the specification of trends and their extrapolation as forecasts, (3) the appraisal of "structural certainties" in organizations, communities, and nations, (4) the breaking of "operational code" for these entities, (5) defining the organization or community operational system, (6) defining the "over-riding problem", (7) isolating and judging the strength of the "prime mover", (8) tracing the "sequences in development", (9) devising appropriate accounting schemes, and (10) sketching scenarios and other alternative futures. The author's basic method is logical analysis. His basic purpose is evaluative; for each mode, an assessment of practical usefulness is offered. Depending upon the purposes for which predictions are needed, he notes, each mode of prediction proffers a mixture of strengths and weaknesses. Choosing the "proper" mode involves both specifying clearly the purposes which are to be served, and acknowledging (and compensating for) concommitant limitations. Taken together, the twelve modes employ a wide range of data: from expert opinion to "synthetic data" manipulated by formal models; from empirical data derived experimentally to intuitive perceptions deduced from "collective experience". Discussion of each mode includes the author's appraisal of the prospects for success, i.e. for achieving the predictive accuracy inherent in each mode. Berki, S.E.: "Demand and Utilization," Chapter of 6 of Hospital Economics. Lexington, Massachusetts: D.C. Heath and Company, 1972, pp. 121-165. The objective of this chapter is to present a theoretical approach to demand analysis which stresses the complex nature of demands for medical care. Need is defined in terms of consumer perspective; that is, need is that which the consumer perceives is necessary to attain his expected status of health, when it is in a state of disequi- librium. It is recognition of this disequilibrium which origi- nates the demand for health services. Demand is determined by a number of elements and factors which include the initiator of the medical care process, the mechanism of initiation, and the rele-, vant social and economic variables which may have played a role in the initiation process. A major portion of the chapter is devoted to describing the failure of many demand studies to consider all the elements of demand, thus making them relevant to only a subset of the total demand. Examples of such study approaches are price and income elasticity, utilization patterns and demand and utilization patterns in single practice settings. The difficulties and the trends in correcting the shortcomings of these approaches are also discussed. The article is most directly concerned with demand in the hospital setting, although alternative settings are discussed as a means of considering the appropriateness of the demand. $ The bibliography which appears in the book, Hospital Economics, 1S extensive. Boulding, K. E.: The Concept of Need for Health Services. Milbank Memorial Fund Quarterly. Part II:44(4) :202-220, October 1966. This article is an overview of the concept of need, laying particular stress upon the way in which needs relate to the econo- omists' concept of demand. Among the reasons for the importance of the concept of need is the weakness of the concept of demand in the area of health services. For example, the demander of health services is usually seeking knowledge from the professional, because he is usually a layman with incomplete knowledge in this area. Professional evaluation of need, in turn, depends upon application of a definition of "the state of health" to the lay- man. Maintenance of this state of health suggest that "certain minimum mechanical, chemical, biological, physiological, even economic and sociological requirements exist for the functioning of any organism or organization." These needs can be divided into those which can be taken care of by the organism itself and those which require professional intervention. The phenomenon of aging, however, creates additional needs for the organism. The problem of maintenance of the organism becomes very delicate. "Should the medical profession devote a relatively large proportion of its resources, as it does now,. in keeping miserable and senile elderly people alive, when their capital value even to themselves has become negative?" At the other end of human life is the problem of the rights and needs of the unborn. The need and state of health of the whole society must be considered. In both of these areas, medicine has con- sidered health in terms of inputs to an individual, not to a society, and has not addressed the possibility of acute conflict between the two. A further question regarding maintenance of the state of "health" is that of the state in which the organism is to be main- tained. A wide range of conditions exists between the "ideal" state and the "functioning" state, including physical health and mental health. And the concept of ill health might conceivably be applied to moral and political ideas, themselves moral judgements. In addressing societal health, one must further analyze the means by which societies provide for the "medically indigent" -- those whose "income is not large enough to provide a demand for the minimum medical care which a society, or a profession, identifies as need." Research into health -- rather than illness -- is encouraged, as is recognition of the necessity of avoiding research strictly on medical care needs without tying them to the system: "If the question is asked, how does one use a combination of the grants economy and the price structure in producing a system of medical care that compromises between needs and demands, a much richer and more satisfactory answer will likely result than if one simply asks what is the need for medical care?" Dunlop, D. W.: The Development of An Output Concept For Analysis of Curative Health Services. Social Science and Medicine. 6: 373-85, 1972 > This paper focuses on a major theoretical issue in medical economics: the conception of output employed in analyses of health services, especially as done by area-wide planners. It attempts to develop a concept of output, particularly for curative services, which is susceptible to quantitative estimation. Homogeneous concepts, such as patient days, are viewed as inadequate. Convenient measures, such as the number of trained manpower or beds per unit of population are discarded as analytically unsound. It is suggested that in order to improve upon the present conceptualization, it is important to remember that the output of health services is not homogeneous. Rather, it consists of a set of individuals who have (a) contracted one or more given diseases, (b) have received various diagnostic and treatment services specific to the contracted disease, and (c) have responded in varying degrees to the services received. The net output for a health services firm (facility) as derived by the author is stated symbolically in the following terms: Where: P = output of a curative health services production unit (i.e., hospital, health center, or dispensary) v = the number of persons seeking and securing at least one health service during a specified period of time v= the number of persons provided with a curative health service who are transferred to another health service facility v,= the number of deaths of v v.. = the number of persons who were treated but were unable to resume their major activity subsequent to treatment The author also reviews recent literature on several associated issues: (a) the necessity for a curative services output measure, (b) past efforts to develop such a measure, and (c) the issue of how to express non-homogeneous output successfully. A brief note discusses an analysis of health services in Uganda, in which the output concept presented here was employed. Etzioni, Amitai.: "On the Process of Making Decisions." Science. May 6, 1966, pp. 746-757. This article is a book review of The Intelligence of Democracy: Decision Making Through Adjustment by Charles E. Lindbloom. Lindbloom explores the relationship between policy formulation through the decision-making process and the democratic value of equality. He attacks widely held concepts about "ration- al decision-making processes" as neither feasible or desirable. Lindbloom believes that due totime and cost limitations, people cannot separate goals from means, values from facts or gain all the information they need to judge rationally. These overriding limitations should be acknowledged in official and professional Liturgy. Lindbloom further maintains that decision-makers in democracies do not follow a pattern of rational decision-making, but one of "disjointed incrementalism." This requires the decision-maker to solve problems by experimenting with policies only marginally different from the existing ones, gradually widening the scope of experimentation until the problem is solved, or until exhaustion sets in. He does this instead of pursuing "goals" in some abstract sense. The key to incrementalism, says Lindbloom, is the democratic values served by this pattern of decision-making. A kind of pluralism shapes and influences the public decision-making process in democracies. Therefore, Lindbloom argues, "good decisions" are those based on consensus. Squeaky pluralists should get public grease. 10 Feldstein, P. J.;: Research on the Demand for Health Services. Milbank Memorial Fund Quarterly, Part II, 44 (3) :128-57, July I966. This article provides a framework for the factors affecting demand for medical care and its components. A detailed bibli- ography at the end of the article provides references for further explanation of the issues presented. The author discussed the differences between "need" and "demand" for health services: "Need is the amount of care believed necessary by medical authorities while demand is the actual use of medical care services." Changes in utilization of (demand for) health services may be explained by any of several factors: ® changes in how services have been defined and how their use has been measured over time = a change in the supply of care + a real change in the demand for care o a combination of the above. Specification and measurement of the various factors whose inter- action results in health service "demand", thus, is a complex problem. In the course of the article, definitions of demand and in- fluencing variable are presented. In addition, a framework for analyzing demand for health services -- basically an economic framework -- is presented. Assumptions underlying the economic approach are described, including the role of choice, and factors affecting the patient's demand -- including incidnece of illness, cultural-demographic characteristics, and economic factors -- are discussed. Finally, factors affecting the physician's use of the health care components are included in the picture. 11 Goldsmith, Seth B.: "The Status Of Health Status Indicators." Health Services Reports. March, 1972, pp. 212-221. Measures of "community health status" are frequently used by physicians and other health professionals in their decision- making processes. Due to potential importance of these indica- tors, a review of the status of health status indicators was undertaken by the School of Public Health and Tropical Medicine at Tulane University. The review centers on the concept of health, the purpose and problems behind the use of indicators, and the outlook for health status indicators. Definition of "health" and clarifica-, tion of the purposes of health status indicators -- their validity, reliability, data sources and cost -- are main problems encountered in their use by physicians and health planners. The author discusses these problems in connection with examples of traditionally used indicators such as mortality and morbidity indices. A list of general evaluation criteria on which the merits of health status indicators can be judged is also suggested in the review. The limitations of this article from the health planner's perspective are: (i) it is concerned with only general community health status indicators; (ii) all indicators incorporate some degree of value judgement which the evaluation criteria do not take into account; and, (iii) conceptual problems exist in the development of health status indicators which should be recognized by physicians and health-care planners alike. The author concludes that action regarding the adaptation of present health status indicators in the decision-making process rests on the professionals' faith in the reliability and validity of selected indicators as conceptually sound and pragmatically acceptable. 12 Grossman, M.: On the Concept of Health Capital and the Demand for Health. Journal of Political Economy 80:223-255, March- April 1972. The aim of this study is to construct a model of the demand for the commodity "good health". It is in part based upon the notion that individuals invest in themselves. The costs of the investments include direct outlays on market goods and the oppor- tunity cost of the time that must be withdrawn from competing uses. The model argues that "a person's stock of knowledge affects his market and nonmarket productivity, while his stock of health determines the total amount of time he can spend producing money earnings and commodities." A second basis of the model is the realization that "what consumers demand when they purchase medical services are not these services per se, but, rather 'good health.'" "It is assumed that individuals inherit an initial stock of health that depreciates over time ... and can be increased by investment ... Health is demanded by consumers for two reasons. As a consumption commodity, it directly enters their preference functions, or, put differently, sick days are a source of disa- bility. As an investment commodity, it determines the total amount of time available for market and nonmarket activities. In other words, an increase in the stock of health reduces the time lost from these activities ... The analysis in this paper stresses that the shadow price of health depends on many other variables besides the price of medical care ... It is shown that the shadow price rises over the life cycle and falls with education if more edu- cated people are more 'efficient producers of health..." The author's mathematical model shows that "health can be viewed as a durable capital stock that produces an output of healthy time. A person determines his optimal stock of health capital at any age by equating the marginal efficiency of this capital to its user in terms of the price of gross investment." 13 Maass, Arthur A.: "Benefit-Cost Analysis-, Its Relevancy to Public Investment Decisions." Quarterly Journal of Economics XXIX (May, 1966) Can cost-benefit analysis be effectively used in analyzing public investment programs? This article deals with this question by examining the trade-off between economic efficiency and other objectives. The proposal of this article is to make use of poli- tical institutions to measure trade-off ratios between a market- determined efficiency and non-efficiency objectives, such as income distribution, in government programs. The public investment decision process should be organized to play to the strengths of political institutions rather than to its weaknesses. The capacity of the legislative process to select trade-off ratios useful in the design and implementation of public projects and programs is examined. The trade-off ratios provide the means to set standards for public investments. Under the heading of public policy, the definiton of cost- benefit analysis is the maximization of the contributions of a project to a national goal, instead of merely discovering the most efficient method of satisfying a fixed requirement. In determining public policy, this article challenges certain basic assumptions of cost-benefit analysis of public programs in terms of market prices: (1) non efficiency considerations are not taken into account. (2) efficiency is based on the premise that price ought to equal marginal costs. This is not relevant to public investment decisions because the market price only shows an aggregate of individual preferences which are not necessarily synonomous with "community interest." (3) «in most programs, particularly health planning, the main objective does not center around economic efficiency. (4) there is a need to measure costs in terms of resource displacement instead of pure market price. (5) the relation of benefits to costs is not necessarily an indicator of the project's worth. 14 Newhouse,: J. P., C. E.: On Having Your Cake and Eating It Too: Econometric Problems In Estimating The Demand For Health Services. The Rand Corporation. R-1l149-NC, April 1974 This report discusses a series of methodological problems common to many estimates of the elasticity of demand for medical services. These problems cause the estimates to be inconsistent. Where possible, the report derives a priori the direction of the inconsistency. It also reports estimates of the magnitude of the inconsistency caused by certain of the problems. These estimates use the 1963 Center for Health Administration Studies National Opinion Research Center (CHAS-NORC) survey. Seven problems are discussed. The first problem occurs with the use of the average price rather than the marginal price. If marginal price is the correct explanatory variable, it is shown in an appendix that use of the average price results in inconsistent estimates of price elasticities. The second problem comes from inferring an elasticity from the gross price variable using a specification with the percent of a population insured (or a zero-one dummy for individuals) plus a gross medical price variable. There is then an error in the health insurance variable, and there is a non-zero covariance between the true coinsurance rate and the gross price. This, in general, causes inconsistency. The direction of the incon- sistency cannot be signed a priori. A third problem occurs when price is estimated by dividing expenditure by quantity, when quantity is measured with error. This specification will result in overstating expenditure demand elasticities. The magnitude of the overstatement depends upon the size of the error in measuring quantity. No empirical estimates of the magnitude of error are made. A brief discussion of the estimation techniques is included, because the studies in the literature frequently do not use the best techniques. The fifth problem comes from aggregating across services and using an average coinsurance rate as an explanatory variable. Because some services are better covered than others, and because covered services tend to be associated with larger expenditures, an average coinsurance rate will show a spuriously high effect on expenditure. Combining hospital and physician services, the elasticity from using the average coinsurance rate appears over- stated by nearly an order of magnitude in the 1963 survey data. 15 Aggregation across services or across individuals, a sixth problem, causes potential instability (that is, the estimated coefficients will not predict future expenditure well). One of the usual sufficient conditions for stability of predicted values -- that the distribution of insurance remains the same -- could not reasonably be expected to hold, either across services or across individuals, if new health financing legislation were enacted. The other usual sufficient condition for stability of prediction, identical responses of the components of the aggre- gate, expirically does not hold across services; and no evidence has been presented that it holds across individuals. There is thus a presumption against using aggregate data. The final problem is omission of a cross-price variable. The direction of the inconsistency caused by this omission is away from zero. Empirically, elasticities appear to be over- stated by a very small amount. 