The Aging Population in the Twenty-First Century: Statistics for Health Policy (1988)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Introduction and Summary of Recommendations In the coming decade, the nation's decision makers will continue to be challenged by changing demands for social and health services due to the anticipated rapid rate of growth of the elderly (65 years or older) and especially the oldest-old (85 years or older) populations. In the 12 years until the year 2000, it is anticipated that the very old (80 years or older) U.S. population will be the largest single federal entitlement group, consuming $82.8 billion (1984 dolIars) in benefits (Torrey, 1985~. The growth of these populations will be even more rapid soon after the turn of the century, as the post World War II baby-boom cohorts become elderly. To respond wisely to the multifaceted demands of an aging U.S. society, it is crucial to have the data and information necessary to make difficult choices about resource allocation and program struc- ture. To produce this information: (1) appropriate data must be collected and (2) appropriate models and statistical methods must be applied to the analysis of those data. Although there is currently a sizable expenditure on data collection activities, the investment is minuscule compared with the size of the federal and state programs that this information is used to manage and direct. Furthermore, recent scientific evidence on the modifiable nature of health and functional transitions among the old and oldest-old populations sug- gests that many aspects of current data collection are inadequate to support policy analysis. Efforts dedicated to methodological de- velopment of unproved analytic and forecasting tools are even more deficient. Data production and analysis, as well as methodological 1

2 AGING POPULATION IN THE TWENTY-FIRST CENTURY research, will need careful and thorough review, and modification in the light of that review, to maximize the utility of the data collected. BACKGROUND Concern about the inadequacies of statistical information and methodology available for policy decisions for the elderly is wide- spread. Seven federal agencies that shared this concern the Vet- erans Administration and six agencies of the U.S. Department of Health and Human Services: the Health Care Financing Admin- istration, the National Center for Health Statistics, the National Institute of Mental Health, the National Institute on Aging, the Of- fice of the Assistant Secretary for Planning and Evaluation, and the Social Security Administration joined forces and sponsored a study by the National Research Council to address these problems. The panel was charged with the following major activities: 1. To determine the data requirements for policy development for health care of the elderly during the next decade; to assess the statistical adequacy of current data sources pertaining to the health care of the elderly; and to identify major shortcomings and recom- mend appropriate remedies and actions; 2. To identify the essential components of a comprehensive pro- gram of statistics on the elderly that can be implemented within a decentralized statistical system (assuming continuation of the cur- rent decentralized system) and that would provide adequate data on aging for all functional areas and to recommend changes and procedures that would facilitate integrating data from the various components; and 3. To determine whether changes or refinements are needed in the statistical methodology used in health policy analysis or in the planning and administration of programs for the elderly and to rec- ornmend actions or further research. The panel approached these charges in a period when budgetary con- straints assumed special significance because of their implications for statistical activities of federal agencies. The panel therefore placed emphasis on modifications to existing national statistical programs and surveys and those that are getting under way an approach that led to numerous recommendations but was designed to recognize cur- rent budget stringencies. Accordingly, the panel's recommendations,

INTROD ACTION AND SUMMARY OF RECOMMENDATIONS 3 whenever possible, are formulated to capitalize on and enhance avail- able data resources, surveys, and administrative records. Neverthe- less, meeting the needs for data is not cost-free, and new budgetary support for existing agencies will be essential. In making its recommendations, the pane! has dealt with data requirements for the immediate future and has also provided a long- range planning guide for the collection and analysis of statistical ata for policy analysis for the elderly population over the next decade. We recognize that many of the surveys and administrative record systems that are the subject of our recommendations are not limited to the elderly and that their implementation will have to fit into the agencies' comprehensive plans for improving their statistical systems. With careful planning, and some additional resources, it should be possible to implement all the recommendations before the twenty-first century. Several recommendations specify the frequency with which surveys should be conducted. The recommended cycles represent the best judgment of pane! members, but we recognize that agencies will have to examine the proposed cycles more intensively, drawing on input from nongovernmental sources. In the course of its work, the panel paid particular attention to overlap and duplication between the major data systems and con- cluded that data gaps are much more serious than overlap. There are surveys that appear to cover many of the same areas, but they differ substantially on specific components of concern to policy makers, particularly in content and the populations covered. Many of the data gaps are the result of program adjustments made by federal sta- tistical agencies in response to the changing fiscal environment of the last decade and the accompanying reductions in budgetary support. In many cases, these changes were made hastily, without adequate consideration of their short-term and Tong-term consequences, and thus have not supported the development of information for well- informed public policy debate. Because of budget reductions, the statistical agencies have been forced to make changes in programs and policies in recent years- changes that affect the availability of statistics on the elderly. These changes involve one or more of the following: (1) changing the fo- cus from policy-oriented statistical programs to those that support the administrative aspects of government; (2) reducing the periodic- ity (or frequency of data collection) of major surveys; (3) reducing the coverage of surveys, through deletion of specific subpopulations from the universes of interest or through reduction of sample size;

4 AGING POPULATION IN THE TWENTY-FIRST CENTURY (4) reducing efforts in the areas of data collection operations, data processing, and data dissemination; (5) reducing the timeliness of data dissemination, of both hard-copy reports and public-use data files; and (6) postponing or eliminating the regular review of data needs in developing areas, usually in the interest of protecting core programs within agencies. Specific examples of the types of adjust- ments cited above are considered in Appendix B. which examines the programs of those agencies responsible for the information bases used for supporting policy development in aging. In short, the panel faced a complex task. Our deliberations took place during a cost-conscious era, there are serious gaps in the data needed for policy analysis, and we recognized the need to strike a balance between short- and Tong-term concerns in formulating our recommendations. As a basis for its work the pane! compiled an inventory (Na- tional Research Council, 1986) of 117 data sets related directly or indirectly to health of the elderly- data sets that remained follow- ing a decade of changes the agencies have been forced to make in response to budget stringency. Simply put, a systematic review of these data sets led the panel to the conclusion that the available information is inadequate for policy analysis. A major concern of the panel relates to remedying this situation and ensuring availability of information adequate in scope and timeliness for policy purposes. Careful attention has been paid to the most urgent improvements needed in the existing surveys and in the use that can be made of systems of administrative records. No new large information sys- tem is recommended: the pane! considered it more cost-effective and more acceptable politically to obtain the new data required for policy use by adding to existing surveys. Recommendations in the report address data requirements that can be met by federal information systems that are national in coverage. The pane! notes that data sets generated in epidemiological studies, longitudinal community surveys, and evaluation studies may be equally important for policy purposes, but they fall outside the scope of the charge of the panel. HEALTH POLICY ISSUES The Changing Policy Agenda Changes in socioeconomic, demographic, and health care trends frequently raise new health policy issues. Other issues come to the

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 5 attention of policy makers because of the efforts of interest groups and public officials or as a result of particular dramatic events. For example, the release of Medicare patients from hospitals who still require skilled nursing care with no provision for such care gave focus to concerns of the Senate Special Committee on Aging early in 1985 about the prospective payment system currently in use by the Medicare program. Such broad issues as the cost, the quality, or the availability of health care are continuously part of the general or systemic policy agenda of the country. When a particular issue becomes defined as a crisis, however, sufficient political interest may be generated to move alternative solutions to the more active policy agenda, where solutions are debated in the form of new legislation, appropriations, or regulations (Elder and Coble, 1984~. Otherwise the issue will disappear from the policy agenda, although it may well reemerge on the active agenda with changes in the economic or political climate. For example, in the 1970s the level and trend of national ex- penditures for health care moved cost containment and payment for health care to the top of the active policy agenda. In fact, in the late 1970s, the two major policy initiatives of the Department of Health and Human Services were hospital cost containment and national health insurance (Stoiber, 1979~. Reduction of waste and duplication in the hospital sector was intended to generate the funds for a national health insurance program that would fill the unmet needs for health care. Legislation for national health insurance was considered by Congress but, after much debate, no legislation was adopted and further consideration of national health insurance was dropped. Concern about hospital cost containment continued and was addressed by the Social Security Amendments of 1983, which gave rise to the prospective payment system for Medicare based on diagnosis-related groups. The second significant trend in the prior decade was rapid change in the relation of health care delivery to the reimbursement for ser- vices. Several special issues were raised concerning the new forms of delivery and reimbursement for services, which included health maintenance organizations (HMOs), surgical centers, the expanded role of hospitals for delivery of ambulatory care, hospices, walk-in centers, and multihospital systems. For example, in 1972, legislation was passed to provide grants or loans to support the development of HMOs. Also during the 1970s the federal Administration on Ag- ing began to emphasize that health counseling and health screening

6 AGING POPULATION IN THE TWENTY-FIRST CENTURY should be included in the social services for the elderly provided at the local level under the Older Americans Act. Yet another exam- ple was the federal government's response to the shortage of physi- cians in the 1960s and 1970s. Federal programs for medical students were legislated programs that stimulated increased enrollments in medical schools and the establishment of new medical services and ultimately resulted in the physician surplus of the 1980s. The full impact of the growing number of physicians on the organization of medical care is not yet clear (Tinier, 1986~. Policy is still being driven by financing and organizational ques- tions, but today there is increased emphasis on examining quaTity- of-care issues, e.g., what health care are the elderly receiving, what good is it doing, and how are the government programs and policies affecting people's lives. This change is a response to issues that have been raised about costs, access to new types of services and technolo- gies, and matching services to the needs of the population generally and, in some instances, to the needs of the elderly specifically. implications for the Panel's Approach The first task of the pane! was to make recommendations to improve the data base for health policy for an aging population during the next decade. The pane! recognized the difficulties arid pitfalls of trying to identify specific policy issues and their related data requirements when looking forward a decade. Such an approach would not only be difficult, but it might also be counterproductive, because policy issues change in unpredictable ways. Accordingly, the panel's recommendations are directed to modifications in the federal statistical system that would ensure the availability of basic data that are relevant and important to the following set of generic health policy issues for an aging society: Who will pay for health care for the elderly and how will it be financed? What alternative health delivery systems can be developed to meet the needs for health care for the elderly in the next decade? How can health promotion and disease prevention be ad- vanced among the elderly? How much health care for the elderly will be needed in the next decade? What will the health status of the elderly be in the next decade?

