U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Panel on Statistics for an Aging Population; Gilford DM, editor. The Aging Population in the Twenty-First Century: Statistics for Health Policy. Washington (DC): National Academies Press (US); 1988.

Cover of The Aging Population in the Twenty-First Century

The Aging Population in the Twenty-First Century: Statistics for Health Policy.

Show details

8The Financing of Health Care Services for the Elderly


Spending on health care services for the elderly has been increasing since 1965; between 1977 and 1984 it increased at an annual rate of 14.5 percent (Waldo and Lazenby, 1984). The increase in expenditures is reflected in the increasing cost to the federal and state governments of operating the Medicare and Medicaid programs as well as in the increase in out-of-pocket payments made by the elderly. As a consequence, health care financing has become one of the more critical policy issues to be addressed by the nation.

The methods used to finance health care services have important effects on the use of health care services by the elderly and their level of health and well being, as well as on the growth and development of the health care sector itself. In addition, the methods used to finance health care services influence the distribution of income between the sick and the well, the old and the young, the general taxpayer and the recipients of care.

Many types of data are required to provide a factual basis for policy issues related to the financing of health care services for an aging population. Although a policy issue is seldom resolved by using only data from a national information system, such data are frequently used to address some elements of a policy issue. For example, in the list of policy issues that follows, the first would probably require an evaluation study, but Medicare records could provide historical information on the cost of the fee-for-service systems and the HMOs.

To address the second issue, data from the National Medical Expenditure Survey could be used to identify the most costly of the various chronic illnesses and help to limit the illnesses to be addressed by the policy analyst. Selected emerging policy questions that need to be addressed follow:

  • Will health maintenance organizations, social health maintenance organizations (SHMOs), and preferred provider organizations (PPOs) serve as less costly alternatives to the current fee-for-service system?
  • How can we control expenditures for health and long-term care in the face of the projected growth of the elderly who are at risk of chronic illness often requiring extensive medical and long-term care services? Will people have to do without medical care?
  • What aspects of the growing health needs of an aging society are most affected by the pressures to constrain budgetary and economic resources devoted to health care?
  • What mechanisms are needed to share the burden of health care expenditures for an aging society more equitably among all members of society?
  • What is the impact of the Medicare prospective payment system based on diagnosis-related groups on providers, patients, expenditures, and access to and quality of care?
  • With the growing financial burden of out-of-pocket expenditures for certain groups of the elderly who are disabled and require extensive treatment, what are the alternative equitable financing mechanisms to pay for these services?
  • What alternative financing mechanisms for the supply of long-term care services should be supported and by whom?
  • What changes in funding mechanisms, legislation, and public policy are necessary to shift the emphasis away from hospital and nursing home care toward less costly alternatives?
  • Can less costly and less restrictive alternative services to institutionalization be developed to maintain the independence of the elderly? Can economic incentives be developed for in-home and community-based services to maintain the elderly at home?
  • What is the viability of tax credits, reverse mortgages, independent retirement accounts, and the like in assisting the elderly to pay for a greater share of their health and long-term care costs? Are there other private-sector alternatives?
  • What are the costs (direct and indirect) of chronic illnesses such as Alzheimer's disease and what are the implications of helping to meet those costs through public funding?
  • To what extent is there an emerging intergenerational inequity with the aging of the baby boom generation—Americans born between 1946 and 1964—and low fertility rates in which a smaller number of the working population will bear the burden of support of the larger number of retired elderly beginning about the year 2010?

This section of the report focuses on specific aspects of the financing of health care services for the elderly population. First, it describes the major public programs that cover health care services for the aged as well as the other sources of funding for health care. Since many of these programs are undergoing significant changes, the direction of that change is outlined. Next it presents information on the level of expenditures and source of payment by type of service for the aged, including the distribution of expenditures across the aged population. Data needs and recommendations are presented with reference to the policy issues emerging from the current proposed changes in financing mechanisms.

Sources of Funding

In describing the current methods of financing care for the elderly, it is appropriate to begin with a brief description of the public programs, in particular Medicare and Medicaid. These are the most important sources of funding, and the structure of these program affects the nature of the insurance policies that are offered by the private sector.


Approximately 95 percent of all people age 65 and over in the United States are covered by the Medicare program. Medicare consists of two separate but complementary programs: Hospital Insurance (HI) for services furnished in hospitals, in skilled nursing facilities, and by home health agencies; and Supplementary Medical Insurance (SMI) for the services of physicians, home health visits (for people who may not be covered by HI), outpatient services, and the costs of durable medical equipment and prostheses. Coverage for outpatient mental health services is very restricted. In addition, some services frequently used by the elderly, such as outpatient drugs, dental services, and eyeglasses, are not covered.