16 Plessas, Demetrius J., Fein, Ricca.: "An Evaluation of Social Indicators." AIP Journal .; January 1972, pp. 43-51. The main objective of social indicators is to provide decision makers with qualitative and quantitative time series data enabling them to view domestic problems in a more emphat- ic precise, empirical, trend-oriented light. This definition rejects the prevailing reliance on purely quantitative economic data in public decision-making processes. Major difficulty arises in separating social indicators from economic indicators, however, for reasons the authors discuss. The expected outputs of developing social indicators are: 1} To help ensure more effective policy implementation; 2) To provide better information for making program choices; and, 3) To ensure that the allocation of scarce public resources is directed more effectively among social programs. This article examines these expected outputs in view of the major conceptional, operational and statistical problems in the construction and use of social indicators in rational decision making processes. The authors explicitly state that although social indicators may be a useful means for demonstrating causality among social phenomena, their use as tools for priority-setting is invalid. Effective priority-setting calls for sound evaluation of programs through the use of systems designs and controls. Indi- cators are not sufficient in themselves due to the fact that more effective methods are necessary to "disentangle the causal relationships among indicators." Indicators are more valuable in the area of descriptive reporting, analytical studies of social change and prediction of future needs. Social indicators for use in determining policy and program planning is a politically charged issue. Policy planners and social scientists are often at odds as to how social indicators should be used in planning and to what extent. Conflicts between competing socio-economic groups make it difficult for planners to distinguish between interrelated social indicators in choosing those relevant to a particular 17 program. Therefore, social indicators although plausible are not politically feasible in providing a basis on which to formulate rational decisions. The potential benefits of social indicators as a means towards rational decision-making do not justify the cost of achieving them, in the authors' view. The potential conflicts arising from their use, the opera- tional and statistical problems in their implementation make social indicators a weak method on which to base and implement rational policy and program planning decisions, at present. 18 Prest, A.R., and R. Turvey.s "Cost-Benefit Analysis: A Survey." The Economic Journal. 75 December, 1965. This survey of cost-benefit analysis explains its appropriate uses and limitations, the framework in which it operates, and its specific application to various fields including water projects, transportation, land usage and health. Cost-benefit analysis (i.e., to enumerate and evaluate all relevant costs and benefits of a project) draws from a variety of economics: welfare economics, public finance, price theory, and resource economics. The analysis takes into account economic choices involved in investment projects based on a given set of prices. Secondary or indirect benefits including intangibles play a large role in determining these economic choices, however, changes in the price of goods and other factors are not taken into account. In the health field, cost-benefit analysis concentrates on the problem of assigning values to the benefits of lives saved or illness avoided. Economic value of a life saved varies accor- ding to a variety of factors. Under cost-benefit analysis, the value or benefit of a life saved is most commonly viewed in terms of the loss of production of that life through death or illness. Cost-benefit analysis, note the authors, operates under the constraints of physical, legal and administrative costs, as well as uncertainty.and budgetary restraints., The main limitations of cost-benefit analysis, find the authors, are as follows: (1) uncertainty and unreliability of costs resulting in the complex nature of benefits; (2) its usefulness in terms of its limits in principle and practice; (3) - statistical deficiencies. The advantages of the method are general in nature compared to its limitations. The method: (1) forces decision makers to quantify costs and benefits, rather than qualify them with personal judgements; (ii) causes programs and policies to be questioned that would otherwise not come into question; (iii) screens out inferior projects; (iv) is more useful in public- utility than in the social service area of government. The article concludes that valuation of benefits poses greater problems than do costs. Enumeration is difficult due to diverse types of benefits. Utility is almost impossible to measure due to the different values and levels of satisfaction obtained by individuals, thereby making evaluation of benefits difficult. 19 Quade, E.S.: On The Limitations of Quantitative Analysis. The Rand Corporation, p.p. 4530, December, 1970. This paper discusses the nature and limitations of quantitative analysis with regard to decision-making and policy formation. Quade questions the validity of the quantitative approach in obtaining solutions to all problems for which there exist alternative courses of action. Dangers arise in this approach when the analyst fails to recognize that quantitative analysis may not be suitable for solving a particular problem, or that it focuses attention on efficiency rather than on goals. Due to the inherently subjective nature of decision- making, the author suggest that solutions arising from a more subjective or political process might prove less difficult to implement and more acceptable to the general public and decision-makers alike. The external limitations of quantitative analysis stem from existing problems which do not lend themselves to a quantitatively derived solution. Internal limitations result from the extent to which analysis is purely quantitative. Quade describes other methods of problem-solving such as: operational gaming and Delphi which are partially quanti- tative in technique, but have other added features which perhaps make them more suitable in certain decision-making situations. The author concludes with the statment, "Of the major issues facing the nation today, it is hard to think of any that can be resolved purely by what we know. Analysis, consequently, should be viewed more as a method for investigating problems than for solving them." 20 Reiner and Davidoff.» - "Choice Theory of Planning." JAIP. May 1962, pp. 102-115. The Choice Theory of Planning rests on the assumption that man controls his destiny and has the ability to exercise sound judgement in the formulation of rational decisions. The frame-, work in which this theory is set, is by nature general - apply- ing to a variety of fields regardless of the substantive or geographical focus. In essence, this article presents a theo- retical framework of "ground rules" against which methods of planning can be tested to determine their validity. Three basic objectives form the foundation on which the theoretical framework is built: 1) Efficiency and rational action--utilization of re- sources measured in terms of the purpose it serves. 2) Market aid and replacement--planning as an act in testing value alternatives and extending to sugges- ting courses of action towards change. 3) Change or widening choice. The third objective is of particular interest in that it deals directly with how individual or collective choices are made re- lative to the available resources. The axiom of scarcity clearly points out the problem of priority-setting. The environment in which planning takes place contains a scarcity of resources. Factors contributing to the production of goods and services are limited in supply which in turn limits output. In short, it is impossible to achieve everything desired or needed at any one time. Scarcity by its definition necessitates choice. The ultimate objective of planning is to widen the range choice by offering value alternatives from which to choose. Therefore, contrary to the law of scarcity and given the willing- ness and ability to meet their costs, planning should operate under the assumption that all objectives are possible to attain. Choices are made through a process of value formation. A value hierarchy, from which priority-setting stems, is a structure of defined levels in which inconsistancies among values are defined, reduced or eliminated in order to achieve the out-, put of certain goals. Value formation must be placed first in the planning process so that a set of goals can be determined and strived for. Thus, the tendency of a planner to merely use 21 the tools of survey and analysis is reduced. Another potential pitfall that must be guarded against is the conflict between de- mands for immediate action and non-arbitrary decision-making. The The planner must also be informed of the range of choices and their implications in terms of cost, data collection, outcome, and resources required. Lack of knowledge as well as techniques will hinder the planning process. The planner acts as an agent for his clients, assisting them in understanding the range of choices available. He is not responsible for transforming values into policy commitments. This limits possible arbitrary decisions and also the planners role in priority-setting. The general nature of this theory makes a specific formula-, tion of a method for priority setting impossible but does pro- vide the areawide planner with basic objectives and insights into the planning process. 22 Wiseman, Jack: "Cost-Benefit Analysis and Health Service Policy." The Scottish Journal Of Political Economy, 10 February 1963, pp., 128-145 The contribution of cost-benefit studies to policy formation depends on the degree to which the value system contained in the analysis agrees with the value system of the policy maker. The study must be formulated so that the relationship between the two value systems is dealt with. In this article, studies of investment in health services are analyzed on the basis of their contributions to the resolution of problems in the area of health needs. The studies use the general techniques of cost-benefit analysis for obtaining information such as: (i) efficient size of programs, (ii) optimal allocation of medical resources, (iii) efficient utilization of the resources. The author discusses the problems involved in determining resources used in health service investments and relating them to increases in the community output. The concern is mainly with the improvement of the decision-making process and the allocation of community resources to provide for adequate health care. The article outlines two general approaches to studies on policy for- mation -- the welfare approach and the direct evaluation approach. Welfare approach: Under the criteria of welfare economics, the policy-maker must adjust his result by taking into account the individual demand for health as a consumption good, instead of merely an investment. A policy adjusted in such a manner would take into account humanitarian concerns instead of purely market considerations. Cost benefit analysis is employed to provide criteria for identifying preferred public policy. Direct evaluation approach: This approach is directed towards arriving at optimal policy decisions. Based on the assumption that the health environment is an element in calculating the whole set of policy aims, the rate of return to the investment in health. services is related to the health environment. No satisfactory solution is available to distinguish between the contribution of health resources and the contribution made by other forms of investment, which makes it difficult to measure the rate of return. 23 mary Emphasis on Methods Abernathy, W. J. and Hershey, J. C.: A Spatial-Allocation Model for Regional Health-Services Planning. Operations Research 20 (3): 629-42, May-June 1972. This paper presents a model which is aimed at assisting the health-systems planner in determining the location and capacities of a specific number of primary health care centers within a region. The authors have placed emphasis on providing a formulation that is a useful characterization of the region. The formulation distinguishes the separate impact of distance on various socioeconomic strata, as well as upon overall utilization. Census statistics and regional utilization data are the source data used in this model. The population is stratified by characteristics which allow each strata a relatively homogeneous pattern of utilization. The two principal parameters of utilization behavior which are concerned are (1) the number of primary-care visits per time period that can be expected when a primary-care center is in immediate proximity to the individual and (2) a measure of the extent to which propensity to seek care decreases as distance to a primary-care facility increases. The criteria built into the example model were the following planning objectives: (1) maximize utilization, (2) minimize distance per capita, (3) minimize distance per visit, and (4) minimize percent degradation in utilization. The Hookes- Jeeves algorithm is used to solve the model with respect to the first three criteria. The model provides the planner with "an explicit statement of information requirements, a mechanism for examining the incremental gains to be expected from adding additional centers, and a means to explore the differential effects of alternative criteria upon solutions." An example of the use of the model is given in the article. 24 Alonso, W.: Predicting Best with Imperfect Data. Journal of the American Institute of Planners 34:248-255, July 1968. The objective of this article is to distinguish two types of error inherent in most data which must be used by area-wide planners and to suggest some rules for coping with such errors when there is no alternative to using error prone data when forecasting. Two common types of error are those of measurement and those of specification. Errors of measurement arise from in- accuracy in assessing a magnitude. Errors of specification arise from a misunderstanding or purposeful simplification in the model of the phenomenon one is trying to represent. The author observes that anyone who plans uses a predic- tion model of some kind. However, the most elaborate and in- clusive predictive model is not necessarily the best for applied work. This may be the case if the cumulation of data errors exceeds the predictive gain from superior specification, a stage which some of the more ambitious forecasting models may have reached. A rule of thumb for choosing and building models which must use error prone data is suggested: When the complication of the model leads to negative returns, a strategy of "netting out" simple, complementary models may be better. In general, poorer data require simpler models. Elaborate models which are poor predictors may serve as useful contexts for partial models and may achieve their full worth if maintained and improved over time. Further, even a poor predictor may contribute to scientific knowledge and to the understanding of processes, and thus can be helpful for making decisions. Several rules of thumb are derived for minimizing the ill effects of potentially erroneous data when forecasting. These include the following: (1) Avoid use of intercorrelated variables in models. (2) Employ models which rely as heavily upon addition as possible. (3) -If one cannot add, multiply or divide. (4) Avoid models which depend upon taking differences or raising variables to powers. 25 (5) Avoid use of models which proceed by chains of quantitative reasoning; the longer the chain, the greater the need to avoid. 26 Anderson, J.G.: Demographic Factors Affecting Health Services Utilization: A Causal Model. Medical Care 11(2); 104-20, March-April 1973. The purpose of this study is to develop a model for predicting the future demand for health services (short-term general hospitals) from changes in various demographic factors of the population. Demand is implicitly defined in terms of patient days, admission rates, and the average length of stay. "A causal model has been developed relating patient days per thousand population and its components, hospital admission rates and average length of stay, to demographic characteristics of New Mexico counties. Data from the U.S. Census 1960, the New Mexico State Department of Business Research, and the Annual Guide Issue of 'Hospitals',...have been used to estimate parameters of this model" through simultaneous regression analysis. "The model makes explicit the...manner in which changes in demographic characteris- tics of the population may directly and indirectly affect hospital use." The study supports the conclusion that the hospital bed supply in an area is a major determinant of hospital utilization in the area. In fact, "changes in supply lead to significant changes in demand since both the admission rate and average length of stay increase, resulting in a large increase in the number of patient days..." This study also revealed that "socio-economic factors such as income level, educational level, and ethnic composition have...little effect on the use of hospital facilities..." but utilization rates are affected by age and degree of urbanization. The study hints, however, that the conclusion that bed-supply affects admission rates and average length of stay may apply only in those areas "...where alternative health services are lacking or inadequate..." 27 Anderson, J. G. and Bartkus, D. E.: Choice of Medical Care. A Behavioral Model of Health and Illness Behavior. 14:348-62, December, 1973. A desire to account for social-psychological variables as determinants of differential patterns of illness behavior and medical care utilization prompted the authors to develop the model described herein. The models (which differ by sex) are structural equation models which are analyzed using regression analysis techniques. The dependent variables included in the model are socio- demographic, economic, ecological, need, and social psychologi- cal. These have been selected from and are justified on the basis of previously performed studies, which are cited in the text. Need as represented by the proxy variable age is considered an important explanatory variable and is included within the socio-demographic factors (sex, age, and social class). In the model, the desire for medical care, as expressed in consumer decisions regarding choice among alternative health services, is estimated using a technique in which a score of total need for medical care is assigned by disease category, taking into account need for hospitalization, diagnostic services, medicine, and other care. The model, however, is basically a demand model, focusing upon actual utilization of services. The methodology is directed toward developing and testing the model. Students enrolled in a university health plan were used as data sources; 61 percent of the 946 sampled completed the questionnaire. The variables included in the models (which differed according to sex) were: Age in years, marital status, family physician, socioeconomic status, insurance, location, symptom sensitivity, medical need, other's attitude toward the student health center, utilization. The results of the study reveal that the model is useful for indicating indirect as well as direct effects of variables upon health services utilization. The article contains an extensive bibliography. 