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 7 What are the differences in health status among subgroups of the elderly? What policies and programs are needed to ensure that the elderly receive health care of appropriate quality? What provisions need to be made for population subgroups to avoid problems of equity and access in health care for the elderly? Are health programs and benefits for the elderly equitably distributed across the states? What is the impact of geo- graphic variation? . Are there unintended effects of government programs for health care for the elderly? As stated, these health policy issues are sufficiently broad so as to include the specific policy issues of both the past and current decades and can be expected to include those of the next decade. The special policy issues that arise during a decade are shaped by the trends of the time. In forecasting health policy choices for the l990s, a variety of demographic, economic, organizational, and attitudinal trends can be expected to shape health policy issues in the l990s and to change the U.S. health care system (Blendon, 1986~. In addition to the growing number of oldest-old, specific trends can be expected to influence policy issues for health care for the elderly, which are detailed below. The continuing problem of rising health care costs. Despite the continuing efforts to contain them, the nation's health care costs "are projected to increase from $387 billion in 1984 to $660 billion in 1990, reaching almost $2 trillion by 2000" (Blendon, 1986:67~. The aging of the population is not the only factor contributing to the rising cost: improved and costly medical technologies and the volume and complexity of new types of services may serve to increase health care costs (Blendon, 1986~. The growing concern with quality of care and cost-benefit ra- lios, i.e., what government and private purchasers are receiving for their money. Efforts to manage costs and distinguish high-quality care from less effective and less efficient care are stimulating efforts to monitor all sites of care. The monitoring of hospital performance was mandated through provisions of the Social Security Amendments of 1972, requiring that professional standards review organizations (PSROs) be established to ensure that health care services provided

8 AGING POPULATION IN THE TWENTY-FIRST CENTURY under the Medicare and Medicaid programs Were medically neces- sary, conformed to appropriate professional standards of quality, and were delivered in the most effective and economical manner possible" (National Health Policy Forum, 1985:4~. Further changes were made in the TEFRA (Tax Equity and Fiscal Responsibility Act) legislation of 1982, replacing the PSROs with professional review organizations (PROs), which have stronger regulatory powers. In 1986 Congress extended the PRO scope to include review of HMOs and other pre- paid plans under contract to provide care to Medicare beneficiaries. The changes also provided for targeted reviews of nursing and home health care services, and (at a later implementation date) services in physicians' offices. As the role of peer review organizations ex- pands, new measures to quantify quality for different services and in different settings will be needed (National Health Policy Forum, 1986~. Continuing rapid change in the organizational structure of the health service delivery system. During the past few years, enroll- ment of Medicare beneficiaries in HMOS has increased because of federal incentives to HMOs to provide capitated comprehensive care for Medicare patients. In addition, partly in response to the prospec- tive payment system for Medicare inpatient services in hospitals, many new types of out-of-hospital health care facilities have received increased attention. These facilities include ambulatory care centers, diagnostic or imaging centers, hospices, rehabilitation institutes, sur- gical centers, and urgent care centers. These facilities may be owned by hospitals or hospital chains, for-profit companies, or groups of physicians (Blendon, 1986~. Such changes are still evolving and there is no clear indication of the extent to which they will affect the pattern and sources of care for the elderly. Since it is not possible to predict the effects of these interacting trends and the specific policy issues they may generate, the pane! focused on providing a general-purpose statistical base for each of the generic policy issues, keeping in mind the current issues and the foreseeable issues on the horizon. We relied on our knowledge about aging and our experience with data requirements and data gaps in past policy analyses to make determinations about the relevance and importance of data for these generic policy issues. General-purpose statistics and statistics derived from adminis- trative records for federal programs cannot be expected to supply all the data needed by policy analysts, although the two types of data taken together can answer many questions. For example, the policy

INTR OD ACTION AND SUMMARY OF RECOMMENDA TIONS 9 question, "Should we be spending more or less on health care for the elderly?" requires information on who is being served and how much is being spent in the aggregate and for which services. Aggre- gate data from administrative records (e.g., Medicare and Medicaid) can provide information on who is being served and how much is being spent under the programs. The National Medical Expenditure Survey, which surveys a sample of the general population as well as people in institutions, will provide information on expenditures not included in government programs and information on health care expenditures by elderly people who are not participants in these programs. The policy question, "Are the Medicare and Medicaid programs meeting the needs of different categories of the elderly (ru- ral elderly, chronically disabled elderly, etc.), as 'need' is defined by different policy analysts?" calls for evaluations and would probably require special studies, although the national statistics on health ser- vice utilization for various subgroups of the population might be the starting point in designing the special study. The policy question, "How much need will there be for health care for the elderly in future years and what will it cost?" may call for a forecast of the quantity of health care that will be needed assuming policy remains unchanged, in which case demographic data, in conjunction with data from the Health Care Financing Administration on Medicare and Medicaid utilization and costs, are essential to answer this question. Policy analysts frequently require forecasts of the quantity and costs of health care under various hypothetical scenarios, which may have never occurred. A current example of this is found in the eight alternative reform proposals under consideration for redesign- ing Medicare, which are discussed in a report for the National Health Policy Forum (Etheredge, 1987~. One aspect of Medicare benefits under consideration is "Should Medicare eligibility continue to start at age 65, or at younger or older ages?" Forecasts of savings by adopt- ing cutoffs at older ages can be prepared from Medicare data, but the quantity of health services that ended if the eligibility age were changed to 62 (as suggested in one proposal) is more tenuous. The basic data source would be the health service utilization data for ages 62-64 in the National Medical Expenditure Survey. However, since utilization of health services might change under expanded Medicare coverage, data from experience under past policy, although relevant, are not conclusive. In fact, many policy issues cannot be adequately addressed by drawing on general-purpose statistics and data from administrative records. In some cases, a special supplemental survey

10 AGING POPULATIONIN THE TWENTY-FIRST CENTURY may be needed. In others, targeted experiments are required to learn about the impact of policy options in many settings. RECOMMENDATIONS The multiplicity of agencies concerned with statistics on the el- derly and the numerous surveys and administrative record systems involved led the panel to make a large number of recommendations. The organizational structure of the report also contributed to the number of recommendations, since several chapters on different pol- icy areas may have recommendations about a single survey. To guide the reader, we summarize the recommendations in three ways: first, we present a set of 12 specific priority recommendations; second, we present 5 general recommendations that summarize the chapter recommendations; finally, we present a table with the 79 individual recommendations that are discussed in the chapters. Priority Recommendations The pane] selected these recommendations from the large num- ber of detailed recommendations as those that should be given prior- ity because they will provide the data most urgently needed for health policy for an aging society. The priority recommendations consist of both parts and combinations of recommendations that appear in the chapters. The recommendations are presented in the context of the policy areas they address and are followed by the rationale for each recommendation. Financing of Medical Care: For the aging population, the major policy issues that wiB confront the United States during the next decade are the cost of supplying health care to the elderly, who wiB pay for that care, and how it wiB be financed. Policy development for these issues requires trend data on the health expenditures of the elderly and also longitudinal data on the use of medical care as a person ages. In addition, evaluation of policy questions related to the Medicare program would be facilitated by improved access to the Medicare statistical system. Priority Recommendation 1: The panel recommends (a) continuation of the periodic survey of national medical care expenditures, the periodicity to be determined in relation to policy needs and timing of other health-related surveys and

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS (b) that the population age 55 and over in the 1987 survey be identified and followed by linking to administrative records, including Medicare reimbursements from Health Care Fi- nancing Administration records, and, to the extent feasi- ble, Medicaid reimbursements from state record systems and work history and pension benefits from Social Security Ad- ministration records. In addition, the National Death Index of the National Center for Health Statistics and state health department death records should be used to identify the year and cause of death of each sampled person. Priority Recommendation 2: The panel recommends that the Health Care Financing Administration develop files de- signed for easy access to the Medicare Statistical System (including the Medicare Automated Data Retrieval System) that would facilitate use by researchers for policy analysis related to the Medicare program. 11 Trend data on health expenditures for the elderly. National health expenditure data are widely used by policy makers to evaluate the extent of coverage of existing public programs, such as Medicare and Medicaid, and to estimate the total health care costs of the elderly population by type of expenditure and source of funds. The data also serve as a basis for assessing the possible consequences of changes in public policy programs. The major sources of such data have been the 1980 National Medical Care Utilization and Ex- penditures Survey (NMCUES) and the 1977 National Medical Care Expenditures Survey (NMCES). These surveys and the Medicare files are the primary data sources for estimating cost and coverage of program changes, such as the various proposals for catastrophic health care coverage currently being considered by Congress. More current data will become available in 1988 from the ongoing 1987 Na- tional Medical Expenditure Survey (NMES). The 1980s have been and continue to be a decade of far-reaching changes in the structure of the health care delivery system, private health insurance, federal and state and local health care programs, as well as in the demo- graphic composition of the nation. How these changes affect the kinds and amounts of health care Americans use, how they will pay for it, and the implications of further changes in health care policy are questions that NMES data and the analyses based on them will help to answer.