The Medicare program was explicitly designed to cover the major costs associated with episodes of acute illness or the acute manifestations of chronic illness. The extent to which the program actually pays for services such as skilled nursing facilities and home health, usually considered long-term care services is therefore very limited.

People are automatically enrolled in HI within a short time of reaching their sixty-fifth birthday. The costs of HI are covered by Social Security payroll taxes paid into a trust fund by employers and employees. Enrollment in SMI is contingent on paying a premium that covers about 25 percent of the cost. There is some beneficiary cost sharing on Medicare-covered services. The cost-sharing provisions were included in the original legislation both to control federal budget expenditures as well as to deter unnecessary utilization. The actual cost-sharing provisions are complicated and vary from service to service.

When Medicare was first enacted, the Congress incorporated into the program many of the standard features of the Blue Cross/Blue Shield plans that were then the dominant form of private health insurance. The most important features were beneficiary freedom of choice of providers, cost-based reimbursement of institutional providers, and fee-for-service reimbursements based on reasonable charges for physicians' services. With the exception of long-term care services, there were few limitations placed on the use of covered services. There were no incentives for hospitals to control costs and no rewards for improving efficiency. The reimbursement provisions are currently undergoing major changes. In 1982, changes were made in the law to stimulate the enrollment of Medicare recipients into prepaid group practices; in 1983 the cost-based reimbursement system for hospitals was replaced with a prospective payment system under which hospitals are paid a fixed amount for taking care of patients based on their discharge diagnosis. In 1985 the Reagan administration proposed replacing the current physician reimbursement system with a fee schedule. The federal government is also evaluating ways for bringing other providers of care under prospective payment (Davis and Rowland, 1986:Chapter 3).

Between 1977 and 1984, payments for services under Medicare increased at an annual rate of 17.3 percent per year. By 1984 Medicare expenditures for the elderly amounted to $58.5 billion, making it the third largest federal program (Waldo and Lazenby, 1984). As a result of the rapid escalation in the cost of the program, the reimbursement changes outlined above were made. In addition to those changes, the administration and Congress are seeking other ways of lowering the costs of the program, such as increasing the cost-sharing requirements, changing the age for eligibility, increasing the premium costs, making Medicare the second payer for those who are employed and eligible for employer-provided health insurance, increasing Medicare enrollment in HMOs, and changing the Medicare program to a voucher program. (This description of the Medicare program and the following discussion of Medicaid are based heavily on Sawyer et al., 1983.)


Medicaid, a program to provide medical services to the poor, is administered by the states under federal guidelines. With respect to the elderly, Medicaid pays for the medical care for those who meet Supplementary Security Income (SSI) standards. States have the option of covering medically needy individuals (those with incomes slightly above the SSI levels) and individuals who have incurred sufficiently high medical expenditures that they "spend down" to Medicaid income eligibility levels.

For dually eligible recipients, Medicare is the first payer for Medicare-covered services and Medicaid is the second payer; that is, Medicaid will pay the cost-sharing amounts that would normally fall to the patient. Depending on the state's program, Medicaid may also pay for the cost of services not covered by Medicare, such as outpatient, drug, optical, and dental services. More important, as Medicare fulfilled its purpose of covering most of the costs related to acute episodes, Medicaid has evolved into the primary public funding source for long-term services for the elderly.

Like the Medicare program, the costs of the Medicaid program have been increasing at a rapid rate. Between 1977 and 1984, Medicaid expenditures on behalf of the elderly increased at an annual rate of 14.2 percent—largely due to coverage of nursing home care. In 1984, expenditures for the population age 65 and over were approximately $12.8 billion (Waldo and Lazenby, 1984). As with Medicare, there is a concerted effort to reevaluate the structure of the program, in particular with respect to long-term services, to revise the methods used to reimburse providers, especially nursing homes, and to change the income eligibility levels.

The Veterans Administration

Until recently, all veterans age 65 years and over were eligible for Veterans Administration (VA) services. A recent law made several changes in a veteran's eligibility for VA health care regardless of age, which became effective July 1, 1986. Actual access to health care services is now determined through an eligibility assessment, with first priority being given to veterans with service-connected disabilities or to those who meet a means test known as the category A means test. These veterans are eligible for hospital care in VA facilities. Outpatient and nursing home care may be provided in VA facilities if space is available. Veterans with nonservice-connected disabilities and with income above the category A means test but below a higher level, known as the category B means test, may receive hospital, outpatient, and nursing home care in VA facilities if space is available. Veterans with nonservice-connected disabilities and with income above the category B means test may receive VA services if space is available, but they must agree to pay a deductible amount for care equivalent to the payment that would be required by Medicare. In addition to the VA-owned and operated facilities, which include hospitals and their associated outpatient departments and nursing homes, the Veterans Administration contracts for skilled and intermediate care at community nursing homes on a per diem basis. However, with the exception of care provided to veterans with service-connected disabilities, the VA will not pay for more than 6 months of care in community facilities. In 1984, $3.3 billion was spent for health care of aged veterans.