28 Anderson, R. and Benham, L.: Factors Affecting the Relationship Between Family Income and Medical Care Consumption, Empirical Studies in Health Economics. - Baltimore, Maryland. The Johns Hopkins Press, 1970. pp. 73-95. "The intent of this paper is to measure and assess the importance of some factors which may influence the relationship between family medical care consumption and family income." Demand is discussed in terms of medical and dental care consumption. Medical care consumption is measured by dollar expenditures and weighted "units" of use. Dental care is measured only by dollar expenditures. Five propositions based on the general hypothesis that "...apparent income elasticities of demand with respect to medical care may be altered significantly when other factors are taken into account" were "...tested for physician and dental care consumption using data from a national consumer survey for 1963. The findings supported three propositions unambiguously, supported a fourth for dental but not for physician expenditures, and lent no support to the remaining proposition...". The five propositions are quoted below: "1. The simple observed income elasticities for physician and dental expenditures will differ Slgnlflcantly from the corresponding elasticities obtained by holding constant price, quality, demographic characteristics and preventive care. 2. The estimates of elasticity of demand for both physician and dental expenditures will be higher with respect to permanent income than with respect to observed income. 3+ Illness simultaneously increases physician expenditures and produces a negative transitory component in family income. 4 . Permanent income elasticities for physician expenditures will increase when family level of illness is standardized. 5. Income elasticity estimates for physician services will be lowered by the substitution of quantity of services consumed for dollar expenditures in the analysis." 29 Anderson, R., Smebdy, B., Eklund, G.: Automatic Interaction Detector Program for Analyzing Health Survey Data. Health Services Research. 6:165-83, Summer 1971. The purpose of this article is to illustrate five uses for the Automatic Interaction Detector (AID) computer program in the analysis of health survey data. Applications are made to examples of increasing complexity. The "forced trees" technique is discussed as a solution to the problem of substitution effects, and results of the AID analysis are compared with those of regression analysis. The Automatic Interaction Detector (AID) computer program is focused on the data-analysis problems typical of many large-, scale social surveys which involve more than the reporting of - descriptive statistics but which may not involve the exact testing of specific hypotheses. The underlying questions the program seeks to answer are: Which explanatory variables and which combinations of these variables are important for reducing the variance in the dependent variables? How much of this variance can be explained by the use of these explanatory variables? The predictors of the AID program can be either classifi- catory variables or more precise scales of measurement. The dependent variable can be a continuum, an equal-interval scale, or a dichotomy. Linearity and additivity assumptions are not required, nor are the restrictions on the form of the predictors that are inherent in conventional regression techniques. The program subdivides the original sample, through a series of dichotomous splits with respect to the predictors, into a number of mutually exclusive subgroups. In the first step the predictor and that division of categories of the predictor are chosen, from all potential dichotomies, which maximize be- tween-group variance. In the second step the subgroup with the largest sums of squares, with respect to the dependent variable, is likewise divided, to maximize the variance explained. The analysis goes on with successive splits until minimum require-, ments concerning size and variance are no longer met by any of the subgroups of the sample. The algorithm for the AID program produces output that lends itself readily to graphic display which enable the reader to see for himself the main effects and interaction effects, an important pedagogic advantage of AID over more traditional multi- variance methods, particularly for readers of varying or limited statistical background. 39 A preliminary examination of some problems involved in the analysis of survey data on health services utilization suggests that the AID approach may have particular merit in this area. While health survey methodology has probably advanced as rapidly as the methodology of any other content area of the behavioral sciences, there has been a relative dearth of accompanying theory of sufficient sophistication to provide adequate guide- lines in the analysis of data collected. The inductive aspects of the AID program can be particularly helpful under these circumstances. Another potential advantage is the AID emphasis on inter- action effects. In utilization studies the illness or health-, level variables are of prime importance. It is well documented that the amount of illness people experience is a basic deter- minant of the medical care they receive. Given this initial importance of illness, the interesting question is: What are the subsequent factors that further divide the ill and the non- ill according to the medical care used? For example, how does the medical care utilization of the low-income sick population compare with that of other sick groups? The distribution of the dependent variables and their relations to some of the most important predictors in health survey analysis cause special problems in meeting the assumptions of regression analysis and may further enhance AID as a potential alternative. The familiar U-shaped relation between age and volume of medical care consumed is one example of the non-linear relationships that must be dealt with in analyzing health survey data. Even more complex relationships are sometimes found among such variables as family life-cycle stage, family income, and health services utilization. These concepts are developed through examination of five examples: (1) One sample, one dependent variable; (2) One sample, one dependent variable, multi-stage analysis; (3) One sample, several dependent variables, multi-stage analysis; (4) multi-sample, multi-stage analysis; and (5) the "forced tree" method . 31 Ault, D.E. and Johnson, E., Probabilistic Models Applied to Hospital Planning. - Growth and Change. 402) 7-13, April 1973. This article presents a model for approximating the geo- graphic service areas of hospitals, to accurately estimate demand for, and revenue generating capabilities of, proposed hospital facilities and to simultaneously determine patient days, occupancy rates and revenues generated by a new facility. Demand for health services is discussed in the paper in terms of number of admissions, patient days, and occupancy rates. The model used in the study is "... similar to probabi- listic gravity models used to determine market areas for re- tailing activity." The variables in the model are surrogate measures of two variables affecting patient admitting patterns: "... the disutility of distance and the uncertainty of the in- dividual to the type and quality of the services offered." The model uses data from the St. Louis SMSA and from a pre- dominantly rural area in west central Illinois. Using the model geographic service areas, "... concentric bands about each hospital complex within which the probabili- ties of- allo- residents utilizing that complex are equal," are generated. The model is also used to predict hospital ad- mission, patient days, and occupancy rates. The average error in predicted admissions compared to actual admissions was 7.91 percent, and 7.25 percent for predicted patient days compared to actual patient days. 2 2 Cooper, J. K., and Corcoron, T. M.: Estimating Bed Needs by Means of Queueing Theory. New England Journal of Medicine. 291(8): 404-5, 22 August 1974. The purpose of this study is to determine the number of acute and intermediate coronary care beds necessary for a major urban hospital in order that a bed in each unit would be available when needed a predetermined 95 percent of the times. The basic model is that of the multiple server queueing model of operations re- search theory. The study assumes from 400 to 800 suspected cases of myo- cardial infarction per year based on the participating hospital's suggested guidelines of 600 such cases per year. While the problem resembles the standard multiple-server queue, ing model, it is complicated by the restriction that the patients in the first (coronary-care) unit are of two kinds (true and false myocardial infarction), each kind staying a different length of time, and that a second service (intermediate-care facility) unit receive a certain portion of the patients discharged from the first unit, because of the mortality rate for true infarctions. Accordingly, the methodology used is a triangular queueing system with two types of customers, those with true and false myocardial infarction. It assumes that both interarrival time between patients and service distribution times are exponentially distributed random variables. By using differential-difference equation methods (not described), seven linear equations describe the probability of the system being in each state. In the author's view, this methodology has general applicability and can be applied in any medical situation in which there is one service unit or two units serving patients in sequence. In particular, it should be useful in planning a progressive coronary care unit. The analysis represents a concrete example of the potential value of the operations research methods theory to health care planning. 33 Cordle, F., and Tyroler, H. A.: The Use of Hospital Medical Records for Epidemiologic Research. I. Differences in Hospital Utilization and In-Hospital Mortality by Age- Race-Sex-Place of Residence and Socioeconomic Status in a Defined Community Population. Medical Care. - 12(7) ;: 596-610, July 1974. The survey effort presented in this article is "intended for the use of measuring hospital utilization patterns at any level below that of a region." Citing the rapid changes in medicine and community public health, the author stresses the value of local utilization statistics for facility, manpower and service planning to supplement National Health Survey statistics. The study was conducted in Charlestown County, South Carolina. The data used are the 1960 United States Census and the results of a hospital survey conducted in 1963 covering all county hospital admissions for that year. Specifically the hospital survey includes hospital numbers, area of residence, age, marital status, height, weight, blood pressure, admitting and final diagnosis, and status at discharge. A subgroup of this sample is "further divided by place of residence (rural, urban)... and socioeconomic group" in a manner similar to the total county population description of the census data. Two major data problems are considered in the survey: (1) the number of hospital admissions of county citizens outside the county, and (2) the time interval between the census data and the hospital survey. As to the first problem, a brief survey of hospitals in surrounding counties revealed only a small number of citizen hospital admissions outside the county. Upon analysis of the local hospital data, its findings are compared with national averages for the same categories of persons as found in the National Health Survey. 34 Davis, K. and Russell, L.B.: The Substitution of Hospital Outpatient Care for Inpatient Care. The Review of Economics and Statistics. 54(2): 109-20, May II77. This paper has several purposes. First, it investigates the properties of the demand for hospital outpatient care. Second, demand functions for inpatient and outpatient care are estimated by regression methods using 1969 data from 48 states. Third, implications of the inpatient and outpatient demand elasticities for total hospital care cost are discussed. The demand for hospital outpatient care is measured by the number of outpatient visits. "The demand for inpatient care is divided into two components: number of admissions and length of stay." ] Each of the three demand equations (outpatient visits, admissions, length of stay) contains the same economic and socio-demographic variables. "The substitution of outpatient care for inpatient care is captured in the model by two variables: the price of inpatient hospital care and the hospital occupancy rate." The substitution of outpatient care for physician care outside of the hospital is represented by the charge for a visit to the physician's office and by the available supply of primary care and specialist physicians engaged in nonhospital practice." One result of the outpatient visits log linear regression estimates is the indication that a decrease in outpatient price increases the visits demanded. The results also showed "... a clear substitution relation between inpatient and outpatient care." Outpatient demand was also found to be sensitive to inpatient occupancy rates, thus outpatient visits increased during high-occu- pancy periods. Among the results of the log linear regression estimates for inpatient care is the negative effect of occupancy rate on inpatient care demand. Outpatient price was found not to affect length of stay. Also, "a given percentage change in the cost of inpatient care has a greater effect on outpatient care than the same percentage change in the cost of ourpatient care has on the demand for in- patient admissions." 35 Feldstein, M. S.: Hospital Cost Inflation: A Study of Nonprofit Price Dynamics. The American Economic Review. 61(5) : 853-72, December 1971. This article presents an "empirically estimable model of the nonprofit hospital industry and uses it to analyze the problem of hospital cost inflation." The model contains twelve equations which seek to explain the behavior of the hospital industry. The equations are divided into four groups: (1) demand relations, (2) price adjustment, (3) components of cost, and (4) expansion of capacity. The demand and price adjustment equations are "estimated using a mixed cross section of time-series sample of ten annual observations for each state." The sample included a total of 470 observations. The variables considered as influencing the demand for admissions and the mean stay are: "price, income, availability of hospital facilities and alternative sources of care, the demographic composition of the population, and general attitudes toward hospital care." In studying demand, the relevant price variable is "cost to the patient net of insurance reimbursements and relative to the price of other consumer goods and services." As to the price adjustment feature of the model, Feldstein says an increase in demand permits the hospital to yield to the internal pressure for higher price methods of care without reducing the percentage occupancy of beds. The model and equations presented in the study depart from the usual explanation of hospital cost inflation. "Increases in the components of cost are seen as primarily the result and not the cause of higher prices. The source of inflation is the pressure of rising demand induced by increases in insurance coverage, personal incomes, the availability of hospital-oriented specialists, etc." The converse of the last named effect seemed to be indicated, also. That is, an increase in the number of general practitioners would induce a large saving in hospital resources. The research described in the article is stated to be continuing into a phase in which the equations dealing with components of cost and expansion of capacity will be estimated, and "the problem of quality change and its impact on consumer demand will be examined explicitly." 36 Feldstein, P.J., and German, J.J.: - Predicting Hospital Utilization: An Evaluation of Three Approaches. - Inquiry. 2(1):13-36, June 1965. The purpose of this article is to test and evaluate three methods of predicting future hospital utilization: (1) "the use of the trend in patient-day/population ratios; (2) the use of the bed/population ratios; and (3) the use of social, demo- graphic, and economic variables to predict future patient-day ratios." In this study, hospital utilization is measured in patient days per thousand population in non-Federal short-term general hospitals. The article includes three major steps. First, the three methods are defined and the basic assumptions of each are discussed. Second, following a brief discussion of model buiding and statistical measures for evaluating results, the alternative statistical models for each approach are presented. Four trend line extrapolation models, four bed supply extrapolation models, and two demand analysis models based on six factors: median family income, projection of population with hospital insu- rance, room rate, population age distribution, urban-rural setting, and race characteristics are developed. From the results of each statistical test, two measures were derived to enable comparison of each model to the otehrs: (1) the standard error of estimate and (2) the correlation coefficient. Third, each method is evaluated using the following criteria: "how well does each technique predict; how flexible is the technique in allowing changes to be incorporated into the model; and how useful is the technique for hospital planners?" The trend in patient days/1000 and the trend in beds/1000 possess greater predictive accuracy than the demand analysis approach. On the other hand, the demand analysis "... had the greatest ability to incorporate charges which may occur". Finally, "... the patient- day ratio would present the least number of problems to the hospital planner; while the demand analysis approach presented the most". 37 Cabrielson, I. W., Siken, E., Sohler, KL. B., and Stockwell, E. G.: Relating Health and Census Information for Health Planning. American Journal of Public Health. $9(7): 1169-76, July L969. This paper reports on experience with coordination of health information from various sources with the 1967 Census pretest invNew Haven, Connecticut. Although demand is not explicitly defined in the paper, it is viewed as the equivalent of utilization. Five types of data are selected for use in the health information system: birth records and matched fetal and infant death records; hospital obstetrical records; family health survey data; 100 percent basic census record; and 25 percent census record. A basic issue was to decide upon the optimal level of aggregation for health planning data. Possible levels: (1) individual; (2) household; (3) housing unit; (4) block face; (5) block; (6) enumeration district or traffic zone; (7) quarter tract; (8) census tract; (9) town; (10) SMSA. The relative strengths of each level of aggregation are briefly discussed. All data in the reported study are aggregated at the quarter tract level in the city of New Haven and at the tract level in the balance of the SMSA. The family health survey, designed to "determine three major areas of concern -- medical care, day care and family planning -- consists of a questionnaire regarding each of these topics." It was sent to 7,449 households selected by stratified, random sampling techniques. The actual linking of the data from the five various sources is accomplished by computer mapping techniques. Two basic types of mapping are shown in the paper: The first is a plot of discrete events; the second displays frequencies within predetermined ranges. The method exemplifying the former type is the Dual Independent Map Encoding (DIME) file. The method described which is of the second type is the Synagraphic Mapping (SYMAP) technique. 38 Grimes, R.M., Allen, C.L., and Sparling, T;iR.: Use of Decision Theory in Regional Planning. Health Services Research. 