12 AGING POPULATION IN THE TWENTY-FIRST CENTURY However, to capture trends in utilization and expenditure pat- terns in response to future changes in delivery and payment systems over time and to meet policy needs for current data, a survey of na- tional medical care expenditures should be conducted periodically. In order to take full advantage of the various types of relevant exper- tise in different agencies, future national medical care expenditure surveys should be joint efforts of appropriate federal agencies, in- cluding the National Center for Health Services Research and Health Care Technology, the Health Care Financing Administration, and the National Center for Health Statistics. Longitudinal data for the elderly in NMES through follow-up studies using administrative records. A major gap in understanding the impact of changes in the financing of medical care services for individuals and their families is the lack of longitudinal data on the use of and expenditures for medical care services as a person ages and is at risk of chronic illness requiring acute medical and long-term care services. Since significant changes in health status frequently start to occur in the decade prior to age 65, it is important for longitudinal studies to start following people at age 55. Extensive additional data about the individuals sampled in the NMES are available from administrative records of the Health Care Financing Administration, the Social Security Adrn~nistration, and the Internal Revenue Service. Information on the cause of death can be determined through use of the National Death Index maintained by the National Center for Health Statistics and death records in the state health departments. Use of the rich information available from administrative records can enhance the usefulness of survey data collected in the NMES for longitudinal analyses at very low cost. Improved Access to the Medicare Statistical System. The Medi- care Statistical System (MSS) was designed to provide data to mea- sure and evaluate the operation and effectiveness of the Medicare program. It has also been a major source of information for eval- uating policy questions relating to the equity and efficiency of the Medicare program. For example, data on the distribution of Medi- care reimbursements for survivors and decedents and for type of service provide useful data on the high use and costs of medical care services in the last year of life. Medical reimbursements per capita by state and county are useful measures of equity. Provider cer- tification data related to population are Report ant measures of the supply of facilities and services and their variation across the country. Geographic variations in surgical procedures among the elderly are

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 13 important indicators of practice patterns. The MSS, a by-product of administrative record systems, includes data on each Medicare en- rollee, institutional providers of service, and records of the utilization of services that can be matched to enrollees and providers. Despite the obvious attractiveness of the Medicare files for an- alytic purposes, these files were established primarily to assist with administration and monitoring of the Medicare program. To make the Medicare administrative data more accessible and less costly for research use, a new file has been designed the Medicare Auto- mated Data Retrieval System (MADRS). The MADRS is intended to reorganize and merge claims files for Medicare Part A (hospital insurance) and Part B (supplementary medical insurance for physi- cians' services, outpatient services, etc.) to shorten search time and will contain all Medicare chains data and patient provider identifiers. The MADRS file will enable researchers to identify groups of special interest and analyze them by age, by sex, and/or by admit- ting diagnosis, for example, and examine the care they have received over time. The development of the MADRS is a positive step toward facilitating the analysis of Medicare data, thus gaining a better un- derstanding of trends in health services utilization among the elderly. The Health Care Financing Administration should develop new approaches to improving access to its administrative files by non- federal users, such as making arrangements for intergovernmental personnel appointments, interns, and postdoctoral fellows. Organization and Delivery of Health Services: Cost concerns ffe- quently lead to msues concerning the organization and delivery of health services. Rapid changes are taking place in the health care delivery system; policy Brokers need to monitor these changes and to assess the ejects they are having on health care for the elderly. Detailed data are needed on changes taking place in the health care delivery system and their effects on health care for the elderly. Priority Recommendation 3: The pane} recommends that federal agencies give high priority to reviewing and modify- ing the contents of administrative record systems, provider- based surveys, and, to the extent feasible, population-based surveys to reflect the rapidly changing patterns in health ser- vice delivery. Standard definitions and formats for recording information on the type of health plan should be used by all agencies collecting such data.

14 AGING POPULATION IN THE TWENTY-FIRST CENTURY The adoption of the diagnosis-related groups and other prospec- tive payment systems for hospital care, the growth of for-profit health care, the adoption of business-oriented approaches by health care providers, and the growth of competition in the medical care mar- ket may cause unintended side-effects, such as barriers to access to health care for certain population subgroups and deterioration in quality of care. The preceding recommendation, if implemented, would provide the data required to monitor the effects of the rapid changes in the health care delivery systems. The modifications rec- ommended should improve the data base for monitoring changes in the organization and delivery of health care service, while adoption of standard definitions will enable interviewers, respondents, and coders to distinguish among the various types of health plans in use, including the varieties of capitated plans; and to detect differences in the cost-sharing provisions of these plans. Provision for [ong-term Care: A major factor In the cost of supply- ~ng health care for the elderly is the need for and cost of long-term care. Such care is also the primary concern in many of the issues pertaining to the organization and delivery of care. The concern about provision of long-term care will lead to policy about its organization, delivery, and financing. Policy makers will need current information about the use of long-term care and data for projecting the need for it as characteristics of the elderly and the nature of their support system changes. Priority Recommendation 4: The panel recommends that the National Nursing Home Survey (a) be conducted on a 3-5 year cycle and (b) that a subsample of admission cohorts in the 1985 survey serve as a panel for a longitudinal study of institutionalized persons and that the records of this subsample be linked with Medicare files. Timely data on [ong-term care. The provision of long-term care for the chronically ill and feeble is a major issue related to the aging population. The complex of long-term care issues has surfaced in the current congressional debates on the need for increased Medicare coverage for catastrophic illness, highlighting the need for data on long-term care. Decisions will have to be made, either now or in the near future, about the organization, delivery, and financing of long-term care. The National Nursing Home Survey, conducted by the National Center for Health Statistics in 1973-1974, 1977, and

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 15 1985, provides information useful for such decision making. This in- formation includes data on nursing homes, their services, staffs, and financial characteristics, and on the personal and health character- istics of residents and discharges (1985 only). These surveys should be conducted more often than every ~8 years, however, because of increases in the costs of nursing home care and frequent changes in policy. A Year interval, or at most, a 5-year interval, is needed to address issues on changes in health for the elderly, outcomes of care, targeting and substitution of services, and the impact of policy changes. Data for projecting the need for long-term care. By definition, Tong-term care is care over an extended time period. To project the need for long-term care, data are required for a relatively Tong period on changes in the characteristics of the elderly population, their use of services, and the nature of their support system, as well as changes in this system, both formal and informal. The 1977 and 1985 National Nursing Home Surveys are useful models for longitudinal surveys of the institutionalized elderly. The 1985 survey included a follow-up of discharged patients. Data were collected from next of kin of both residents and discharged patients on functional disability at time of admission, caregiver stress, and the family's view of the reasons for admission. Surveys of this type conducted at relatively frequent intervals with longitudinal follow-up as recommended above would provide much of the data required for policy decisions on Tong-term care. Over time, the National Nursing Home Survey should be ex- panded to include all types of long-term care institutions (i.e., chronic disease hospitals, mental health facilities, rehabilitation centers, board and care homes, psychiatric halfway houses, and residential facilities. Expanding the coverage of the survey in this manner would make it possible to compare the costs of other modes of providing Tong-term care services, costs that are essential to planning the or- ganization and delivery of long-term care services. The National Center for Health Statistics took the first step in expanding coverage with the 1980~1981 development efforts to expand the coverage of the Master Facilities List. It would also be desirable to link the records of the longitudinal subsample with Medicaid records, since Medicaid is a major source of support for nursing home care. At present, how- ever, there are problems in linking into Medicaid information on a national scale. Health Promotion and Disease Prevention in the Elderly: A wise

16 AGING POPULATION IN THE TWENTY-FIRST CENTURY investment strategy would support spending health care doBare to lengthen the period of life spent in vigorous health. Expenditures for health promotion may wed lead to more years of good health than equal expenditures on medical care directed to treating disabilities and serious or irreversible disease. :Related policy issues include how health promotion activities can be advanced among the elderly, whether the federal govermnent should fund health promotion and disease prevention activities, and, if 80, which ones. The paucity of data concerning health promotion and disease prevention activities among the elderly creates barriers to the devel- opment of policies that recognize this aspect of health care. Priority Recommendation 5: The pane] recommends (a) that modules of health promotion and disease prevention items (including those concerned with attitudes, knowledge, and behavior) be developed that are appropriate for the el- derly and subgroups of the elderly population at risk for par- ticular diseases, illnesses, disabilities, or conditions, which can be used with a variety of population-based surveys; (b) that these modules be tested on relevant segments of the elderly population; and (c) that successful modules be in- corporated in population-based surveys such as the National Health Interview Survey and the National Health and Nu- trition Examination Survey, or as supplements to them. Health promotion and disease prevention are a major emerging theme in geriatric medicine and health care generally. Although such efforts have typically been targeted at younger persons, there is growing evidence that this approach is both appropriate and feasible for persons 65 and over. Health promotion and disease prevention are one of a number of possible strategies to deal with the prevalence of chronic illness and multiple chronic illnesses or functional impair- ments among the elderly. While this approach will neither replace medical care for the treatment of acute diseases, nor for acute flare- ups of chronic illness, it has promise for reducing or delaying the incidence and prevalence of chronic and acute disease among both the general population and the elderly. Despite the obvious importance of health promotion and disease prevention, data are not routinely available through national data systems on the extent to which the population is informed as to the causes of preventable illnesses and conditions and the actions they might take to reduce their own risks of developing such illnesses