The actual use of the VA system by elderly veterans is contingent on the characteristics of the elderly that define eligibility, the availability of VA-owned facilities, and the access to the private sector by VA eligibles. Since most veterans are covered by Medicare, the service that is likely to be most attractive to the older veteran is nursing home care. However, the main factor that will influence the veteran's use of nursing home services is the number of nursing home beds available within the VA system. Before July 1986, approximately 10 to 12 percent of elderly veterans actually used the system (U.S. Congressional Budget Office, 1984). This number may change as a result of the current and proposed changes in Medicare and the new eligibility procedures for VA health care.

Other Government Programs

There are a number of other programs that have been developed to finance health care expenditures for the elderly. Included are the Department of Defense Civilian Health and Medical Program of the Uniformed Services, which provides care for active and retired military forces and their dependents, state and local government hospitals providing community and psychiatric hospital services to older citizens, federal grant programs, and state and local public assistance programs providing services that are not eligible for federal matching funds under the Medicaid program. In 1984, a total of $3.4 billion was spent for the elderly by government programs other than Medicare, Medicaid, and the VA.

Private Insurance Programs

There are a number of private insurance policies that have been specifically designed for the elderly. In general, most of these policies fill the gaps in Medicare-covered services, such as the coinsurance and deductible provisions, rather than extending insurance protection against uncovered services such as long-term care services or outpatient drugs (Rice and McCall, 1985). With the projected large increase in the number of people age 65 and over, especially those age 85 and over, there is increased interest in finding mechanisms for stimulating the development of private insurance coverage for long-term care services (Meiners, 1985a).

Of the noninstitutionalized elderly, approximately 65 percent have private policies that supplement Medicare, 10 percent are covered by Medicaid, 20 percent are covered by Medicare only, and 4 percent had some other form of coverage (Garfinkel and Corder, 1984). Among the elderly not eligible for Medicaid, those who are better educated, have higher incomes, and are in slightly better health are more likely to purchase a supplemental health insurance policy.

Expenditures on Health Care

In 1984, personal health care expenditures for the elderly amounted to $119.9 billion, or $4,202 per person age 65 and over (Waldo and Lazenby, 1984). Of this total, 25.2 percent was paid by the consumer directly, 7.2 percent through private insurance, 48.8 percent by Medicare, 12.8 percent by Medicaid, and 5.6 percent by other government programs, primarily the Veterans Administration. However, these averages mask the wide differences in the funding sources for different types of medical services. The extremes are represented by sources of expenditures on hospitals and nursing homes. In 1984, hospital expenditures for the elderly amounted to $54.9 billion ($1,900 per capita) of which 3.1 percent was paid by the patient directly, 7.9 percent through private insurance, 74.8 percent by Medicare, 4.8 percent by Medicaid, and 9.1 percent by other government programs. Nursing home expenditures in the same year amounted to $25.1 billion ($880 per capita), of which 50.1 percent was paid for by the patient directly, 1.1 percent by private insurance, 2.1 percent by Medicare, 41.5 percent by Medicaid, and 4.4 percent by other government programs (Waldo and Lazenby, 1984).

Since the elderly population is heterogeneous, average expenditures provide an incomplete picture of the cost of illness and the sources of funding. For example, consider the 1980 National Medical Care Utilization and Expenditure Survey expenditures data for three groups of the noninstitutionalized elderly: the "low-cost" users, those with expenditures less than $500, "medium-cost" users, with expenditures between $500 and $3,000, and the "high-cost" users, those with expenditures over $3,000. Although these categories are somewhat arbitrary, they help to establish the well-known fact that health care expenditures are concentrated on a small number of sick people. As shown in Table 8.1, only 14 percent of the noninstitutionalized elderly had health care expenditures over $3,000, but they accounted for 75 percent of total expenditure of health care services made on behalf of the elderly.

TABLE 8.1. Health Care Expenditures for the Noninstitutionalized Elderly, 1980 (percentage).


Health Care Expenditures for the Noninstitutionalized Elderly, 1980 (percentage).

Five percent of the elderly in the survey were institutionalized or died during 1980. Expenditures for this group emphasize even more strikingly the concentration of dollar costs for the seriously ill. During the part of the year that these persons were in the community, they accounted for 22 percent of the total health care expenditures for the elderly.