7173-78, 1974. The article reports the methodology used by the authors, who are associated with the University of Texas, in assisting the members of the Houston-Galveston Area Coulcil (HGAC) deter- mine the number of hospital beds to be built in the 13-county region over the next seven years. The article describes statis-, tical analyses which can be performed on probable projections which present alternatives to individuals as a basis for decision making. The costs of the alternative decisions are illustrated: too few beds would imply inadequate care for the sick; an exces- sive number of beds would mean substantial burden on the tax-, payer in support of unused beds. Need and demand concepts are defined only indirectly by their relationship to current and projected utilization. Methodology for the project was initially influenced by a framework established by HGAC: need projections were to be for seven years (1974-1980) for the 1l3-county service area using projections based on hospital days/population/year using an 85 percent occupancy factor and stated trends in utilization rates. The authors then prepared four basic tables to assist HGAC in the decision making process. A sample decision matrix illustrates what the potential shortage or excess of beds would be if one decision was made and another turned out to be correct. A chart shows the dollar costs of excess beds, using the least squares method ($202,000 per year per bed for 1974-1980 period). This indicates the potential costs of various decisions. To make this matrix more precise, the expected values technique was used which relied upon equal probabilities of occurrence of the outcomes. The probabilities of the occurrence are weighted by the length of time on which the projections were based. By using marginal analysis, which showed that even a modest reduction in expected bed shortages would have a high economic cost, the HGAC selected Decision I as the alternative to be followed. This means that "unless use rates rise rapidly, no hospital will receive a certificate of need to construct short term beds in Houston before 1980, assuming no existing beds are removed from service." 39 Hill, Morris.: "A Goal-Achievement Matrix for Evaluating Alter- native Plans." JAIP. September 1968, pp. 19-29. This paper examines cost-benefit analysis - its application as an established technique of plan evaluation in the private and public sectors. After ferreting out their shortcomings, it offers the goal-achievement matrix as an alternative method of plan evaluation. In the private sector, each project is evaluated on the basis of revenues and costs in order to choose the profit maximizing project. In the public sector, cost-benefit analysis attempts to achieve the efficient allocation of resources by a public agency without resort to market pricing. However, certain problems arise when using cost-benefit analysis in the public sector. Public resources are limited and barriers to the flow of funds exist. The value of benefits and costs cannot be equated in terms of market prices. Cost- benefit analysis does not always take into account social along with monetary costs and benefits. The author offers the goal-achievement matrix method as an alternative to cost-benefit analysis in meeting the require- ments of program planning and evaluation. A brief outline of the method follows: A set of goals with relative weights attached to each is established. Objectives are defined operationally and the consequences of each alternative course of action are determined. Costs and benefits of each action are measured in terms of the achievement of each goal established. In this manner, the effect of a plan on the achievement of a set of objectives can be determined. Another approach of the goal achievement method is to weigh the objectives and their incidence in order to measure the achievement of an objective. The goal-achievement matrix method is more suitable for ranking courses of action than for testing the desirability of a project. However, the disadvantage lies in the inability of the method to register interaction and interdependence between objectives. 40 Kalimo, E., Sievers, K.: The Need for Medical Care: Estimation On The Basis of Interview Data. Medical Care. VI (1): 1-13, January-February, 1968 In 1964, the Research Institute for Social Security of the National Pensions Institute of Finland organized a program for collection of health interview material. The purpose of the information to be gained was to describe morbidity and the utilization of health services in Finland before the establish-, ment of national health insurance. Further, the survey was aimed at measuring the need for utilization of medical services created by perceived morbidity. The questionaire includes questions concerning the inter- viewee's state of health from which qualitative data are ob- tained on the interviewee's morbidity. The data on the nature of the diseases and disorders found are coded in accordance with a classification of diseases which included diagnostic categories prepared expressly for this survey on the basis of the International Classification of Diseases. The article outlines a method for converting the qualitative data yielded by the classification of diseases into quantitative variables describing the need for care for various disease states. It also presents results concerning the relative needs for care in various diagnostic categories and discusses the reliability and the validity of the method used. Finally, it described the average need for care in the Finish population interviewed. The need, say the authors, for an index suitable for ex- pressing the state of the population's health, and the difficulties involved in the formation of such an index have been emphasized by numerous students of the problem. The present survey reports the construction of such an index, how it was developed and applied to the results of an interview survey. Of the consequences of disease, only one aspect is taken into account: the need for medical care these consequences create. But, the authors con- tinue, Lt is just this aspect of morbidity which is the most relevant from the viewpoint of health policy. It is conceivable, it is concluded, that by applying similar rating methods to measuring other influences of illness, other useful information could be acquired which would describe various aspects of the population's health and changes taking place therein. The value of the index does not depend solely upon the perception of the respondents. 41 VKlee, A.J.: "Let DARE Make Your Solid Waste Decision." American City, 95(1), February 1970, pp. 100-103. ~ Decision Alternative Ration Evaluation (DARE) is a decision weighting model used to evaluate score and rank sets of competing alternatives. The alternatives are scored according to established decision criteria. Since factors are usually not equal in impor- tance, to assure that the model is accurately and fairly constructed, the scores are weighted according to their relative importance in the evaluation process. DARE operates under the assumption that the decision-maker has selected his criteria in such a way that independence among them is assured. The method is based on pair-wise comparisons but reduces the required number of such comparisons. Factor weights and subscores are determined by a procedure in which each alterna- tive is compared against the criteria, producing a rank ordering of alternatives. The method produces the "best" alternative but does not take into account the risk, profitability or utility of the alternative. 4 2 Kong-Kyun Ro: Interactions Among Variables Affecting Hospital Utilization. Health Services Research. 8(4): 298-308, Winter 1973. ’ The purpose of this article is to describe a method for improving predictions of the amount of services used by inpatients by using multiple regressions of explanatory demographic variables in all combinations rather than in terms of pair-wise comparisons. Demand is not explicitly defined in the article, but is equivalent to utilization. Two indexes are used to measure hospital use: length of stay and a weighted number of special services provided. The indexes for hospital use are "disease-adjusted" by grouping all cases into 30 categories defined by the WHO International Classification of Diseases. Each of these categories is further divided into four categories: surgical or non-surgical, and single or multiple diagnosis. The data used in the study are a one-in-nine sample of 9,000 patients admitted to 22 short-term general hospitals in the Pittsburgh area. The demographic variables of age, sex, and race are analyzed to determine their relationship to hospital use as defined above. This data classification scheme includes four age categories, two categories for sex and two categories for race. As opposed to the conventional cross-tabulation analysis, and the use of dummy variables in a multiple regression with the usual constraint that one of the variables be omitted, the methodology presented allows a systematic search for possible interaction among all 16 possible combinations of the demographic variables considered. Additionally, says the author, "Since the method uses a typical regression program, the method can easily be adapted to existing computer programs". Use of this method is especially valuable, the author observes, when one suspects interactions among cate-, gorical variables (i.e. in this case age, sex, or race). 43 Leibenstein, H.: Allocative Efficience Vs. X-Efficiency. American Economic Review 1966, pp. 392-415. Making use of numerous studies on the allocation of resources, the author discusses the concept of efficiency. He argues that allocative efficiency is used to the exclusion of other types of efficiency (called X-efficiency) which are in many instances more significant in determining economic growth. In his assessment of the importance of allocative efficiency, Leibenstein draws on empirical evidence of the studies to show that the benefits gained from eliminating restrictions to trade in the U.S. by resource reallocation are insignificant. Due to small social or economic gains resulting from allocative-efficient methods such as cost-reduction, the author concludes that other methods of achieving economic efficiency such as plant layout reorganization, waste control and work methods should be implemented to increase benefits. The data on which this conclusion is based is gathered from the ILO reports developed by Kilby which show the magnitude for misallocation effects from other types of efficiency to be larger than effects from allocative efficiency. The empirical findings presented in this article, therefore, are used to confirm Leibenstein's argument that the improvement of X-efficiency would likely increase output. In comparison, the amount gained by increasing allocative efficiency is trivial. 4 4 Litchfield, Nathaniel: "Cost-Benefit Analysis In Plan Evaluation." The Town Planning Review, XXXV (July, 1964), pp. 165-174. This article concerns the use of cost-benefit analysis in evaluating the contents of a plan in terms of advantages disad- vantages, and comparison against alternative plans. Cost-benefit analysis of planning evaluation is usually based on the law of scarcity. Methods are devised comparing resources with the benefits they produce in a proposal plan. The author, therefore, directs his attention towards an explanation discounting the above premise as it pertains to planning in the area of social services, such as health care. Benefits of social service programs are most often valued in terms of cost minimization. However, representatives of the public when spending public funds must grapple with the diffi- culties involving social choice by spending money on projects they consider in the public interest. In improving choice between various social services, cost- benefit analysis may be useful in deciding which proposal makes best use of scarce resources and thereby contributes to the overall social welfare of the community. However, when benefits cannot be measured in terms of economic costs, the method must include broader criteria for adding the choice between alterna- tives. Such criteria: (i) analysis must deal with multi-use projects; (ii) analysis must deal with systems, as well as, independent projects; (iii) benefits must also be measured in terms of social costs and benefits instead of in purely economic terms. The main advantage of cost-benefit analysis of cost-benefit analysis is its flexibility. The method can be applied: (i) at any stage in the planning process, from the design stage to the formal consideration stage; (ii) to any kind of plan, regional to local; and (iil) to any particular proposal. Cost-benefit analysis is not the final and only means of determinating choice between alternatives concludes the author, but it is useful in aiding decision-makers in their choice because it provides a logical framework for analysis. 45 Meadows, D.H., Meadows, D.L., Randers, J., et al: The Limits To Growth: A Report For The Club of Rome's Project On the Predicament of Mankind, New York: New American Library, Inc., 19723. This book reports the construction and testing of a com- puter simulation model of world population growth and resource utilization for foreseeable future time periods. The objective of the model is to forecast the inter-active effects that five major variables will have upon the world under a number of assumed conditions. The five variables are population growth, food production, capital investment, non-renewable resource availabilities, and environmental pollution levels. The various chapters of the book discuss how one develops models for specific purposes which are suitable for computer simula- tions. ; Once the major variables of the "world" model are speci- fied, the important causal relationships among the five are listed and the feedback relationships among them are traced. The relationships are quantified as accurately as possible using global data where available and characteristic local data where global measurements have not been made. With the com- puter, the simultaneous, inter-active operation of all these re- lationships is calculated over time; that is, the future is fore- casted as a time series of data readouts. The effect of changes in the basic assumptions of the model are tested to find the most critical determinants of the system's behavior. Finally, the effect on the global system of various policies that are current- ly being proposed to enhance or change the behavior of the system are tested against the objectives of those who advocate such policies. The authors note that when modelling complex social systems for forecasting purposes, these steps are not always followed serially; often new information coming at a later step leads back to alter the basic feedback loop structure. There should not be one inflexible model. Instead there should be an evolv- ing model that is continuously criticized and updated as under- standing of the phenomenon being modelled increases. Advantages of the modelling process for planning are dis- cussed as are some of the disadvantages and conceptual pitfalls to be avoided. The latter include the hazards of using frag- mentary, partially representative, or sequentially non- discriminating data, and the difficulties in estimating function- al relationships under conditions of unspecified stochasticity. 46 Navarro, V., Parker, R., and White, K.L.:- A Stochastic and Deterministic Model of Medical Care Utilization. Health Services Research S:342-57, Winter 1970. A previously reported stochastic model (see Navarro, V.: Systems Approach to Health Planning. Health Services Research 4:96, Summer 1969) predicting the future distribution of a stable population in various states of health care is modified in the present article and expanded to take into account the stochastic process of varying rates of utilization among differ- ent age groups and the stochastic and deterministic processes that determine changes in size and age structure of the popula- tion. The functioning of the model is described and its use illustrated by application to a hypothetical population. Demand is defined implicitly as equivalent to service utilization. "Need" is not directly considered. The authors present a basic Markovian model which relates the various health service components to the probability of a patient going from any one of these components to any other of them during a fixed period of time (in this case, one day). This model is augmented by two modifications: (1) a modification to the Markovian matrix which allows taking into account the effect of the population's aging (and hence shifting patterns of service utilization), and; (2) a modification which permits taking the birth/death process and its effects upon utilization into account. The Markovian matrix theory is illustrated employing a numerical example involving hypothetical data. The article notes several limitations of the model. First, immigration and emigration are not considered. Second, .... "the degree of simplification in the assumptions underlying the arithmetic language and logic used....limit the usefulness of the model". This weakness can be lessened in the author's view by enlarging the number and type of health service states uti- lized in the model. Third, the model requires data on the population's distribution among the various health states by age groups, the population's health resources utilization rates, and its growth parameters. These data, however, are often unavailable. If the above data can be collected ".... if parameters are available, one can predict further the amounts of resources required by each age group for each component of the health ser- vices system". 47 Nomile, F. R., and Ziel, H. A. Jr.: Too Many OB Beds? Hospitals. 44 (14): 61-4, July 16, 1970. This article presents a methodology for determining the number of obstetrical beds needed by a hospital and describes its application in a metropolitan area of Michigan. The methodology consists of several steps. First, the number Of births is predicted by "...a projection of the historical trend in the number of community births, modified by pertinent demographic data and by local conditions". Second, "the average daily census of postpartum mothers (ADC-PP) can be estimated with acceptable accuracy by use of the following formula: ADC-PP equals the number of annual deliveries multiplied by the average length of postpartum stay divided by 365". Third, the variation in ADC-PP is estimated by using a normal curve and the ADC-PP's standard deviation. The number of OB beds required to meet this variation in ADC-PP a given percent of the time can be then be estimated by adding "...multiples...of the standard deviation..." This methodology is then used to determine the number of OB beds required in a metropolitan area. Based on this calculation the number of existing surplus beds is derived and the cost- savings achieved by consolidating OB services can be shown. 48 Phelps, C. E. and Newhouse, J. P.: Coinsurance, the Price of Time, and the Demand for Medical Services. The Review of Economics and Statistics, 56 (3): 334-42, August 1974. The purpose of this article is to measure the changes in elasticity of the demand for health services due to changes in coinsurance and due to upward shifts of the price of medical services including the price of the time required to obtain such services. Demand is measured by the utilization rates of various types of health services including physician, hospital, and dental services. Five implications are drawn and tested against recent studies by other economists and existing data collected by the authors. I The general conclusions of the article are that coinsurance does affect demand and "... the impact of coinsurance varies across medical services in a systematic fashion depending upon the time-price of the service". The specific conclusions, based on the five implications are: "Services with a relatively high time price, especially physician office visits, exhibit relatively low coverage (or price) elasticities and relatively high time price elasticities... services with relatively high money price such as home visits, show considerably higher own-price elasticities... Money price elasticities appear to fall with coinsurance rates", The results of this analysis have led the authors to believe that coinsurance should be considered in making decisions about consumption of medical care, because "consistently, across a number of studies based upon diverse data, coinsurance has been found to exert an impact on utilization of various services". 