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 17 and accompanying impairments. Development of the recommended modules will require cooperation and coordination of effort by sev- eral agencies, including the Office of Health Promotion and Disease Prevention, the National Center for Health Services Research and Health Care Technology, and institutes within the National Insti- tutes of Health. Monitoring and Forecasting Health Status and the Utilization of Services: Assessment of the cost of supplying health care to the elderly will require estimates of the health care needs of the elderly, what types of care wiD be needed, and to what extent needs are met. This requires periodic monitoring of the health status of the elderly and their use of health care services to detect trends and to forecast *nature health status and utilization of services. The information needed for such monitoring and forecasting includes longitudinal data on aging and information on the relation between health status and health services utilization. In addition, an evaluation of the theory, methodology, and data requirements for forecasting the characteristics of the aging population is needed. Priority Recommendation 6: The pane} recommends that the National Center for Health Statistics (a) continue to col- lect, every two years, health and other relevant information (hospital and physician care, income, housing, informal sup- port, and the use of community services) from all persons age 55 and older in the original sample of persons in the 1984 Supplement on Aging of the National Health Interview Survey, now known as the Longitudinal Study on Aging, (b) use the National Health Interview Survey as a base for a periodic Supplement on Aging for the noninstitutionalized population age 55 and over, and (c) conduct follow-up inter- views of the supplemental sample to provide longitudinal as well as cross-sectional data on the older population. Priority Recommendation 7: The pane! recommends that (a) linkage with Medicare records be performed on a routine basis for persons age 65 and over who are respondents to pop- ulation surveys that collect health data and (b) the Health Care Financing Administration and the National Center for Health Statistics explore linking the Health Interview Sur- vey with the Medicare Automated Data Retrieval System, when the latter becomes operational.

18 AGING POPULATION IN THE TWENTY-FIRST CENTURY Priority Recommendation 8: The pane! recommends that a study to evaluate theoretical, methodological, and data requirements for forecasting the characteristics of the aging population be undertaken. This would include theoretical and practical considerations for evaluating the sensitivity of forecasts to underlying assumptions. Longitudinal data on aging. Many questions concerning the health and functional status of the elderly remain unanswered at this time. In particular, the dynamic process of how individuals change in health and functional status as they grow older, and the nature of the critical transitions they experience as they pass from one health or functional status to another, are not well understood. This information is basic in determining whether and how various subgroups of the elderly population are changing in their patterns of morbidity, functioning, and quality of life, information essential for determination of future health care needs and expenditures as longevity of the population increases. Repeated measures of the same individual over time are needed to elucidate these processes and changes and provide information for health policy making. A promising medium for obtaining such information is the 1984 Supplement on Aging (SOA) of the National Health Interview Sur- vey. This supplement collected data on a sample of 8,000 persons age 55 and over in the community. In addition to information on health status and hospital and physician care, this survey collected infor- mation on income, housing, informal support, and the use of social services. Persons age 70 and over in 1974 were reinterviewed in 1986. Additional reinterviewing of this sample and periodic repetition of the SOA with longitudinal follow-ups would permit measurement of some of the transitions occurring in the elderly population. The follow-up could be carried out by a combination of telephone calls and personal interviews with varying frequencies appropriate to the age of the sample member. Rapid changes in health, medical care needs, living arrangements, and available support are characteristic of the older population and must be understood to forecast the health services needs of the future elderly. Relationship between health status and health services utilization. It is Report ant to be able to relate the health and personal status of the elderly to their utilization of health services so that trends in health status can be used to forecast patterns of change in health service utilization. Many population surveys, including the National Health Interview Survey, the National Medical Expenditure Survey,

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 19 the National Nursing Home Survey, and the Survey on Income and Program Participation, obtain important demographic and health status information that can be linked with Medicare file data for the same individuals. Such linkages could provide a rich source of longitudinal information on health services utilization by the elderly, making it possible to relate health status and other characteristics to subsequent use of health services. Forecasts of the characteristics of the aging population. Many of the factors that are of interest for an aging population inter- act with one another. One way to account for such interactions is through global models that attempt to integrate different submodels or modules into a common projection framework. Integrated efforts are necessary to forecast health status, functional limitations, and support systems available for older persons. These forecasts can be useful not only for probing important policy issues related to health care expenditures and welfare programs, but also for improved mor- taTity forecasts in general population projections. Ensuring That Health Care Meets Quality Standards: Questions about the quality of care are an integral part of the debate on cost containment policies for health care for the elderly and the conse- quences of adopted policies. Policy makers will require indicators on quality of care to inform this debate. A concerted effort is needed to develop measures of quality of care of elderly patients that are suitable for national data systems. Priority Recommendation 9: The pane] recommends that agencies having cognizance of national data systems, whether based on administrative records or survey data, (a) jointly review these data systems to identify items that can be used in developing quality of care measures and to identify miss- ing items that are needed to produce such measures and (b) make provision for collecting the needed data and for producing quality of care measures. Quality of care is emerging as a matter of central interest for policies directed at controlling costs or improving services for specific subgroups of the elderly. An Office of Technology Assessment report to Congress (Office of Technology Assessment, 1985a) pointed out that assessing the effects of Medicare's prospective payment system on quality of care is critical. The Institute of Medicine's recent report on quality of care in nursing homes (Institute of Medicine,

20 AGING POPULATION IN THE TWENTY-FIRST CENTURY 1986) identifies the need for a system of acquiring and using resident assessment data. The context for quality of care issues will broaden as policies to control costs are extended to sectors other than hospital care and as alternative sources of care become available, particularly for long-term care patients. National data systems can produce the information needed to de- rive quality of care measures to determine how quality varies within the population and how it may be influenced by sources of care and policies that affect services received. Many of the specific items re- quired are included in surveys that are already operational or about to start; others are dependent on the extension of existing items or the introduction of new ones. Candidates are found in the Na- tional Health Interview Survey and its supplements on aging and prevention, the National Nursing Home Survey, the National Health and Nutrition Examination Survey, the National Ambulatory Medi- cal Care Survey (NCHS), the Long-Term Care Survey (HCFA), and in the forthcoming National Medical Expenditures Survey (NCHSR and HCFA). Not only should national data systems be able to measure qual- ity directly, but they should also provide broad indicators of quality of care that will alert policy makers that other information is needed. For example, in 1986, when the Health Care Financing Administra- tion released mortality data (an outcome measure for quality of care) for 2,300 hospitals and identified 269 institutions in which death rates were significantly higher or lower than others, the response was to fo- cus attention on how patients are managed in hospitals. The hospital industry argued that the data were misleading because they did not take into consideration factors such as severity of illness or the extent to which high-risk procedures are performed in specialty hospitals. This led to increased efforts to develop measures of severity of illness (Horn and Horn, 1986~. The pane] considers the development of measures of quality of care of elderly patients that are suitable for national data systems to be a high-priority effort. At the same time, we recognize the inherent difficulties because there is no accepted definition of quality of care and because of the conceptual problem of quantifying good quality. Improving the Quality of Health Policy Analysis: Procedural rec- ommendations to unpro~re access to data or the capability to analyze data from multiple sources cut across policy moues. Increased detail In publication of data, improved access to data, and pawning And

INTRODUCTION AND SUMMARY OFRECOM:MENDATIONS 21 coordination of data collections to provide the capability of ~nter- relat~ng the data Mom different sources can improve the quality of policy analysis and extend the scope of policy-oriented research. The pane! makes three recommendations on these activities. Priority Recommendation 10: The pane! recommends that agencies that collect and disseminate data on the total popu- lation or administrative data by age (a) use Year age groups for the publication of data up to age 90 and larger intervals thereafter, except when limited by privacy or confidential- ity regulations and (b) provide data on geographic areas, income, and other economic characteristics at the greatest level of detail consistent with the protection of confidential- ity. Priority Recommendation 11: The pane! recommends that the plans of a federal agency for the creation or continuation of a data system include a plan for issuance of publications and public use tapes following its creation or updating. The plan shouIcl include explicit procedures for disseminating the data and should provide for widespread access and timely, thorough, and accurate data dissemination. Priority Recommendation 12: The panel recommends that a mechanism be provided for discussion and coordination of data needs, standardized definitions and classifications, priority identification, and production of data relating to an · · . aging society. Data detail. As the elderly population increases in size and as a proportion of the total population, it becomes increasingly impor- tant to know more about specific groups within that population. For different purposes, additional detail may be needed on the oldest-old, five-year age groups among the elderly, women, minorities, persons with health problems, the poor elderly, and other special popula- tions. Frequently information issued in aggregate form does not in- clude enough detail to identify these subpopulations. For data from administrative records or from the decennial census, the failure to meet these needs of policy analysts is a matter of presentation. The single most useful standard that agencies could adopt is to provide information for standard age categories in publications and public use files.