49 Phillip, Joseph P.: Some Considerations Involved in Determining The Optimum Size Of Specialized Hospital Facilities. In- quiry,(VI(4), December 1969, pp. 44-48. A case study was done by the Greater Detroit Area Hospital Council, using the Martin Place Hospital emergency unit, of the pattern of demand for emergency services to determine the opti- mum size of the emergency facility. The method defines opti- mality in terms of two efficiency criteria: demand and cost. Optimality is achieved when an "acceptable balance" between these two factors is struck. The method is divided into two stages: choosing the theo- retical distribution used to determine the demand for a speci- fied facility; and calculating the comparative efficiency of vari- ous sizes of EMS facilities employing the two efficiency cri- teria. In this case study, two theoretical distributions, the Poisson and Normal distributions, are compared against the observed distribution of demand to determine which best "des- cribes" the demand for the emergency facility by a specified population. A chart shows that the Normal distribution provides a better fit for the model in determining demand. The object of this stage is to determine which theoretical distribution best fits the situation, and therefore, requires a knowledge of statistical methods. The method relies to a certain extent on trial and error in testing models with different theoretical distributions. This could prove to be a lengthy time consuming process, depending on the staffs' knowledge of statistical data and its interpretation. The second stage of the method is conceptually more compli- cated but statistically more adaptable to the layman's use. However, the decisions to be made from the outcome of this stage require competent and extensive knowledge of the area in which one is testing (in this model, hospital facilities). Criteria of efficiency are formulated which reflect both the need of the community for EMS services and the requirements of the hospital as an economically viable organization. The two criteria are: e the ratio of the number of emergency facilities utilized to the number demanded. En the ratio of the number of emergency facilities utilized to the number allocated. An increase in one ratio tends to reduce the other. For example: the increase in the use of the facility to meet the need factor reduces the ability of the hospital to provide ade- quate care at an economical cost. 50 Through this method a range of alternatives is provided so hospital authorities can select the size which best represents a compromise between the community's philosophy of adequate health care (expressed purely in terms of demand), and the hospital's ability and policy in meeting the costs of fulfilling these demands. This method provides a possible means for determining pri- orities, but does not deal with how to rank these priorities once determined. The model purposely leaves open the question of emphasis: where to strike a balance between meeting demands and controlling costs. The standard of "appropriate balance" is left undefined by the author which makes it difficult to define and set priorities. Perhaps a system for priority setting can be derived by testing priorities in terms of their optimality using the appropriate statistical data. However, in order to rank priorities, an agreed standard of appropriate balance" between demand & cost must be established. 51 Pollack, J. and Taf, M.: "Urban Planning; Ripe for Systems Analy- sis." Journal of Systems Management. January 1971, pp. 12-17. The authors present an overall systems design and general- ized planning model which serve as possible tools in the analysis of urban planning. The basic objective of the model is to solve planning prob- lems through system analysis by breaking down the components of the area of planning into easily managable subsystems. This technique provides a useful framework in which to view health care planning problems as well as urban planning. The breakdown of health care planning into various subsystems poses an interest- ing if not viable possibility for rational and effective decision- making. Through analysis of the various subsystems which compose the overall system, (for example urban subsystems - shopping centers, industrial parks, residential neighborhoods) a more con- cise picture of the total needs and objectives can be assessed. The ability to measure problems of specific subsystems (employ- ment opportunities, walking or riding distance, accessibility of recreational areas) enables the analyst to construct standards for achieving certain objectives. The need for priority setting is clearly implied within the general framework of analysis, but not specifically stated. The framework leaves the formulation of specific criteria on which to base value assessments up to the "imagination" and ability of the person using the model. Only general criteria are given on which to base a priority-setting method--long range goals and short range objectives. Therefore, an exact method of priority, setting would have to be extracted from the general planning model. This would require a large degree of expertise and fam- iliarity not only with the methods and theories used as inputs in the planning model, but also with the interrelationships between various subsystems in the particular field of planning. Quade, E.S.: Cost Effectiveness: Some Trends In Analysis. The Rand Corporation, P-3529-1, March, 1970 Quade discusses new approaches and techniques of analysis which aid decision-making. Most new approaches, he asserts, are computa- tional in nature. However, in this article stress is placed on the few that are concerned with those aspects of problems which cannot be handled in a purely quantitative manner. Computers, mathematics, the use of expertise, and procedures to improve acceptance and imple-, mentation are the four trends noted by the author. To solve military defense problems, systems analysis is pre- dominately used, however, decision-makers involved with nonmilitary problems are more inclined to recognize the merits of an approach based on analysis but supported by intution and experience. The computer is advancing its role as a decision aid due to advanced technology resulting in a closer relationship between analyst and machine. The author stresses that even the most sophisticated computer employing the most advanced mathematical techniques cannot aid in all decision-making. Under the category of mathematical techniques, Quade notes the trend towards the use of game theory as a valuable quantitative method for handling cooperative and competitive elements of problems. In a model where moral considerations dominate and the use of quantitative methods would be inadequate, the use of techniques involving experts is necessary. Finally, according to Quade, the analyst must devote some attention to problems of acceptability and implementation. 53 Reinke, W.A.: Analysis of Multiple Sources of Variation: Comparison of Three Techniques. Health Services Research . 8:311l>-21, Winter 1973 Three tecnhiques of statistical analysis -- multiple regression, automatic interaction detector (AID), and multisort -- are applied to the analysis of data from an investigation of health services utilization and expenditure patterns in a national sample of 45,000 residents of Chile. The inherent weaknesses and strengths of each method are described and illus- trated with special application to the socio-demographic factors associated with physician utilization. The overall results given the three methods are compared. Because of the limitations imposed by the circumstances under which the Health services utilization survey was conducted, the analyses are confined to the following 13 utilization factors and six independent variables: Utilization (dependent) variables Physician visits in 2-week period: Percentage of population with visits Visits per capita Cost per capita Dentist visits in 2-week period: Percentage of population with visits Visits per capita Cost per capita Hospitalization in 6-month period: Percentage of population hospitalized Hospital days per capita Cost per capita Drug expenditure in 2-week period: Percentage of population with expenditure Expenditure per capita Laboratory service expenditure in 2-week period: Percentage of population with expenditure Expenditure per capita Sociodemographic (independent) variables Residence location (Santiago, other urban, rural) Monthly per capita income in escudos (0-74, 75-149, 150 and over) Sex Age (0-4, 5-14, 15-49, 50 and over) Insurance benefit status (not covered, national social insurance, other coverage) Years of education (0-3, 4-9, 10 and over) 54 The first technique, multiple regression analysis, is a well-established technique commonly used when large numbers of independent variables are being considered. It requires, however, detailed advance specification of the form of the mathematical relationships expected to be present. Because such relationships are often poorly understood, multiple regre551ons are too often based on arbitrary formulations that result in naive analysis. Specifically, the approach encounters difficulties in handling nonlinear effects (e.g., those induced by age), nonordinal variables (e.g., classifications of insurance benefit status), and interaction effects (e.g., those between income level and age). When each cell or block of the study design contains the same number of observations, these difficulties are handled routinely by analysis of variance. In such circumstances ortho- gonality is preserved with relatively simple calculations. With varying numbers of observations per cell, however, the calcula- tions and interpretations associated with analysis of variance become extremely complex. This is inevitably the case when there is correlation between independent variables. The second approach, involving the use of the automatic interaction detector (AID) program, has already been applied to the analysis of health survey data. The AID technique can conveniently accommodate large numbers of independent variables, but identification of interactions is not as automatic and unambiguous as the name of the technique suggests. The third approach, multisort analy51s, was developed by the author in analyzing health services utilization in Taiwan. This technique cannot handle as many independent variables as the AID technique, but the results are more definitive if the number of independent variables is small enough to permit obser- vations to be recorded for nearly all possible factor combina-, tions. In the sections that follow, the application of these techniques to the per capita physician visits variable is des- cribed in detail. The results obtained for the other dependent variables are then summarized in a section on comparative results of the three approaches. 55 Rosenthal, G. D.: The Demand for General Hospital Facilities. Hospital Monograph Series, No. 14. . Chicago, Illinois: American Hospital Association, 1964. This book develops a model for estimating the general hospital facilities needed, by State, to meet a given demand for actual days of care. Demand is not explicitly defined in the book but is equated in the model with the utilization of hospital facilities in a State area as determined by hospital characteristic variables and parameters of the demand relationship. The model used to estimate the demand for hospital facilities, described in chapter 3, is a least-squares linear multiple regres- sion. The purpose of this method is to demonstrate casual relation- ships and to obtain forecasts. The analytic framework allows socio- demographic and economic characteristics to be associated in some predictable way with differences in utilization. Then these asso- ciations can be used to calculate estimates of utilization and to forecast future use patterns. Chapter 4 of the book discusses each of the independent variables and their expected relationship to the use of general hospital facilities based on logical reasoning and special studies. These characteristics include age, marital status, sex, degree of urbanization, race, educational level, popu- lation per dwelling unit, price variations in hospital charges, in- come of consumers and population with insurance. "In addition to the total demand relationship (patient days per 1000 population, the dependent variable), the components of demand--admissions per 1000 and average length of stay, also dependent variables--were individually regressed against the same set of independent variables ...". Chapter 5 discusses the observed interrelationships. Chapter 6 derives an estimate of beds needed per State "...generated by the peculiar characteristics of each State with respect to both demand and the average size hospital. Part III of the book applies the estimates of chapter 6 to current methods of predicting bed-need. The appendixes contain a bibliography and curvilinear model of demand. 56 Rosenthal, G.: Price Elasticity of Demand for Short-Term General Hospital Services, Empirical Studies in Health Economics. Baltimore, Maryland: The Johns Hopkins Press, 1970. 'pp. L01l~17. The purpose of this analysis is "to examine the degree to which the length of stay is associated with various price-payment factors" for a fairly homogeneous group of patients. Demand for hospital facilities is described as consisting of two aspects: the admissions rate and the length of stay. The analysis presented, however, deals only with the latter. The data were obtained from a sample of medical records and financial information pertaining to 15,685 admissions to 68 New England non-Federal, short-term general and special hospitals. The sample was stratified by institutional size. The actual analysis of the data is based on two hypotheses. The first hypothesis is that "... length of stay is a function of some price-payment measure. Two specific price-payment measures are examined: cost outlay as a percentage of the total bill and average daily room charge"., The second hopothesis is that "... certain diagnostic categories will be more likely to show a high price elasticity of length of stay than others." The data are grouped by patients with similar age, sex, and diagnostic characteristics in order to better reflect the effect of the two price-payment factors. The actual analysis then consists of two regressions for each group. These regression analyses were of the following form: J 1. Y = aX b a + bl log X1) 2% Y = aX b a + b2 log X2) In this formulation, the dependent variable Y represents length of stay, the independent variable Xl represents cash payment as a percentage of total bill, and x2 represents average room charge. The "bs" represent elasticities of length of stay with respect to each price variable. 57 The analysis provides "...evidence of significant price elasticity of length of stay when the price variable is the average daily room charge," and "... far less price elasticity... with respect to relative cost outlay..." Additionally, "the results suggest that for a shorter length of stay there is likely to be less price elasticity". 58 Schneider, J. B.: Measuring the Locational Efficiency of the Urban Hospital, Health Services Research. 2(2): 154-69, Summer 1967. This paper presents a methodology which measures the locational efficiency of an individual hospital by determining the minimum aggregate travel point of certain hospital users and, by means of a locational imbalance factor, measures any distance between that point and the actual hospital location. The term locational efficiency is defined as "...the costs of operating a hospital which may be attributed directly to its location. A major portion of such costs includes the out-of- pocket and time costs incurred by various groups of hospltal users for travel." The decline of locational efficiency is defined as equivalent to a rise in monetary and time costs of travel by various groups of hospital users. The demand for hospital services is viewed in terms of utilization patterns which may be adversely affected by poor locational efficiency. The paper argues that locational efficiency can be approximately measured by the pulling force of just one of several possible groups, hospital inpatients. This pulling force is the product of ".I. (1) the value of its travel time and (2) the frequency of its trips to the hospital..." This methodology is used to analyze nineteen Cincinnati hospitals using inpatient data collected between April 1, 1964 and March 31, 1965. Some policy questions which arise from the study's measure of locational efficiency are: A Is it desirable to have several hospitals compete in in the same service area? R Is a particular area "over-served" in relation to other parts of the urban environment? Can a change in patient mix among centrally located hospitals improve locational efficiency? 59 Schonick, W. and Jackson, J.R.: An Improved Stochastic Model for Occupancy-Related Random Variables in General-Acute Hospitals. Operations Research. 21(4): 952-65, July- August 1973. The objective of this paper is to present improvements to an earlier stochastic model. (See Schonick: Understanding the Nature of Random Fluctuations of the Hospital Daily Census: An Important Planning Tool. . Medical Care 10 (2): 118-142, March/April 1972.) for appraising the distribution of the daily hospital census, or more precisely, the daily census of a distinctive patient facility (DPF). A DPF is defined as roughly equivalent to a service such as med/surg or obstetrics. The earlier model was developed by the author to answer the following question: "How many beds does a community 'need' to meet a previously determined demand for hospital service?" The improved stochastic model reported in this article is designed to be used most appropriately in areawide health planning, to correct the particular shortcomings of the earlier model which "...was likely to impair the usefulness of the model for a DPF that was able to queue a goodly proportion of its patients," and to provide the health planner with greater flexibility "...by increasing the number of decision choices available for achieving given outcomes. This increased flexibility is attained via a generalization of the queueing disciplines, permitting admissions of elective patients to be suspended when the number of occupied beds reaches a predetermined level (which may be less than the total bed complement). For any emergency/elective mix demand for hospi- talization, say the authors, this model permits the computation of many measures of operating efficiency including expected overfill rate, percentage occupancy, waiting-list length, and loss of emer- gency patients. The article also discusses the assumptions of the model, the derivation and solution of the steady-state equations and provides examples of, and a comparison of, the model with its precursor. The article, furthermore, discusses the use of the improved model where the bed complement is effectively infinite (i.e. where emergency cases arrive during "full" periods and are treated in non- approved facilities rather than turned away). 60 Wirick, G.: A Multiple Equation Model of Demand for Health Care. Health Services Research. 