22 AGING POPULATION IN THE TWENTY-FIRST CENTURY Improved access to data. To make data more accessible to re- searchers and to increase the usefulness of data for policy purposes, tape files for research use need to be created in federal statistical and administrative programs without violating confidentiality. The agencies should also conduct more outreach programs to inform po- tential users of the availability of tapes and provide training on their use. For example, the National Center for Health Statistics has a program for informing potential users of the availability of data tapes and how to use them; a series of "Advance Data" releases are also distributed before final figures are available. Planning and coordination of data collections. In a decentralized federal data system, such as exists in the United States, it is especially important to have some mechanism for planning and coordinating the disparate activities of different federal agencies that are engaged in collecting information about the elderly. The purpose of the newly established Forum on Aging-Related Statistics, which includes representatives from all the federal agencies that collect statistical information on the elderly, is to encourage cooperation among these agencies in the development, collection, analysis, and dissemination of data on the older population. General Recommendations The preceding priority recommendations come from a substan- tially larger set of specific recommendations that reflect the full range of the panel's findings and conclusions. The detailed recornmenda- tions are intended to provide the basis for making current data col- lection efforts more effective and for providing necessary supplements to those efforts. Table 1.1 at the end of this chapter lists short titles of all the recommendations given in each chapter with the agencies whose involvement would be required to implement each recornmen- dation. The priority recommendations discussed above are indicated by a P in the table. Recommendations that entail conducting surveys appear in the table under the appropriate statistical agency~ies), but not under research sponsoring agencies that might provide support for the surveys. The panel recognizes that such support is essential for near-term implementation of some recommendations, particularly those assigned a priority. In this section, five general recommendations provide an inte- grated picture of the full set of recommendations. They represent

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 23 the collection and integration of a number of specific recommenda- tions that follow in subsequent chapters. . Develop and maintain a core group of national longitudinal health surveys to study health transitions and health service needs among the elderly. Nationally representative longitudinal data are needed to accu- rately describe the health and functional transitions of the old and oldest-old U.S. populations. Purely cross-sectional data may pro- vide erroneous and incomplete insights to these processes and miss important underlying changes. To develop accurate nationally rep- resentative data on the health and functional changes of the elderly, and the relation of those changes to recent life expectancy changes, existing national surveys of the major population groups of the el- derly need to be supplemented by longitudinal follow-up of selected survey cohorts. And these reinterview programs need to be supple- mented by the linkage of the survey records to Medicare Part A and B data files and to death certificates identified through the National Death Index. The minnnal core of longitudinal surveys for adequate popu- lation coverage are the survey programs that produced the 1984 Health Interview Survey, the Supplement on Aging (SOA) and the 1986 follow-up Longitudinal Study on Aging (LSOA), the 1985 Na- tional Nursing Home Survey (NNHS) and the 1987 follow-up, and the 1982 and 1984 National Long-Term Care Surveys (NETCS), and the follow-up studies of the National Health and Nutrition Examina- tion Survey (NHANES) populations. The first three of these surveys cover the major elderly population components (i.e., the SOA-LSOA, the general elderly population; the NNHS, the institutionalized pop- ulation; the NETCS, the community-disabled elderly). The fourth, the NHANES with expanded foBow-up studies, would provide longi- tudinal biomedical and epidemiological information. Although these four surveys provide an excellent basis for a national set of longitu- dinal surveys, they have been hampered in design and coverage by inadequate and unstable funding. More adequate levels of funding could make these surveys more effective by: ensuring follow-ups at regular periods, improving instrument design and measurement, ensuring cross-sectional representativeness of the total U.S. elderly population by enrolling samples of the population that become elderly between survey dates,

24 AGING POPULATION IN THE TWENTY-FIRST CENTURY making the sample design more effective by conducting more intensive sampling of special target populations (e.g., over- sampling the population age 85 and older in future supple- ments on aging in the Health Interview Survey), providing for assessment of short-term changes by more fre- quent follow-ups of sample subpopulations. l~trocluce design changes in other major survey programs to improve their usefulness for studying the health of the elderly, health care expenditures, and quality of care. A number of existing major surveys such as the National Med- ical Expenditure Survey (on the health expenditures of the general U.S. population), the National Ambulatory Medical Care Survey (NAMCS) (on the use of office-based physician services, health prob- lems, and diagnostic and therapeutic services received by the general U.S. population), and the Survey of Income and Program Partici- pation (SIPP) (on the income, assets, and program participation of both young and old Americans), provide valuable information on a series of special topics. Furthermore, certain surveys like the NHIS, the NNHS and the NMES represent long-term and continuous survey efforts. Thus these surveys merit continuing support both because of the information they currently generate and because of the in- creased value of that information in the context of a long time series of such surveys, extending from 1957 in the case of the NHIS, 1963 in the case of the NNHS, and 1977 in the case of the NMES. Changes have been made in these surveys recently and additional changes are needed. Though crucial to continue, these surveys must be carefully assessed for needed changes in the measurement, survey, and sample design to improve their effectiveness. Specific examples of possible needed changes are: (a) increased oversampling of the old and oldest-old in the NMES and the SIPP, (b) repeating the health module at least twice in the SIPP survey cycle, (c) improving the quality of care measures in the NHIS, and (~) developing and introducing health promotion and disease prevention items in the NHIS, the NHANES, or supplements to them, and the NAMCS. In general the adequacy of sample sizes for special subpopulations (such as the 85 years and older population) should be evaluated in all such surveys, and their periodicity and the quality of their health instruments need to be assessed. Before such changes could be introduced two types of research

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 25 will be needed. The first is evaluation studies of the current con- tent, sample design, periodicity, population coverage, and overlap of these surveys. Such studies will require special and adequate funding to the appropriate technical groups within the National Center for Health Statistics and the Census Bureau. The results of such stud- ies must be reviewed for technical merit by an independent group before instituting changes to ensure that (a) changes will improve the technical quality of the effort and (b) the substantive continuity of the existing time series will be maintained. Part of this evalua- tion process should involve assessment by the appropriate agencies of whether the necessary data are being collected to serve national health policy needs, to monitor the health of the nation, and for scientific purposes. For example, in terms of meeting public health and scientific needs, such surveys should be reviewed by technical groups appointed by the federal Forum on Aging-Related Statistics, which involves the Census Bureau, the National Center for Health Statistics, the National Institute on Aging, and other agencies. The second type of studies is scientific investigations of data col- lection methodologies (e.g., studies of instrumentation, sample de- sign, and periodicity). The studies should be supported through the appropriate extramural research programs at the National Institute on Aging and the National Science Foundation using supplementary funding identified specifically for these purposes. . Standardize definitions and instrumentation across data collec tion and data dissemination activities. A number of efforts should be made to ensure comparability between different data collection and dissemination efforts. For ex- ample, a standard definition of long-term care should be adopted. To give another example, in disseminating population data adequate age detail is necessary (e.g., by five-year age groups to age 90 and larger intervals thereafter). By establishing standard categories with adequate detail in published documents the utility of multiple data collection efforts will be enhanced. There is much to be gained from the standardization of certain basic content modules to facilitate comparability across surveys. For example, basic income and assets questions might be standardized, as might certain basic instrumentation for functional assessment, for health promotion and disease prevention activities, and for measuring quality of care. The efforts of interagency committees to generate

26 AGING POPULATION IN THE TWENTY-FIRST CENTURY such standardization should be facilitated by appropriate levels of support. . Enprove mechanisms for the broad disee?n;nation of aD types of data collected with federal support. With both the existing and proposed federal data collection efforts, there should be rapid and broad circulation of data generated from all types of federally supported research and data collection. That is, part of our current inability to respond to certain major policy and scientific questions is that data already collected are not available until several years after collection. Federal policy should mandate that all federally funded survey efforts be made publicly available an soon as possible. Included in such releases should also be all linked information, e.g., linked Medicare service use records. Federal data policies must be changed to make such provisions. Provide an adequate ferret of support for statistical and forecast- ~ng research. Federaldata collection efforts, to tee maximally effective, must be supported by appropriate methodological research on the analysis of longitudinal data from repeated cross-sectional and longitudinal surveys. Such methodological research should incorporate research on design of data collection such as the dimensions of measurement, survey design, sample design, linkage to continuous longitudinal ad- rn~nistrative records, the issue of the correct temporal spacing of repeated surveys, treatment of missing data in longitudinal studies, and the calculation of power for multivariate cross-temporal data. Analytic research issues center around retrieving estimates of pa- rameters from process models of health and functional changes at advanced ages from incompletely observed processes. In addition, research is needed on how multiple surveys can be integrated in co- ordinated analyses and on the use of such data in forecasting. There is also a need for research on the forecasting of health, disability, and mortality in the U.S. elderly and oldest-old populations. The research should examine such questions as how to represent the ef- fects of social and economic changes on health transitions, how to determine the uncertainty of forecasts, and how new data will affect the degree of uncertainty. Such methodological research should be supported by extramural research programs at the National Science Foundation ant] the National Institute on Aging.