1(3):301-46, Winter 1966. The purpose of the methodology presented in this article is to describe alternatives for decision-making by identifying demands for various components of the health services field. Using a simultaneous equation model, five components of demand for medical care (hospital, doctor, dentist, medicine, other) are represented by five equations into which are also included measures representing "the forces thought to influence consump- tion (need, realization, motivation, resources, and availability of service)." Demand is measured in terms of the physical amount of service desired. The "need" for services -- one of the five forces affecting consumption -- is when "a person suffers from a condition requiring medical attention, or [when] he has some other reason for seeking the supplies or services classified as medical care." Need also depends upon a recognition or realiza- tion of need in order to be operationalized. Using statistical information about demographic character- istics of the population and actual utilization statistics, the methodology of the model depends upon a computerized analysis of interaction of the components and forces to generate esti- mates of demand. Because the goal was to develop a planning tool, it was determined that the measure of demand in terms of the physical amount of service desired was more meaningful than use of an expenditure variable. The statistical method applied to the data is an application of one-way analysis of variance, called the Automatic Interaction Detector (AID). It is basically a stepwise optimizing technique, which assures that only those partitions of the entire sample which substantially contribute to explanation of variance will be made. The results of model development and testing on data from a sample survey of Michigan's population in 1958 (1,000 families, comprising about 3,500 individuals) reveal that there was support for the model structure and its variables, and that it forms a good basis for development of a simultaneous equation model. 61 Yett, D. E., Drabek, L., Intriligater, M.D.: A Macroeconometric Model for Regional Health Planning. Business and Economic Bulletin. 24(1); 1-21 Fall 1971, In view of today's widespread and justifiable concern over the expense and performance of the nation's health care system, areawide planners urgently need operational tools for the eval- uatfon of alternative policies. The purpose of this paper is that of describing an econometric model which, it is hoped, will provide such a tool. The goal of this model is to assist CHP agencies by allow-, ing them to: (1) deal with both supply and demand in their analyses of the economics of their health care system, (2) treat the three major elements in this system -- health services, man- power and facilities, and (3) analyze a wide variety of alter- native policy instruments as these simultaneously act upon the many types of elements and participants in the system. The model is constructed in three phases. The first phase is that of conceptualization of the model, identifying the major economic features of the health care system and their demand, supply interrelationships. The second phase is that of estimat-, ing the variables of the model, using California Statewide data, determining the quantitative importance of the interrelation, ships and examining "outliers." The third phase is that of simulating the model, both replicating the recent period and providing conditional forecasts of the variables under alter- native policy initiatives. The relationships among the participant groups in the model is illustrated. Consumers, providers and manpower (circles) are linked through services and labor markets (diamonds). The lettered arrows represent groups of equations in the model., while the symbols in each market are the variables to be determined. Demand relationships (A,C) point to the right while those pertaining to supply point to the left (B,D). As of the date of this article, the third phase was awaiting completion. 62 Vietorisz, Thomas.: "Quantized Preferences and Planning By Priorities." American Economic Review. 6(3) August 1975. pp. 65-69. The author relates project planning by priorities to planner's preference in an attempt to discover some relationship between con- sumer utility functions and quantitative models. The inability to measure utility causes priorities to be derived independently from resource allocation, or payoffs in terms of resources. Indifference curves graph the optimization of satisfaction with a given priority subject to a budget constraint. Preference maps derived from priorities order utility functions ordinally. Indifference curves order project sets whose utility increases as additional projects are included. Resources are valued for what they contribute to different entities of project sets. This model assumes that projects are implemented in rank order regardless of resource allocation but in order of utility maximization. The object of the method is to maximize the number of projects implemented in terms of utility over a time period. 63 CHAPTER II--GOAL HARMUONIZING APPROACHES Primary Emphasis on Concepts Bender, Douglas A.: "Delphic Study Examines Developments in Medicine." Futures, pp. 289-303. Smith, Kline and Franch Laboratories conducted a study on the Delphi forecasting technique as an aid in planning a pharmaceutical company's operations. In this study, the Delphi technique produced a consensus of medical experts' opinions about: future developments in five areas of medicine: biomedical research, diagnosis, medical therapy, health care, and medical education. Before Delpni was tested on an outside panel_of experts, Smith, Kline and French Laboratories conducted an in-house test on a panel chosen from the company's scientific research and development staff. From the questionnaires answered by these staff members, a table of consensus projections was formulated out- lining areas where research was considered urgently needed and feasible within the next 50 years. Thirty-three panel members from various medical background participated in the complete Delphi study. Two sets of question-, naires were used. Questionnaire I gave a brief working definition of each of the five areas of medicine and asked participants to list important discoveries, breakthroughs, changes in methodology and other events that might occur in each of the next 50 years. Questionnaire II grouped the statements in the five areas of the medical field being analyzed, each broken down into several subdivisions. Panel members were asked to estimate the year in which the event had a 50 percent, 90 percent or 0 percent chance of occuring. Figures 1-5 in the article list all statements on which consensus was reached among panelists and show graphically the median and interquartile range in each major area of medicine for the 0 percent, 50 percent and 90 percent estimated probability levels. Data obtained from this study are also used to construct a scenario of the education, training, practice and research desir- able for an American medical student entering school in 1978 and entering practice in the early 1980's. The author concludes (somewhat fatuously) that Delphi is one methodology useful in estimating projected changes in the medical field, whether these changes will be evolutionary or revolutionary. 64 ponabedian, , A. v.: "Assessment of Need", Chapter 3 of Aspects of Medical Care Administration: Specifying Requirements for Health Care. Cambridge, Massachusetts: Harvard University Press, 1973. pp. 80-251. This chapter of the book provides an extensive survey of topics relating to needs, with emphasis upon clinically deter- mined needs. The introduction to the chapter presents an illustration of the conceptual framework of the medical care process. "Need" is grossly defined as "same disturbance in health and well-being." A more camplex view of need, the author notes, recognizes that there are at least two perspectives on need: that of the client and that of the provider. The differences between these viewpoints are indicated, and the importance of some degree of congruence between the two viewpoints is illustrated. Because "need" is often used to refer to other than states of health or illness, the author points out that "need" refers to three concepts: "(1) states of health or ill nealth; (2) use of services; and (3) levels of supply". The term "need equiva-, lents" is proposed to apply to the services or supplies (re- sources) required for the patient who requires care. A model is proposed for the assessment of need which represents a wedding of the medical care process and the notion of need equivalents. Represented as a three colum-three row matrix, the model includes most of the issues and complexities of needs deter- mination, including utilization ("effective demand"), unment needs, resources, and the complexities of the two viewpoints on need. These reasons for assessment of need are given: First, need defined as states of health and illness may be considered to constitute the outcome or end product of medical care. Hence, it can be used to indicate, in part, the success or failure of the medical care system in any given population ... The first use to which estimates of need may be put is, therefore, evaluational. A second use is in the formulation of priorities for health action ... A third use is for the purpose of translation into equivalent units of services and resources necessary to satisfy the varieties and levels of need prevalent in a given population." The chapter and the book contain extensive bibliographies relevant to need assessment concepts of health and illness, measurement techniques, etc. 1 65 Jeffers, J. R., Borgnanno, M. F., and Bartlett, J. C. +. On the Demand Versus Need for Medical Services and the Concept of 'Shortage'. American Journal of Public Health. 61(1) : 46-63, January 1971. Mucllor, H. F., Uphoff, W. H., and Zocliner, H.: Medical Services Demand Versus Need - A Comment. American Journal of Public Health. 64(1): 54-57, January 1974. Jeffers, J. R., Borgnanno, M. F., and Bartlett, J. C.: Demand Versus Need and the Concept of Shortage - Rejoinder. American Journal of Public Health. 64 (1): 58-60, January 1974. The purposes of the first article, which is theoretical in nature, are to differentiate between the concepts of "need" and "demand" and to interpret the concept of "shortage" as it applies to medical services. The second and third articles are coments on the first and main article. The perspective in all three articles is that of a theoretical economist. Need is defined as the quanitity of medical services that a given population ought to consume over a given time period based on the standard of normative professional judgement. This judge-, ment is assumed to be precise, authoritative and not subject to interclinician variability. Differences between this concept and a consumer's concept of need ("wants") are discussed also. Demand is defined as a "multivariate functional relationship between the quantities of medical services that the member of a group desire to consume over a relevant time period at given levels of prices of goods and services, financial resources, size and psychological wants of the opoulation as reflected by consumer tastes and preferences for (all) goods and services." As defined, "need" is independant of prices or financial resources of consumers, but "demand" is directly influenced by these and other factors, although not necessarily in a rational pattern. Taken together, the three papers consititute a debate over the validity and consequences of these concepts for economic. theory. Minimal reference is made to empirical validation or policy consequences of the major ideas. "Shortage" of medical services is discussed as having two quantities to consider - normative shortage and market shortage. 66 Shortage can be generally defined as that quantity which is the difference between medical services consumed ("demand"), and medical services needed. Market shortage is used as a shortage which may be expected to work itself out through upward price adjustment. Normative shortage, however, requires market inter- vention which alters the basic character of the industry. The alteration may occur on the sides of supply and demand. From these basic concepts, a discussion of the effect of costs for medical services, as influenced by market alteration on the sides of both supply and demand is pursued. The bibliography at the conclusion of the first article is extensive. 67 ' Kahn, H. and Wiener, A. J.: The Year 2000: A Framework for Speculation, New York: MacMillan Company, 1967. The study of the future, say the authors, once the exclusive domain of science fiction writers and prophets, has become a vital part of the conduct of public policy. Drawing extensively upon the use of scenarios, the book develops methods for drawing a comprehensive picture of the shape of probable, possible, "ideal" and "nightmare" worlds. It illustrates the usefulness of these methods to those charged with planning for the future. Economics, demography, history, political science, sociology and the physical sciences have, all provided statistics, projections and "surprise free" information about the changes in technology, science, population and international power balances which are likely over the next several decades. The book is concerned with demonstrating methods for effectively coping with such information in an effort to forecase more effectively. While not specifically devoted to health planning or the forecasting of needs or demands, the book is of interest on two grounds: it illustrates the use of scenarios and canonical pro- jections and it includes a stimulating chapter on forecasting methodology. Major topics in this chapter are shown below: CHAPTER X. POLICY RESEARCH AND SOCIAL CHANGE A. Introduction B. Ways to Go Wrong 1. Criteria Too Narrow 2, Decisions at Inappropriate Point in Structure 3. Inadequate Thought 4. Bad Luck: Unknown Issues 5. Bad Luck: Unlikely Events 6. Changes in Actors 7. Inappropriate Models 68 neenee ee on c. 10. Inappropriate Values Over - or Underdiscounting of Uncertainty or the Future The Best May Be the Enemy of the Good (and Some- times Vice-Versa). The Objectives of Future-Oriented Policy Research l. 2. Stimulate and Stretch the Imagination and Improve the Perspective Clarify, Define, Name, Expound, and Argue Major Issues Formulate and Study Alternative Policy "Packages" and Contexts Create Propaedeutic and Heuristic Methodologies and Paradigms Improve Intellectual Communication and Cooperation Increase the Ability to Identify New Patterns and Crises and Understand Their Character and Significance Furnish Specific Knowledge and Generate and Document Conclusions, Recommendations, and Suggestions Clarify Current Choices -- (Hedging, Contingency Planning, and Compromising) Broaden and Improve the Basis for Both Political Decision-Making and Administrative Actions in Dealing with New Trends and crises. Conclusion: Man's Increasing Faustian Power over Nature (Including Man). 69 Kaufman, H.: - "The Politics of Health Planning." AJPH, 59 (5), May, 1969, pp. 795-813. In the introduction of this symposium on the political aspects of health planning, the author observes that "Politics in the broadest sense involves efforts to mobilize and utilize power to attain ends. From this point of view public health has always been involved in politics." Public policy decisions are by definition political decisions. Politically sophisticated groups outside the health field are voicing their concerns about public health services, therefore planners who alienate them- selves from politics, create a role of political impotence by depriving themselves of the means of shaping public policy. Dr. Kaufman's introductory remarks are central to the themes of the three subsequent articles. Dr. Mott, in The Myth of Planning Without Politics, maintains that most models of community health planning, such as the rational decision and the community action models, do not adequately deal with political realities in the health planning field. The rational decision model is politically naive because it views planning as an exclusively technical process. The community action model is not politically feasible due to its failure to recognize the limits of consensus in formulating a health plan enmeshed in conflicting interpretations of health data. In the second paper, Dr. Feingold discusses the background of health planning and its outlook for the future. Health planning has been plagued by lack of agreement on goals, lack of priority ordering of goals, and lack of the means to implement them. This disagreement is attributed to the decision-maker's lack of understanding of the political arena in which health planning takes place. The remaining paper by Robert Binstock discusses different types of political influence and how these are (or should be) manipulated by the planner to achieve desired goals. 70 Mason, R.O.: A Dialectical Approach to Strategic Planning, Management Science. 15 (8), 1969, pp. B-403-414. This study begins with the author's assertion that organizational structure and planning are based on management's assumptions about the world in which they operate. These assumptions include predictions, value systems, and choices among behavior patterns. The author is concerned with the development of a planning technique that "tests" these assumptions by exposing hidden assumptions, and suggesting new relevant assumptions upon which the manager can base his future planning strategy. Mason points out the failure of the expert approach and the Devil's Advocate approach for testing assumptions. Both destroy the organizational plan without replacing it with an improved one. The design for the dialectical approach to planning stems from Hegel's triad--thesis, antithesis, and synthesis The method uses identical health data to construct a plan and "counter plan". The manager as an observer of the two conflicting plans integrates them together by exposing hidden assumptions and developing a new conceptualization of planning. Evidence supporting the dialectical approach is obtained from a field study experiment by RMK Abrasives. Mason concludes his article by listing conditions necessary for the implemen- tation of the dialectical approach. 71 Newhouse, J. P., Phelps, C. BE., and Schwartz, W. B.: Policy Options and the Impact of National Health Insurance. New England Journal of Medicine. 290:1345-1359, 13 June 1974. This article addresses three questions in the face of early enactment of a Federal health insurance program: " (1) how much change in demand for health services will be created by any given insurance program and how much will a change affect the nation's health care bill?; (2) how readily can each component of the health-delivery system respond to the anticipated increase in demand, and what will the consequences be if the demand for ser- vices cannot immediately be satisfied?; (3) how much return on investment can be expected from increased provision of services?" The methods employed by the authors are drawn from public expenditures analysis, price theory, and capital investment analysis. None are directly relevant to appraising the demand for specific health services. The data indicate that under full coverage or 25 per cent maximum insurance, demand for hospital services would rise modestly. However, either program would greatly increase demand for ambulatory services and would stress the delivery system, with resulting increased price of physicians' services, queuing, or less physician time per patient -- all without increasing total delivery of ambulatory services. Ambulatory services woud be redistributed from the affluent to the poor. A catastrophic health insurance program would not stress the ambulatory system. Reorganization of the delivery of ambulatory services into pre- paid groups will probably not increase productivity, nor will emphasis on preventive medicine reduce overall demand for health services. National insurance which provided more health services benefits would not appreciably affect objective indexes of health (life expectancy), but should improve subjective but relatively unquantifiable elements such as quality of life. 72 Schlesinger, J.R.: Systems Analysis And The Political Process. The Rand Corporation, P-5284, September 1974 . Schlesinger evaluates the role of systems analysis and its function in a political bureaucratic environment. He assesses the quality of information, methodology, bias (especially within a large organizational gtructure) and the impact of politicized environments, such as Congress and the Defense Department. Particularly stressed is the bureaucratic impact on systems analysis. Often bureaucracies are pressured to produce new pro- grams before analysis has been performed to determine the potential problems and value of such a program. Many agencies do not release all the information available so that analysis and, therefore, decisions are based on deficient knowledge of a problem. The author lists the causes of bias in a bureaucracy: a symmetry in sources of information, disproportionate attention by the analyst to preferred information sources, prior intellectual commitment on the part of the analyst, selectivity in organizational recruitment, and other bureaucratic pressures. Schlesinger believes that effective policy research must be carried out at the highest managerial level if analysis is to be used to its full advantage. However, this is dependent on the modernization of the bureaucratic structure. In conclusion, the author states that systems analysis cannot in the long run be expected to solve all problems confronted by decision-makers in the political arena, but it can in the short run serve an educative function by reshaping how agencies view their own problems. 73 Simon, Herbert.: "On the Concept of an Organizational Goal." Administrative Science Quarterly. June, 1964, pp. 1-22. In this article, the author discusses the concept of "organ- izational goal" and its effect on individual behavior within an organizational structure. The author defines an organizational goal as the set of‘ constraints that define the roles played by decision-makers in different levels of the organization. In this analysis, individual behavior is usually explained in two ways: (i) in terms of the individual goals of members of an organization, or (ii) in terms of the existence of organizational goals higher in priority than goals of the various individuals. To clairfy the collective decision-making process, the author makes a distinction between "goals" and "motives." He suggests that a course of action must satisfy a whole set of requirements in the decision-making process, whether or not the action is of an individual or organizational nature. In the determination of the course of action to be pursued to satisfy this set of require-, ments, the article examines motivations, role behavior, inter- personal differences and other artifacts of the organizational decision-making system. The author concludes that the selection of requirements or constraints is based on their relation to the motivation, and to the search process generating particular courses of action. 74 Woods, D. H.: Improving Estimates that Involve Uncertainty. Harvard Business Review 44(4); 91-99, July-August 1966. Despite the fact that all forecasts-gontain elements of un- certainty, they are often handled within (and between) organiza-, tions in ways which do not take this uncertainty into sufficient account. In particular, the author. notes -that forecasts: are frequently misused: (1) Misinterpretations and ambiguities occur simply because judgments about uncertainty are difficult to transmit easily from one person to another. (2) Biases and distortions occur because (a) the "experts" are not always sure what forecasts are desired by those who must make decisions employing them; (b). the capacity for taking: (or appraising) risks differs among individuals. (3) The estimating procedures used by the experts them- selves are often unsystematic, muddled and communicated in a way which suppresses or misrepresents the un- certainty inherent in them. The purpose of the article is to present a number of devices for improving intra-organizational communication about forecasts involving uncertainty. The author recommends four formats for presenting forecasts so as to minimize communication problems: (1) A table showing the range of possibilities and their associated probabilities of occurence; (2) A histogram showing the numberical values associated with each possibility and their probability of occurence; (3) An "Uncertainty Index"; (4) A scale showing the character of the forecast (i.e. the extent to which it is viewed by its developers as conservative vs. optimistic). Though the concepts and examples in the article are selected from a business context, the methods should be of general applicability in organizational situations where decisions must be made on the basis of uncertainty. 75 Primary Emphasis on Methods Delbecq, A. and Van De Van, H.: "A Group Process Model for Problem Identification and Program Planning." The Journal of Applied Behavioral Science. pp. 466-491. The authors develop a group process model for identifying strategic problems and developing appropriate programs to solve them. The process is geared towards innovative solution stra- tegies by bringing together various constituency groups with different ideological concepts and vested interests. These con- stituencies are involved in successive phases of the process as follows: I. Problem exploration-clients; II. Knowledge exploration-external resource people and internal specialists; III. Development and adaptation-administrators; IV. Program proposal building-organizational staff; V. Approval and evaluation-all constituencies. The use of the nominal group approach to stimulate and increase creativity is applied to the five successive phases of plan development. Phase five reinvolves consumer representatives, knowledge resource people, and administrators in the review of the plan proposal around concerns developed in prior phases. This enhances the opportunity for majority approved last-minute adjustments and minority reservations. 76 Lamanna, Richard A.: "Value Consensus Among Urban Residents." JAIP, XXX, No. 4, 1964, pp. 317-322. This study in urban planning employs an attitude survey for investigating and obtaining information on concensus building useful in urban planning. The survey was conducted in Greensboro, North Carolina, in 1958. The author believes that the critical problem of consensus building lies in the relationship between the "consumer" population's desires, and the knowledge and values of planning experts. The effort to bridge the gap between the two occurs in the process of planning not in the plan itself. The main goal of the planning process is, therefore, to maintain mean- ingful communication and consensus among participants on goals. The Greensboro study analyzed the value orientation of urban residents by means of rank ordered correlation coefficients. The larger the coefficient of rank order correlation, the greater the consensus between the population segments on the relative import- ance of a given value. Other tables rank order the "livability values" in selected population segments. The results show a high degree of consensus among various population segments concerning the relative importance of the thirteen values studied. According to the author, the planner's task, therefore, is not hopeless because the evidence from the Greensboro survey indicates communities can be planned in terms of a common core of value assumptions. However, the author warns planners that sound planning must take into account significant differences in the values of urban residents. 77 Lawson, J.S.: How Many Beds? Problems in Estimating Requirements for Hospital and Nursing Home Beds. Medical Journal of Australia. 1(2):70-73, 8 January 1972. The objective of the article is to attempt to define ideal bed to population ratios in order to develop a basis for estab-, lishing a balance between adequate services and excess bed capacity. "Demand" for hospital and nursing home beds is loosely de- fined in this article. Somewhat synonymous with utilization of facilities, demand is suggested to be "a ratio of beds to pop- ulation that will provide sufficient accommodation for the ma- jority of patients most of the time". The methodology used for developing the need ratio is quite judgmental. The author analyzes British, American, and Austra- lian literature for bed-population ratios, and then applies his experience and judgment to current utilization rates to deter- mine the proposed rates for each type of facility. The results of the author's efforts are a series of ratios, presented below. He suggests that, although arbitrary, the ratios are an essential planning tool. The suggested ratios are as follows: "(i) acute beds (excludes obstetrics, psychiatry, geriatrics; includes medicine, surgery, tuberculosis, gynecology, etc.)}, 3.4 beds per 1,000 of the total population; (1i) obstet- rics beds (excluding gynecology), 30 beds per 1,000 births per year; (iii) geriatric beds (a) nursing home beds for patients requiring active nursing care, 25 beds per 1,000 of the popula- tion aged 65 and over, and (b) welfare beds for the 'frail' aged, 15 beds per 1,000 of the population aged 65 and over; (iv) psy- chiatric beds, until more experience in the care of psychiatric patients in the community, 1-8 beds per 1,000 of the total population ". 78 Moore, W. S. and Bock, H. B.: Estimating the Demand for Medical Care. Inquiry 9 (4) :64-66, December 1972. This research effort has as its objective the design and development of a Health Services Simulator which predicts demand for medical care, and measures resulting requirements for care. Demand is defined in terms of level or extent of care, duration and disposition of cases. The model developed in this project "processes by age groups the transitions of patients between levels of care, recovery, and death by classification of conditions based upon morbidity and time ",. Because the transition data are not available 'from published statistics, the technique of subjective likelihood has been used. This method involves multiple esti- mates, collation with statistical data, and resolution. The statistical data which can be used in the model are demographic and utilization data. The demand of a system (age group) is estimated using multivariate regression equations which reflect demographic, economic and social factors, and existing health resources of the population. Using this technique, the demand for medical care for that element of a given population over 65 years of age is developed. 79 Navarro, V." A Systems Approach to Health Planning. Health Services Research 4 (2) :96-111, Summer 1969. The planning model described takes a holistic, "system-, structure approach". It plans: for the different parts, or subsystems, of the total health system, taking into consider- ation the interdependence among the parts. The model requires data on internal functional relations among components, as well as on the number of services provided. Basically a demand model, its purpose is threefold: "prediction for the ordinary statistical problem of forecasting; parametric study or simu- lation, for determining the effect on the whole system of simulated changes in its parameters; and goal seeking, for calculating the optional utilization strategy to achieve a specified goal under given restraints such as minimization of costs or resources". The methodology of the model embodies a Markov chain in which health services states of population not under care, under consultant care, under primary medical care, nursing home care, hsopital care, or domiciliary care are postulated, and in which the probabilities of going from one state to another determine the number of patients in the various States through-, out time. "Transitional probabilities" -- the probability that a person who is in one State at the beginning of the defined period will go to another State during that period -- are the heart of the model. These probabilities are considered as known in the model; no methodology for their calculation is provided. The model is applied in the three areas on prediction, simulation, and goal seeking, and its utility is shown to the planner in these examples. 80 Pill, J.: The Delphi Method: Substance, Context, a Critique and An Annotated Bibliography. Socio-Economic Planning sciences. 5(1):57-71l, 1971. The objectives of this paper are to provide a comprehen- sive review of the Delphi technique, a procedure which offers certain advantages in the systematic use of expert opinion. Since this method may be considered a specialized part of the whole field of subjective scaling (i.e. construction of mea- suring devices for mental processes, particularly preferences and normative concepts), it is treated in that context. The paper includes a brief review of the historical development of the Delphi methodology, a description of the procedure itself, and a fairly complete annotated bibliography. Delphi's ad- vantages and disadvantages are discussed. A comparative analy- sis of its role vis-a-vis other techniques such as "revealed preference" as used in the econometric theory of consumer be- havior and in various psychological scaling procedures is pre- sented. An analytical framework is proposed for the larger question of scaling subjective variables and an attempt is made to delineate the usefulness of Delphi in this context. The author reaches the following conclusions about Delphi given his purposes: 2% The Delphi technique should be used on forecasting problems which feature high levels of uncertainty, vague, ness and which must employ information which is subject to variable interpretation. One must accept the diffi- culty of gauging its usefulness (i.e. the confidence one can have in its validity and precision of its results). 2% It should be profitable to apply Delphi in con- junction with a more concrete procedure which works from empirical "real world" data instead of from subjective, albeit expert, perception. 3. Its eventual usefulness will be judged by its performance rather than by any abstract analysis of its worth. 4 . Research in Delphi should stress its psychological aspects in terms of improved communication among experts concerned with forecasting and establishing objectives for the future rather than in mathematical terms. 81 Putnam, S.H.: Intra-urban Employment Forecasting Models: A Review And Suggested Model: Journal of the American Institute of Planners. 38:216-30, July 1972. This paper is interesting to health planners chiefly as an example of large scale urban models employed for forecast, ing purposes. A number of problems faced by the forecasters of intra-urban employment levels -- such as availability of labor and proximity to markets -- may be analogous to those of the projectors of health service needs and demands, where util- ization of facilities is affected both by their availability and their proximity to high density population centers. The article is divided into three sections: (1) Classifi- cation of employment types, in which a key issue -- which types of employment are to be included in the model and how each is to be handled -- is discussed; (2) Review of existing models, in which those models for which facility location has market sensitivity implications are compared with those for which lo- cation has significance for ' the model's resulting employment projections; (3) A suggested intra-urban employment model. In this section, the author develops a model which "should be capable of producing reliable and accurate forecasts of vir- tually all intra-urban employment distributions". Since the modelling process developed in this paper is concerned with the issue of where to locate facilities in order to take demand for their employment (services) into account most effectively, it may serve as a useful analog for nealth planners in developing regional models of service demand. 8 2 Richardson, A. H. and H. E. Freeman: Evaluation of Medical Care Utilization by Interview Surveys. Medical Care. 10 (4) 1357-62, July-August 1972. Efforts to improve the delivery of health care services and to provide medical care for all groups in the community have utilized a large number of interview surveys on the utilization of medical resources. Though the use of interview survey data is justified and defended on the basis of feasibility and cost, there is considerable skepticism about their value. This study of octogenarian retirees from the automobile industry, with access to both survey interview and medical record data before and after implementation of a broader health befefit plan, supports the use of interview survey data for reasonable conclusions of utilization at various points in time when health benefits, or access to them, have been changed. Though skepticism should continue in the use of self-reports for making precise estimates, if the alternatives are no evaluative research or the use of self-reports in studying the impact of health programs, the data suggest that the decision to proceed with solely interview survey data is the correct alternative. 83 Richardson, J. D. and Scutchfield, F. D.: Priorities in Health Care: The Consumer's Viewpoint in an Appalachian Community. American Journal of Public Health. 63(1) :79-82, January 1973. The objective of this study is to determine at the grass roots level the health care priorities of the consumer in an Appalachian community. "Needs" in this study are evaluated in two ways: according to the perceptions of the consumer himself and by clinicians evaluating the health status of the consumers. The methodology of the study included interviewing a stratified random sample of households in Rowan County, Kentucky. The sample included 110 households (349 people), or 2.5 percent of the population. Results indicate that clinicians and consumers focus upon different areas of need: consumers perceive their needs and problems to be in the area of economics and delivery of services, while the clinical evaluation focuses upon apparent disfunctioning of consumer's physical organ systems. 84 Sackman, Harold: Delphi Critique: Expert Opinion, Forecasting a and Group Process, Lexington sooks: Lexington, Mass.) 1975. This monograph questions the validity of the Delphi method with respect to its use of building and systematic incorporation of expert opinion in forecasting. Since the results obtained from the method are widely used to produce forecasts, develop quantitative estimates and make qualitative evaluations of situations in which expert opinion appears to apply, the method should be subjected to methodolog- ical critique in the author's view. In critiquing Delphi, the author compares and evaluates the Delphi method against key standards in professional questionnaire design, obtained from "Standards for Educational and Psychologi- cal Tests and Manuals". published by the American Psychological Association (1966). The standards are divided into six classi- fications: Interpretive Standards, Empirical Validity, Standards for Use of Experts, Theoretical Standards, Questionnaire Reliabil- ity and Experimental Sampling Standards. When analyzed against these criteria, Delphi reveals its deficiencies in both scientif- ic questionnaire development, and in the experimently controlled and replicable application of questionnaires. The author discusses both advantages and disadvantages of Delphi. The advantages which make this method desirable and extensively used are: low cost, administrative ease, versatile application, ease of respondent understanding, and a minimum amount of respondent time and effort. Disadvantages of the method lie, says the author, in the weakness of its underlying assumptions. Delphi theory main- tains that consensus among "experts" improves forecasting accuracy and is superior to individual opinion; and that expert opinion is in and of itself scientifically tenable, hence the results reached through Delphi are scientifically valid. These assumptions are disputed in this critique. The author believes that in view of the tenuous nature of assumptions, the results from Delphi frequently will prove to be flawed in the short-run and worthless in the long-run. The advantages are, therefore, thought by the author to be inconse- quential in light of the many flaws and inherent weakness of the method . The overall objectives and output of the method on the surface appear to meet the needs of area wide planners. However, as Sackman's analysis of Delphi shows, the method re- 85 quires substantial refinement if it is to be suitable for use by professionals, funding agencies, institutions, corporations, and the government. The potential benefits of Delphi are small and do not justify the costs of achieving them as the method is generally employed. 