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS A Word About C08t8 27 When grouped according to their expected costs, the panel's recommendations consist of 4 groups: (1) recommendations that fall within the scope of routine operations, either administrative or research and development, and could be implemented within the current staffing configurations of the relevant agency; (2) recom- mendations that require special studies, analyses, or projections; (3) recommendations for research to develop additional survey method- ology or methodology for statistical and policy analysis; and (4) recommendations to augment the scope or content of data sources, to increase the periodicity of surveys, to follow samples of existing surveys to obtain longitudinal data, and to increase survey sample size to provide age detail for the elderly. Half the recommendations fall in the first group, and although these recommendations are not cost-free, they do not entail addi- tional costs. A fifth of the recommendations fall in the second and third groups and require special studies or research projects that would probably cost between $200,000 and $300,000 (in 1987 dol- lars) each. For many of the 25 recommendations in the fourth group, the associated costs may exceed $500,000 each. It is important to remember that the costs would be spread over the 21 agencies to which recommendations are addressed as well as the breadth of ef- forts covered and the range and importance of issues to which these data can be applied. The panel discussed the feasibility of estimating the financial and staff resources that would be required to implement this fourth group of recommendations. However, there are several arguments against estimating resource requirements. The panel's recornmen- dations were intended for implementation over the next decade as they fit into agency plans, and estimates made today could be unre- alistic in the future and might even be a barrier to implementation. The panel might have requested the federal agencies designated to implement the recommendations to make cost estimates, but it is unlikely that they would be willing to do so unless plans for the rec- ommended activity were currently under discussion. On the basis of these arguments, and since our charge did not call for cost estimates, the pane] decided against making estimates of the funds required to implement the recommendations in group 4. In any event, the sup- plemental funds required to implement the panel's recommendations are minuscule compared with the amount of money expended on

28 AGING POPULATION IN THE TWENTY-FIRST CENTURY major federal and state health programs- far less than would be rea- sonable for the combined purposes of policy development, program management, evaluation, and research for a major public enterprise. It is logical to ask whether the benefits that would accrue from implementing these recommendations warrant the cost. Some of the benefits are obvious. The information that would be produced would enhance policy makers' ability to target health care to those in need, increase equity in the allocation of Medicaid funds, control the escalating costs of Medicare, and identify alternative forms of Tong-term care that might provide cost savings or increased quality of life for the elderly. Since it is particularly difficult to place dollar values on quality of life and access to health care, it was not feasible for the panel to quantify this type of benefit. Therefore, we assigned priority to recommendations that respond to the most pressing needs for information to meet the overall goals of medical care for the elderly. These goals include the following: to enable the elderly to stay healthy and functionally independent as long as possible, to provide access to good medical care of whatever type is appropriate (preventive, long-term, short-term, acute), and to provide care in the least restrictive and most cost-effective and appropriate environment (Somers, 1987~. ORGANIZATION OF THE REPORT Structure of the Report The panel's three charges are addressed in separate chapters of the report. Chapter 2 summarizes the social, economic, and demo- graphic trends that trigger policy issues in several topics related to the health of the elderly. Chapters 3-9 discuss the data requirements for policy development for health care for the elderly during the next decade and recommend actions needed to ensure that these data will be available. These chapters correspond to the topic "health policy issues for the elderly" discussed earlier in this chapter. Chapter 10 addresses statistical problems that cut across the federal statisti- cal system. Chapter 11 discusses methodological issues common to many of the data collection efforts and to the use of the data in policy analysis. We summarize the chapter contents below.

INTROD ACTION AND SUMMARY OF RECOMMENDATIONS 29 Chapter 2: Social, Economic, and Demographic Changes Among the Elderly Not only is the population aging, but also other major life-course changes are occurring. Age-related roles are less predictable than they were in the past. Late adulthood has a perplexing number of possible life patterns. Trends that should be taken into account in policy analysis on aging include demographic trends, morbidity and mortality patterns, changes in family structure, an increase in the interruption of marital careers, increased mobility, changes in living arrangements, labor force participation, and other activity patterns of the elderly. The magnitudes of these changes and the acceleration of some of them underscore the importance of well-designed statistical programs to monitor the characteristics of the aging population. The trends and projections based on the trends not only influence program size, but also, in some cases, may influence policy. Chapters 3 :'nd 4: Health Status, Quality of Life, and Health Uan . . . SlilOllS In order to develop policy for health care for the elderly, infor- mation is required on trends in health status and utilization of health care to be able to make projections. While there is no doubt that life expectancy is on the rise, a major unanswered question is whether increased longevity will be accompanied by vitality and "active life expectancy," or rather by a prolongation of the period of morbidity and disability and corresponding increments in health services uti- lization and health care expenditures. Will the added years of life be spent in wellness and the enjoyment of high quality of life, or in relative deprivation due to the presence of irremediable chronic ailments and functional impairments? Chapters 3 and 4 review the data series available for monitoring health status, quality of life, and health transitions and make recornrnendations to fill serious data gaps. Chapter 5: Health Promotion and Disease Prevention In this chapter health promotion and disease prevention are defined and their relevance and the appropriateness of this approach for the elderly are discussed. The chapter explores the need for data concerning the extent to which the elderly are knowledgeable about

30 AGING POPULATIONIN THE TWENTY-FIRST CENTURY and engage in health promotion and disease prevention activities and the services in these areas provided to the elderly by health care personnel and institutional providers. Chapter 6: Quality of Care A major concern in health care is the quality of care. Policies di- rected at controlling costs, such as hospital diagnostic-related groups, or at improving services for specific groups of the population, such as the elderly, have made this a matter of immediate interest. In assess- ing the quality of care provided, not only is the technical performance of health care providers at issue, but also the extent to which the health care system reaches those in need. This chapter underscores the importance of developing indicators of quality of care. , , ~ Chapter 7: [ong-term Care The purposes of long-term care services are to assist persons who have lost some capacity for self-care to cope with their disabilities, to decrease their dependence on others, and to narrow the gap between their actual and potential functioning. This chapter reviews the capacity of national data systems to track users of Tong-term care services over time and to measure and report the services they use. Chapters 8 and 9: Fmanc~ng and Utilization of Health Care Services Chapter 8 describes current mechanisms for financing health care services for the elderly, including the federal Medicare and state Medicaid programs, Veterans Administration services, and private health insurance programs. Health care expenditures for the elderly are also summarized. The chapter explores the potential for extract- ing more timely and comprehensive financing and expenditure data from the Medicare Statistical System and from planned and existing national health surveys. Particular attention is paid to data that can be used to document the impact of recent changes in health care delivery systems and financing mechanisms, and the effects of the "spend-down" phenomenon on family members. Chapter 9 reviews the many factors in addition to need that in- fluence the amount and types of health care services utilized by older persons. The focus ~ on improvements in national data, particularly

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 31 in the areas of changing service delivery patterns, dental care, reha- bilitative care, mental health services, and access to care on the part of poor and other special populations such as the rural elderly. Chapter 10: Enhancing the Usefi~Iness of National Statistical Sys- tems Attention is drawn to the need for the federal government to provide greater interagency coordination, cooperation, and planning among those federal agencies and programs that produce statistics on the elderly. Since the United States has a decentralized statisti- cal system, special efforts are required to ensure that data policies across federal agencies and programs are consistent. Coordination of statistics on the elderly is needed with respect to content; coverage; age detail; definitions, concepts, and classification; periodicity; and access to data. The chapter also argues for increased use of linkage among data sets as a mechanism to enhance information about the elderly and suggests ways to improve the vital statistics program and data on causes of death. Furthermore, health-related policy, like other domains of public policy in recent years, is jointly planned and carried out at state and local levels (e.g., Medicaid and the Older Americans Act); consequently the chapter also considers the importance of the collection and timely dissemination of subnationa] data. Chapter 11: Methodological and Statistical Moues The final chapter discusses methodological and statistical issues relevant to an aging society under four headings: the design and analysis of longitudinal studies, linkage of data bases, forecasting the characteristics of future elderly cohorts, and the quantification of uncertainty of projections. Projecting and Dete~n~ng Statistical Needs Several considerations cut across the data requirements for health policy issues addressed in subsequent chapters of the report and pro- vide a unifying framework for the report. These considerations in- clude the need for information on the healthy as well as the dependent elderly, data on functioning to supplement diagnostic information, longitudinal clata on the elderly, information on the responsiveness

32 AGING POPULATION IN THE TWENTY-FIRST CENTURY of the elderly to both acute and long-term care systems, better in- formation about long-term care services provided to the elderly, data on risk factors for the elderly, better data on special subgroups of the elderly, and forecasts of the characteristics of future cohorts of elderly. Formation on the Healthy Elderly Contrary to popular belief or stereotype, the elderly are highly diverse with respect to just about any characteristic, whether it be health or functional status, economic or social circumstances, or living arrangements. Frailty, ill health, and disability increase, on the average, with age, and data on this process are needed. But many elderly persons are in excellent health and function independently in most if not all areas of their lives. These persons often experience a high quality of life and engage in satisfying productive activity, underscoring the need to focus greater attention in data collection efforts on the well and independent elderly. The level of dependence of the elderly in activities of daily living, their physical and/or mental impairments, and their illnesses and disabilities have been the focus of many studies or surveys. Collecting more information on the well and independent elderly will contribute to a better understanding of the factors that lead to positive health and the conditions under which the elderly are able to cope electively with advancing age. Data on Functioning to Supplement Diagnostic Formation While health status and the presence or absence of illness and disease are major sources of satisfaction or dissatisfaction with life for older persons, it is the ability to function independently-physically, cognitively, emotionally, and socially that is crucial to a feeling of well-being and a high quality of life. Diagnostic information is therefore important, as is increased understanding of the effect of different conditions on functional status. However, specific diagnoses often bear no relation to the ability to function nor to the needs and capabilities of an elderly person. New measures of functional status and their incorporation into surveys involving the elderly are needed. Longitudinal Data on the Elderly There is a dearth of information about what happens to the elderly as they age-how their health, functional status, economic