86 Sears, D. W.: Elderly Housing: A Need Assessment Technique. The Cerontologist. - 12(2):182-87, April 1971. Although the technique presented in this article for determining the need for housing for the elderly is not related to medical care services, it does provide a logical, step-by-step framework for gathering data and making decisions along the process. The need determination technique, if adopted for health services, provides a method for evaluating the necessity and the distribution of additional health care facilities. 87 Shillif, TI.E. and Smith, R. D.: A Forecasting Method for Setting Short-Range Research Objectives. Research Management. 12:24-34, March 1972. This article develops the use of morphological analysis in forecasting the short-range future. While research and development in plastics is the subject being forecasted, the basic method can be applied as well in other fields such as health. The authors suggest that unstructured, intuitive think- ing such as used in delphic analogy and scenario methods of forecasting tends to overlook a proportion of feasible alter- natives. Trend extrapolations oversimplify reality by limit- ing consideration to only a few parameter combinations. Morphological thinking, say the authors, provides a systematic classification of all significant possibilities and thereby holds great potential for short-range, normative forecasting. Morphological analysis is defined as the study of relations apart from functions with an attempt to classify agcording to similarities of fundamental characteristics. Analysis of the illustrative problem proceeds in two phases: (1) Demand Analy- sis, and (2) Morphological Analysis. In Phase One, multiple regression is used to identify those independent variables which best explain differences in demands. It is noted that for some situations computer pro- grams; such as the Automatic Interaction Detector System, which employ non-symetrical brancing based on variance analysis to determine which set of independent variables best explains changes in a particular dependent variable may serve better than multiple regression. Once explanatory variables are defined and the range over which their value may vary is specified, one enters Phase Two. In this phase, a matrix is constructed, arraying the various variables by row and their respective ranges of value by column. From this matrix, one can determine the set of all possible interactions (if this set is not too large), the effect the independent variables are likely to have on the dependent variable in each set of interactions examined, gnd the rough likelihood of any particular combination of variables. This information equips the analyst to forecast short-run changes 88 in demand more effectively. Other planning environments may be extremely complicated, admitting a large number of variables. However, the basic method can be applied equally well in most of them, the authors believe. 89 Swain, R. W.: "A General Model for Hospital Census Prediction and Control." Abstract of a Paper Presented at the Operations Research Society Convention, 1974; Boston,. Massachusetts. This paper describes a general system for predicting the demand on hospital resources in the near future and describes a control strategy which can be employed to achieve the goal of maximum hospital revenue subject to limitations on the use of hospital resources. Resources considered may include beds, or time, staffing and similar types of resources. Although an explicit definition of demand is not presented in the paper, several bases for predicting the demand for resources, including beds, are presented. These bases include the patient's length of stay and his diagnosis or service. The model developed in the paper "... is an additive combi- nation of rather simple models. This additive composition allows for the simple tailoring of the resulting system to a particular institution's needs Prediction formulas in the overall model include those for: (1) discharges (2) admissions and (3) transfers. Another aspect of the model's flexibility is that it uses nearly any data pertaining to length of stay, including subjective estimates by health care professionals and historical records and "... permits it to be used in controlling total hospital beds or in controlling the beds used by a particular service. The model may be used to specify the policy of elective admissions which would yield maximum census while permitting only a specified chance of violating prespecified census limits in any of the upcoming days A simple version of the described model has been used on a daily basis to provide planning information for the Ohio State University Hospital and the results are discussed in the paper. Future uses of an modifications to the model are also presented. 90 Van de Van, A., and A. L. Dbelbecq:- "Nominal Vs. Interacting Group Processes. For Committee Decision-Making." Academy of Management Journal. 14:3, 1971 The authors analyze the effectiveness of interacting and nominal group processes for use by problem-solving committees. The differences between interacting groups-interaction takes place spontaneously among participants, and nominal groups-individual efforts are "summed up" to produce a group outcome; are examined from a theoretical viewpoint to determine the justification and need for greater behavior structure in decision phases of the group process. The effectiveness of the two methods is measured by comparing their performance in terms of: (i) mean number of unique ideas, (ii) mean total number of ideas, (iii) quality of ideas. The outcome produced by these measures of performance indicates that nominal groups are superior to brainstorming groups in producing ideas relevant to a particular problem. Interacting groups are not as effective due to their tendency to inhibit individual participation which consequently generates fewer creative and original ideas. The nature of the nominal group process tends to overcome the inhibiting influences by providing a more con- ducive atmosphere for creative decision-making. In problem-solving effectiveness, the authors conclude that nominal groups produce a greater number of dimensions in the understanding of the problem and, therefore, a higher quality of suggestions. On fact-finding tasks, the final choices made through the interacting group process are better than decisions based solely on the summation of individual judgements. A series of guidelines are formulated after examining committee decision-making effectiveness which stress an optimal combination of the two processes in creative problem-solving. 91 Van de Van, A.H. and Delbecq, A.: "The Nominal Group As A Research Instrument For Laboratory Health Studies." AJPH. March 1972, pp. 337-342. The Nominal Group Process is a qualitative approach toward judgemental problem exploration. It is particularly applicable to health planning situations that are subjective and judgemental in character. One disadvantage in group problem-solving, specifically in the health planning field, is that each group has its own per- spective on the qualitative and quantitative parameters under investigation. This often results in an inadequate understanding of the problem area, an understanding necessary before the collec- tion of data can begin. The Nominal Group Method seeks to alleviate this structural problem by providing planners with a depersonalized and therefore equitable method of problem solving. Small groups, composed of individuals, whose backgrounds differ in expertise, vested interests ideological outlooks, respond to exploratory questions, in order to define critical elements of a problem. The individuals contribute to the process separately in writing and then verbally without interacting with other group members. Only after the individual responses are recorded, does a discussion between the members on the items proposed take place. The potential uses and advantages of the method include: (i) it provides a mechanism for critical problems to be singled out, ranked, rated and clarified by a target group; (ii) it explores subjective and objective dimensions of a problem; (iii) it develops hypothesis for formulating survey questions; (iv) 'it is low in cost:" (v) it requires a short time period for implementation; (vi) it is a' tool to aggregate individual inputs at one time while eliminating unbalanced participation and dominant personalities. 9 2 ¥ett, D.E., brabek, Lr, Intriligator, M.D., et. al.: "A Microsimulation Model of the Health Care System in the U.S." Paper presented at the Econometric Society Meetings: 29 December 1973; New York. This paper summarizes the efforts of "...the Human Resources Research Center (HRRC) to develop a microsimulation model of the U.S. health care system which will permit forecasts of the complex interrelations between the demand for and supply of health services and health manpower". The paper does not explicitly define demand but utilizes various measures such as patient days and outpatient visits to clinics and emergency rooms. The microsimulation model described in the paper consists of five components or submodels. "The first submodel generates a population of consumers or individuals who demand medical ser- vices." This model generates annual estimates of the U.S. pop- ulation according to age, sex, race, income, and condition or diagnosis in response to "...exogenously determined birth, death, and immigration rates". The second submodel provides annual estimates of the physician population by specialty, age, activ- ity, and domestic-foreign trained. The output of these two submodels provides input for the remaining three submodels. The physician services submodel provides the linkage between the physician and consumer populations. The submodel treats "... (1) the demand for outpatient physician visits; (2) the supply of such visits, and the derived demand for full-time equivalent RNs, LPNs, allied health professionals, and non- professional personnel; and (3) the market interactions for supply and demand for services via price adjustment. "The hospital services submodel, which treats the inter- actions between consumers and hospitals, consists of three components: (1) the demand for inpatient days; (2) the supply of such services, and the derived demands for full-time equivalent RNs, LPNs, allied health professionals, and non- professional personnel; and (3) the resulting price and wage adjustments over time." The nonphysician manpower submodel calculates the supply of such manpower based on the demands by 28 types of physician practices and 16 types of hospitals calculated in the previous two submodels. "The interaction of the demand and supply for each type of health manpower results in changes in wage rates, which in turn, affect labor force participation rates." Partial results of a test of the model during a historical period, 1960-1970, are presented, as are possible future uses of the model. 93 CHAPTER III--ADDITIONAL REFERENCES REVIEWED Following is a bibliography from which the literature review has been conducted. This bibliography is organized under the following headings: books, chapters from books, journal articles, and papers. Books Anthony, Robert N., Planning and Control Systems: _ A Framework for Analysis, (Boston: Harvard Graduate School of Public Admin- istration) 1965 Blum, Henrik L., Planning for Health. Development and Application of Social Change Theory, (New York: - Human Sciences Press) 1974 Braybrooke, David and Charles Lindbloom, A Strategy for Decision, (New York: The Free Press) 1963 Cyert, Richard, and James G. March, A Behavioral Theory of the Firm, (New York: Wiley) 1963 Diesing, Paul, Reason in Society: Five Types of Decisions and Their Social Conditions, (Urbana: University of Illinois Press)} 1962 Donabedian, Avedis, Aspects of Medical Care Administration- Specifying Requirements for Health Care, (Cambridge: - Harvard University Press) 1973 Friend, J. K., and Jessop, W. N., Local Government and Strategic Choice, (Beverly Hills: Sage Publications) 1969 Goldman, Thomas A., Cost-Effectiveness Analysis: New Approaches to Decision-Making, (New York: Praeger) 1967 Gross, Betram, The Managing of Organizations, (New York: Free Press of Glencoe) 1964 Haefele, Edwin T. (ed.), The Governance of Common Property Resources, (Washington, D.C.: institute for the Future) 1974 Haveman, Robert H., Julius Margolis (eds.), Public Expenditures and Policy Analysis, (Chicago: Rand McNally) 1970 Hecht, Karl, What is Done to Protect the People's Health? The Health Service of the German Democratic Republic, (Dresden: Grafischer Grossbetrieb Volkerfreundschaft) 1974 9 4 Hitch, Charles and Roland McKean, The Economics of Defense in the Nuclear Age, (Cambridge: Harvard University Press) 1960 Isard, Walter, Methods of Regional Analysis, (New York: John Wiley & Sons) 1960 Lindblocm, Charles, The Policy-Making Process, (Englewood Cliffs: Prentice-Hall) 1968 Masse, P., Optimal Investment Decisions: Rules for Action and Criteria for Choice, (New Jersey: Prentice-Hall) 1962 Mishan, E.J., Cost-Benefit Analysis, (London: George Allen & Unwin Ltd.) 1971 Olson, Marcus Jr., The Logic of Coleletive Action:. Public Goods and the Theory of Groups, (Cambridge: Harvard University Press) 1965 Schultze, Charles, The Politics and Economics of Public Spending, (Washington, D.C.: The Brookings Institution) 1969 Simon, Herbert A., Administrative Behavior: A Study of Decision- Making Processes in Administrative Organizations, (New York: MacMillan) 1957 Weisbrod, B.A., Economics of Public Health-Measuring the Impact of Diseases, (Philadelphia: University of Pennsylvania Press) 1960 Book Chapters Arnold, Mary F.: Health Officer Decision-Making -- A Case Study , in Arnold, Mary (ed.), Administering Health Systems: Issues and Perspectives, (Chicago: Aldine-Atherton) 1971, pp. 377-97 Blackman, Allan: Philosophy of Planning as an Instrument of Social Change. Health Planning 1969, H. L. Blum and Associates (San Francisco: Western Regional Office, American Public Health Asso- ciation) 1969 Blankenship, Vaughn L.: Organizational Decision-Making, in Arnold, Mary, F. (ed.), Administering Health Systems: _ Issues and Perspec- tives (Chicago: 'Aldine-Atherton) I971, pp. 377-97 I Déane, Burton W., S. J; Mantel, Jr.: -A Model for Evaluating Costs of Implementing Community Projects. Analysis for Planning Program- ming, Budgeting, M. Alfrandry-Alexander (ed.), (Washington, D.C.: Operations Research Council) 1968 95 Klarman, Herbert E.: Syphillis Control Programs, in Robert Dorfman (ed.) Measuring the Benefits of Government Investment, (Wash- ington, D.C.: The Brookings Institution) 1965 March, James G.: Some Recent Substantive and Methodological Developments in the Theory of Organizational Decision-Making, in Austin Ranney, (ed.}, Essays On The Behaviroal Study of Politics, (Urbana: University of Illinois Press) 1962 Ozbekhan, Hasan: Towards a Geneal Theory of Planning, in Perspectives of Planning, (Paris: Organization for Economic Cooperation and Development) 1969 Schelling, T.C., The Life You Save May Be Your Own, in Problems in Public Expenditure Analysis, Samuel B. Chase, (ed.), (Washington, D.C.: The Brookings Institution) 1968 Stimson, David H.: Health Agency Decision-Making: An Operations Research Perspective, in Arnold, Mary, (ed.) Administering Health Systems,, (jaw York: Aldine Atherton. Inc.) 1971. pp. 398-433 Journal Articles Alonso, W.: Predicting Best With Imperfect Data. Journal of the American Institute of Planners, July, 1968, pp. 248-55 Amara, R.C., et al.: Some Views on the Use of Expert Judgment. Technological Forecasting and Social Change, 3 (3) 1972 Blum, H.L.: Priority Setting for Problems, Solutions, and Projects by Means of Selected Criteria. International Journal of Health Services, February, 1972, pp. 85-99. Coleman, James S.: Colelctive Decisions. Sociological Inquiry, 34 (1) 1964, pp. 166-80 Curran, William J.: Are Planning Agencies Ready for Community Decision-Making Roles? Health Planning Perspectives, 6 (December, 1970) pp. 1548-53 Hertz, D.B.: Risk Analysis in Capital Investment. Harvard Business Review, January/February, 1964 Hodge, Gerald: Use and Misuse Of Measurement Scales in City Planning. JAIP, XXIX, May, 1963, pp. 112-21 Lindbloom, Charles: The Science of Muddling Through. Public Administration Review, XIX (1959), pp. 79-88 96 Mayo, Louis, H. and Ernest M. Jones, Legal-Policy Decision Process: Alternative Thinking and the Prediciton Function. The Georgetown Law Review, XXXII (October) 1964, pp. 318-56 Mushkin, Selma: Health as an Investment. Journal of Political Economy, 1962 Newell, Allen, J.C., Shaw and Herbert Simon: Elements of a Theory of Human Problem Solving. Psychological Review, 65:3 (1958), pp. 159-63 Pill, Juri: The Delphi Method -- Substance, Context, A Critique and an Annotated Bibliography. Socioeconomic Planning Sciences, 2 (1971) pp. 57-71 Reiner, I.8., et al.: Client Analysis and the Planning of Public Programs. JAIP, November, 1963, pp. 270-282 Sullivan, D.F.: A Single Index of Mortality and Morbidity, HSMHA Health Reports, April, 1971, pp. 347-54 Turoff, M., An Alternative Approach to Cross-Impact Analysis. Technological Forecasting and Social Change, Vol. 3, No. 3 (1972), pp. 309-39 Wildavsky, Aaron: Does Planning Work? The Public Interest, (Summer, 1971) pp. 95-104 Wildavsky, Aaron: The Political Economy of Efficiency; Cost- Benefit Analysis, Systems Analysis, and Program Budgeting. Public Administration Review, 26 (December, 1966), pp. 292-311 Pagers Bainbridge, J. and S. Sapirie, Health Project Management: _ A Manual of Procedures for Formulating and Implementing Health Projects (Geneva: World Health Organization) 1974 Feldstein, M.S., et al., Resource Allocation Model for Public Health Planning, Supplement to Volume 48 of the Bulletin of the World Health Organization, 1973 Grundy, F. and W. A. Reinke, Health Practice Research and Formalized Managerial Methods, Public Health Papers, No. 51, World Health Organization, 1973 Hilleboe, H.E., et al., Approaches to National Health Planning, Public Health Papers, No. 46, World Health Organization, 1972 97 Participants in Study Tour Sponsored by the World Health Organiza-, tion, Health Service in the USSR, Public Health Papers, No. 3, World Health Organization, 1960 Popov, G. A., Principles of Health Planning in the USSR, Public Health Papers, World Health Organization, No. 43, 1971 Reutlinger, S.: Techniques for Project Appraisal Under Uncertainty, (World Bank Staff Study No. 10) Johns Hopkins Press, Baltimore, 1970 Second Report of the Expert Committee on Public Health Administra-, tion, Methodology of Planning an Integrated Health Programme for Rural Areas, Technical Report Series No. 83, World Health Organiza- tion, 1954 U.S. Congress, Senate, Committee in Government Operations, Planning- Programming-Budgeting Initial Memorandum, Selected Comment, Official Documents. (Washington, D.C.: USGPO), July, 1967 Winslow, E. A., The Cost of Sickness and the Price of Health, Monograph Series, No. 7, World Health Organization, 1951 World Health Assembly, Modern Management Methods and the Organiza- tion of Health Services, Public Health Papers, No. 55, World Health Organization, 1974 9 8 *U.S. GOVERNMENT PRINTING OPPICE : 1978 O-723-019/386