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 33 status, living arrangements, need for long-term care, and utilization of health services interact and change. There is now considerable evidence that there is plasticity in the aging process that disease processes can be significantly altered or reversed with appropriate therapies. Improvements in health and function over time, as well as decrements, need to be determined for the elderly population. Only repeated measurements of these phenomena over time as a person gets older will yield information on the transitions in health status that occur and circumstances that influence these transitions. Information on the Responsiveness of Both Acute and Long-term Care Systems Too often the public equates geriatric medicine with long-term or chronic care. A substantial portion of the health care utilization and expenditures of the elderly, however, is for acute care provided by short-stay hospitals or in the physician's office. The elderly are likely to need both acute and long-term care services at different points in their lives. Because of the many changes in the health care delivery system, in the costs and financing of health care services for the elderly, and in the types and qualifications of health care providers available to them, it Is especially important to monitor en cl document the impact of these changes on the types, amounts, and quality of acute and long-term care services utilized by the elderly. With the introduction of prospective payment systems, based on hospital diagnosis-related groups, there is some evidence that elderly persons are being discharged from acute care facilities sooner than was the case previously, and thus there is the strong possibility that long-term care facilities are receiving more acutely ill persons than they did in the past. Furthermore, the acute care needs of the elderly are different from those of younger persons, and improved data regarding the use and costs of acute and long-term care services by the elderly could assist health care facilities and providers in planning for increased demands on them from the growing elderly population. Better Formation About [ong-term Care Services Long-term health care refers to the professional or personal ser- vices required on a recurring or continuous basis by an individual because of chronic or permanent physical or mental impairment.

34 AGING POPULATION IN THE TWENTY-FIRST CENTURY There is considerable diversity in the types of long-term care ser- vices used by the elderly, the settings in which they are provided, the costs of such care to the elderly and their families, and how costs are met. These services may range from informal supports provided by family or friends in the person's home, to paid medical and related social services provided on a periodic or routine basis in the home, to community-based adult day care programs, to cus- todial, intermediate, or skilled care provided in residential facilities. Because the elderly are the population group most likely to need long-term care, it is unport ant that surveys- and particularly longi- tudinal surveys provide more complete information concerning the changes they experience in their use of different levels of care. Such information would provide data needed in planning how to reduce the financial impact of long-term care and make the financing of insurance coverage for institutional care, in particular, feasible. Data on Risk Factors for the Elderly Variables identified as risk factors for disease among middle- aged persons, i.e., for the development of stroke, cancer, and heart disease, often differ for persons age 65 and over. Many questions are being raised about how applicable the knowledge developed from studies of persons in their young and middle years is for the elderly. Risk factors at advanced ages have not been adequately studied. The importance of these questions increases as the elderly become more numerous and the proportion at the most advanced ages grows. Better data are needed to identify the specific geriatric syndromes that predispose the elderly to frailty and to establish the risk factors for particular diseases, as well as the risk for developing functional impairment. Better Data on Special Subgroups of the Elderly The elderly population that is age 85 and over, the oldest-old, is growing very rapidly both in absolute numbers and as a percentage of the population over age 65; it is expected to grow even more rapidly in the next several decades. Whereas between 1950 and 1980 the population age 65 and over more than doubled, the subpopulation age 85 and over quadrupled in that period, increasing in number from 577,000 in 1950 to 2.2 million in 1980. Inadequate provision has been made for developing information through national surveys for various subgroups of the elderly population.

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 35 Blacks, Hispanics, and the poor and near-poor are likely to dif- fer from the remainder of the elderly population in health status and other characteristics. These subgroups should be carefully mon- itored, and oversampled in surveys if need be, for their health status and health-related service utilization because of their particular vul- nerability and potential difficulties in obtaining access to care. The veteran population, mainly men, is of interest because of the current rapid increase in the number who are elderly. In 1980 there were 3 million veterans age 65 and over, constituting 10.5 percent of the veteran population. It is estimated that by 1990 there will be 7.2 million veterans age 65 and over, constituting 26.6 percent of the veteran population. By 2000, there will be some 9 million veterans age 65 and over, and by 2020 the veteran population age 65 and over will grow to about 45 percent of all veterans. (Veterans Administration, 1984:4~. In 1980, 27 percent of all American mates age 65 and over were veterans, while in 2000, some 63 percent of American males of that age group will be veterans. In the veteran population, as in the general population, espe- cially large increases are occurring among those age 75-84 and 85 and older. These subgroups consume the greatest amount of healths resources on a per capita basis. Accordingly, the Veterans Admi~- istration will face the problems of increased health needs and costs. The data needs identified in this report for the elderly population are relevant for the Veterans Administration, modified by special conditions such as benefit entitlement provisions and availability of health care resources. Characteristics of Future Cohorts of Elderly Public policy makers need disaggregated projections of the char- acteristics of future cohorts of elderly persons- their size; age, sex, race, and ethnic composition; marital status; morbidity and mortal- ity rates; educational and economic status; labor force participation; and housing and living arrangements. Successive cohorts of elderly will differ from each other in these and other respects. Planning for the types and modes of delivery of health care and related services to be provided, how they wiD be paid and financed, and the numbers of institutional and professional health care providers required to meet the needs of the elderly in future years are heavily dependent on accurate projections of the characteristics of future generations of elderly.

36 AGING POPULATION IN THE TWENTY-FIRST CENTURY Important Data Bases Outside the Scope of the Report This report addresses data requirements that can be met by federal agencies responsible for producing policy-relevant information on the state of health of the elderly population, their use of health care resources for acute and long-term conditions, and government and private expenditures for their health care. The information may come from administrative records, statistical surveys, or the decennial census. The strength and significance of the information systems devel- oped by federal agencies derive largely from the fact that they are national in coverage and produce data on a continuous or periodic basis for demographic, social, and economic subgroups of the pop- ulation. These systems have been and will continue to be the main source of data needed to understand the current status and trends in health affairs. The recent increased emphasis on longitudinal data has added a new dimension to their potential for answering questions about our aging society's health needs. Although these systems are an essential source of data for health policy makers, public and private planners, and administrators, there are limits to the scope and detail of the information on the elderly such systems can be expected to provide. Our comprehension of many aspects of health among the elderly is extended beyond what can be gained from multipurpose national data systems by research projects funded by grants or contracts. The panel is cognizant of the breadth of research on the elderly funded by the National Institute on Aging (NTA), the National Insti- tute of Mental Health (NIMH), the National Center for Health Ser- vices Research and Health Care Technology Assessment (NCHSR), and other government agencies and foundations. The symbiotic re- lationship that exists between these sources and the national data systems has paid dividends and needs to be preserved. In fact, many of the recommendations from the panel regarding data gaps, the need for longitudinal studies, and improved methodologies are relevant for both types of data sources, despite their differences. National systems can identify major characteristics of health- related problems and progress in dealing with them. Research projects can probe more intensively into selected issues and influence the national systems. Research is often directed at specific health conditions, their incidence and prevalence, associated functional im- pairment, and factors affecting amelioration or deterioration. Eti- ological questions and some issues of transition may best be dealt

INTRODUCTION AND SUMMARY OF RECOMMENDATIONS 37 with through research projects that utilize local cohorts monitored carefully at frequent intervals. Many investigations are theoretically based and are designed to test hypotheses; others break open new fields of inquiry through advances in methodology and instrument development. A detailed discussion of specific projects is beyond the scope of this report, but a few examples illustrate their significance. The Epidemiologic Catchment Area (ECA) research program, conducted cooperatively between NIMH and academically based investigators in five local areas of the country, is the source for the most extensive data on the epidemiology of mental health disorders and related health services available in the United States. Oversampling of the elderly in the surveys carried out has made it possible to give them major attention. Prospects for periodic reexamination of mental disorders are increased through transfer of portions of the Diagnostic Interview Schedule in the ECA to national surveys. The surveys called Established Populations for Epidemiologic Studies of the Elderly (EPESE) are sponsored by NTA under contract in four areas. New knowledge is being developed concerning medical and social factors in health conditions among the elderly (e.g., pain, sleep, hearing, vision) through annual, intensive surveys of panels of subjects. The Durham Older Americans Resources and Services (OARS) Community Survey at Duke University has conducted To- cally based longitudinal studies of older people and has adapted and tested Instruments for measuring their functional status. To these can be added other highly targeted studies supported by agencies inside and outside government. An important develop- ment in more recent years to increase the utility of data collection in research projects has been the establishment of archives of infor- mation for secondary data analysis by scholars here and abroad. An example of this is the National Archive of Computerized Data on Ag- ing, currently sponsored by NIA. The availability of public-use data tapes from national information systems is equally important and, as noted previously, is a subject that requires increased attention. The pane! also recognizes that cross-national and international studies that deal with the elderly can be a valuable source of infor- motion. Examination of the characteristics of the elderly in countries with different risks for disease, different lifestyles, and different health care systems contribute to a better understanding of factors that af- fect health status, utilization, and expenditures for health care of the elderly in the United States. Review of such studies and surveys

38 AGING POPULATION IN THE TWENTY-FIRST CENTURY was beyond the scope of the panel's study; however, this subject is receiving attention. The House Select Committee on Aging con- ducted a Workshop on Cross-National Data on Aging in October 1985, at which presentations were made by researchers and repre- sentatives of government and international organizations conducting studies or surveys of the elderly. The Census Bureau is developing and automating an international data base to permit comparative analyses of demographic social and economic characteristics of many countries, based on their population censuses and surveys. With funds from NIA, the Bureau has been compiling age-specific data on socioeconomic status and mortality among the elderly in a large number of developed and developing countries. The first publication from this project, An Aging Wo rid (Torrey et al., 1987), provides an overview of trends and a guide to the use of this data base. Valuable as the three types of data sources discussed are for policy makers, the proving ground for determining the effectiveness of alternative proposals to meet the health care needs of the elderly is demonstration programs with evaluation components. The Health Care Financing Administration has funded projects that test the cost-effectiveness of community-based case management approaches for altering dependence on nursing homes among the elderly and has introduced waivers of restriction in Medicare and Medicaid benefits to assess the value of coordinated long-term care through capitation reimbursement. The Robert Wood Johnson Foundation has provided funds for demonstration programs to address the needs of the health- impaired elderly and, in collaboration with government agencies, the foundation is initiating a project aimed at improving linkages to health and human services, including suitable housing arrangements, for the chronically mentally ill, a significant number of whom are elderly. The Commonwealth Fund has established a Commission on Elderly People Living Alone to assess the health and social problems they face and identify demonstration programs that seek out effective ways to relieve them. Their impact on policy and the diffusion of results are essential features of these efforts, yet national data systems remain the primary sources of information on the changes that occur among the target groups in the population generally.

~ - - r, o ra x £ o o o ._ C. - x - o ~Q ,< ct bO ._ - o v a, ._ ec o o ._ C. 4, o c) o. - - P~ - m E~ U) ._ bO ¢ - o 0O a, . _ U' ~: C~ - C~  :: o 4) O ~ ~ O ~ V :> CQ z V, ~_ O Q o ~ l_, 1 a, O m v CQ co x x x _~< ~oo X :t ¢ V 0 z X _~= ~oz X :t ~ Z Z X X P~ co x v ~X .. V) V :: V z ~X X X X X .- _ O ~ ~c9 ~, ~3, ~e ~ ~c ~ z ~ u E~ E ' ,_~ ~ Z u, .= o ~ ~ O .cn ~O _ ~ ~ ~ ~ ~.O ~ o ~ ~t ~(D ·= V CO ~ o ~CO CO CO CO CO CO 39

DO G) . - ~o - a, o oO cq :t - o ~ - c o v - - m :> CQ z CO - O ~ ~0 ~ O ~ 4= ~ O V o m v co co _~< ~ 0 0 CQ co~ V 0z L. _ ~ ~  Z ._ z z z U) v :~: .. V) V =: V V Z . . o ._ ~:5 o q} L. v x x x x x x X X x x X X X X x X X X X x x x x x X X x x  X X X X X X ~ >,4 XX O ~ ~ - ~ ~ 0 0 ~.°, ~ > E ~ :^ E ~ ~ -~ ° . Z Z . · v ~, o <> oo ~_ ~ o _ CO o: oo CO ~ V 40

x x x x x x x x p" pa x xxx x me x o xx x ~x x , c =~ ~ C , ~ ~ - 1 ~ ~ ~ ' ~ G D . z ~* ~ .= or ~ ~ ~ 41

~q .~ c) ~a - c~ s~ s" o oD .- a, o E ~:5 ._ o v m E~ U) z CQ ~_ ~ 8 ~ o V o :: o C) C~ a-O~ S" a, o  ~ a' m ~ U) co P: 0 CO t: V z = z ~4 z ¢ z P" v .. CQ V ~ ~ V V z . . o ._ ~¢ o CJ I. ·- a ,, ~ 0 V co P~ x x P4 X X P. ~X x ~X X x  x E ~ ~· e ~ i 5 Z c ~O 0> CD CD CD 42

x x x x x x x x x x x x xpa ~x x x ·~- 9, ~ ~ ~. °it,, ~ 9 ~ ~ . ~ ~_ ~ ~Go a, ~ ... . . .. ~ ~rim ~ 43

oO . - bO - c. o . - o v m ~: :> U) z co - ~ O o S" a, ~ ~ ~0 O ~ ~ O ~ V V, ¢ o m v  co cn 00 .t co $ C~ C~ $ o 4, $ - z e, $ _ ~ ~ ~ 0 0 CO $ V z $ Z $ z - z CO v $ .. CO V $ V z . . o ._ o v 4} ~ Cd _ ~ . - ~, 0- l, ~ O V U] P~  X P4 ~X x ~P" X cn V P~ - _ ~ - ~. ~W ~ ~ , z z _ 0 ~C ~CV~ ~ ~00 00 00 ~00 44

x x x x x ~x x x x ~ x Hi ~' o Go, O ~ ,~ ~ ~ ~ ~ ' ~ _ . · ~ of ~co CD ~ _ ~_4 _/ ~03 oo oo DoGo Go Go Go Go 45

¢ - A L. o . - o v m CQ a In I_ 4 ~0 C: ~0 ~ O ~ ~ O ~ V o m V _~< ~oo ~V) U) ~ V 0 z .t ~n ~: ~: o ~V C~ a L. _~=  Z ~_ z ~4 z P~ v cn v z . . o ._ o $.e - ~ O V U) X X ~X X ~X x X X ~X X P4 X x x x X X X X X X X f ~ ~ ~ ~ e ~ ~ ~ ~ ~ ~ ~ - ~ ~ ~ _ ~ u 0' ~CO C;, 46

x pa pi x pa ~x ~x ~x x x pa x pi x x pi ~ x ~x ~x pi ~ x ~x pa ~x pi ~x ~pa x ~x x x ~x ~x .- ~ ~ c ' ~ ~ ~° C e ~2 ~ ~ E ~ C - O ~ ~i, ~ ~ ~ ~ - = ~ o ~ ~, - .CQ ~_ en co ~CD ~a) a ~O O O O O O ._ cry cn ~ O ~~ ~ 47

.a) be ¢ s" a o o' G. ~ - q) u) ~: - o ~ - o v m ¢ E~ :> CQ U) ~_ a ~0 O ~ ~ O ,av CO o. P~ V CO U) _ 0 CO := V z S" _~=  z _ Z :: - z - z o~ v .. co V :t av V Z . . o ._ a, ... - ~ O V eQ x X X X x x x x x x X X x 00 ~47 . _ . =,,= ~ c,~,~ E ' ~,, ~ ' ~ ~ o oo .... . . oooo o o ~- 4 ~ 48

x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x o ~ · ~- ~ 0 cn ~ ~ c. O ~. . . ~ . i 1 , 5. 04 ~.o cO O ~ ~_ _ _ _ _ _ cq ~_ _ ~_ - 4 _ 49

oO v bO - a, o v u) s o ~ - o v - m E" U) z CO ~ O a-O~ V ~o a' . ~ O V o z L' o m v ~n CO ~ 0 0 CO v z - - z ~: z - P4  U) v co z . . o ._ C~ o C) 4) _ t.. - ~c V CO X ~X x x x X X P ~X X X X ~X X X X x x a~ ~ng) ~ ~.= X ~ o ~ ~ ·O o~o '0~ ° 'o ° m- ~ _=~ ~o ~ ~ ~ b0 ~ tD4I, a.~ a, 0 {0 c 0 ~E ~ O ~D. ~ ~ ~E ¢~ c : ,~! ce ~ ~ ~ ~ ~ ~ ~ ~ O ~0 0 0 ~ :^ ~ 0 ~ 0 0 ~0 Pe ~ ~ ~ ~ ~ ~ o 0 ~ c ~co .. . . . . 50

q, v) .- -,4 ~ :~: o ~o o ~ L. ~ - - ~ . - - ~ o p: ¢ o co o ·> S ~C al . - ~ - ~ ¢ L: * ~ - ~ oq ~ .o " u) - . ~u) u) o ~ - ~- - ~ ~s ' ~ ~ ' ~ ~ ~:, ~ ~ E _ _ _ _ _ _ _ _ _ _ _ ~, 4~ o~ ~ ~ C~ ~ ~ ~ ~ ~ ~ ~ o o o o o o o o o o o ~ ~ ~o c,, ~ ~ ~ ~ aV ~ ~ ~ ~ ~ ~ ~ ._ ~ C: . - o ~ ~ C~ ~ ~ ~ ~ ~ ~ ~ o ~ ~ Z zzzzzzzzzou'=> C ~- 4 { ~V] o ~ ¢ ¢ o ~Z Z Z Z C ._ ~ oo o oq P~ ~C -S =3 ~ t. ~_ o C o b. r~ 03 ~' ._ _ O ° O V ._ - C~  ~ =.o oq .. q) aa~ |-, a O ~ ~ O O 4, .~ .~ q, O O V V V CO O ~ V CQ ~S CO ~ ~ ~ Z ~ ¢ ~ ~ ¢ ZZZZZO00~;> E C u: ~ ~ , , C C, ~ ~ ~ ~ ~ o o ~ C~ (U ~ ~ ·_ ._ ._ _ ~ ~>ZZZ~ co cn v ~ =: ~ a <,, ~ ~V=¢m V V V :¢ _ ~ ~ z U) co V V z z 51