As noted earlier, the share of expenditures paid directly by the consumer also varied by the nature of the health care received. Elderly persons who were not hospitalized paid, on average, 67 percent of their medical care charges, while the elderly who had been hospitalized paid 18 percent. Because the total hospital charges are high, this out-of-pocket charge for the hospitalized group was about $650, while the nonhospitalized paid $202 (Kovar, 1986). It should be stressed, moreover, that these numbers conceal differences in out-of-pocket liability for the institutionalized population. People in nursing homes, who incur large out-of-pocket expenditures—because both the cost of the service is so high and insurance is so limited—are not included in the National Medical Care Utilization and Expenditure Survey, the source of the above data.

The Frail Elderly

One group of elderly that is receiving considerable attention are those with potential needs for long-term care services—the frail and dependent elderly. In 1982 approximately 19 percent of the elderly age 65 and over (4.6 million people) needed help in activities of daily living or in instrumental activities of daily living, and the size of this group will increase with the aging of the population (Liu et al., 1985). We actually know very little about the total cost of care for this population: they represent such a small part of the general population that they have not been adequately represented in national household surveys (such as the 1977 National Medical Care Expenditure Survey and the 1980 National Medical Care Utilization and Expenditure Survey), the major sources of information on the distribution of health care services across the aged population.

This group is of particular interest because there is currently limited public and private insurance to help them meet the cost of long-term care services and because it is believed that fostering home-based care may offer a cost-effective alternative to nursing home care. Thus, one major health financing issue is the extent to which public insurance programs should support home-based programs.

This is a very complicated issue because most of the home care services received by this group of people are provided by family members or by members of the helping organizations. In fact, approximately 72 percent of the services rendered are provided by family members (Liu et al., 1985). Although the use of formal (i.e., paid for) services increases with the level of disability, 65 percent of people with many limitations in activities of daily living relied solely on informal care (Liu et al., 1985). The current extensive use of unpaid health care makes it very difficult to design a public financing program for health care services that will encourage the efficient substitution of home-based services for nursing home services, because any such program is also likely to lead to a substitution of services provided by the family or already paid for by the individuals or their families for publicly funded home care services.

Thus in evaluating proposals to change the financing of long-term care services for the elderly, data are needed to assess the extent that financing mechanisms will lead to an increase in the use of services or to a substitution of publicly funded services for services previously provided by the family free of charge or paid by the individuals and their families. If new services are used, to what extent will they act as complements or as substitutes for nursing home care? In addition, what is the effect of the use of new services on the health status of the elderly, to what extent do these services lead to an increase in their quality of life, and what is the effect of this new financing on the proportion of the family's income that is spent on health care?

Future Changes

All aspects of health care financing programs are currently being reevaluated: program eligibility, in particular the extent to which income and assets should determine access to public programs; appropriate roles of public versus private sources of funding; the scope of the covered services, appropriate levels of beneficiary cost sharing; the level and basis of payments to providers; the level of quality of care that should be guaranteed by public funds; and the types of services to be promoted by public funds.

As the organization, delivery, and financing of medical care services are changed, we need data that will allow us to address the following questions: How do the changes affect the cost of care, the distribution of the burden of paying for the care across public and private sources of funding and among individuals, and the health status and quality of life of the elderly?

Data Requirements

In order to develop and guide policy for financing health care for the elderly and to monitor the impact of changes in such financing, it is necessary to have a variety of data, including person-based survey data, administrative record data, and actuarial data. Survey data should include information on health status, income, assets, medical care expenditures, and sources of payment for medical care. Data need to be sufficiently detailed so that reliable statistics can be created for the most vulnerable of the elderly—the frail elderly, low-income elderly, and minorities. This information should be collected on both people residing in institutions and people living in the community.

Additional data that are useful in the analysis of financing medical care services are the administrative records of the Health Care Financing Administration, the Social Security Administration and the Internal Revenue Service. These data are by-products of administering large programs, and their statistical systems are reasonably inexpensive. There are two major constraints impeding the usefulness of these data for policy purposes. The first is that the agencies do not have adequate staffs to analyze the data, a fact that results in significant lags in the production of data and analyses. This problem is likely to become more acute over time. The second is that there are many restrictions imposed on making public use of data tapes, a subject that is discussed in more detail in Chapter 11.

Actuarial data from private insurers are also needed for establishing the costs of alternative long-term care policies. Minimal data are available on costs and utilization of long-term care insurance and very little of the data reflects actual experience. A further complicating factor is that private insurers are reluctant to share their limited data bases to enable the development and marketing of long-term care insurance. Although the number of companies providing long-term care insurance is growing, there is also active discussion of federal coverage of long-term care costs for the population not covered by Medicaid.

Timely Data on Expenditures for the Elderly

The Health Care Financing Administration periodically publishes data on health care expenditures of the population age 65 and over by type of expenditure and source of funds. These data are estimated from secondary sources, including administrative records of the Medicare and Medicaid programs and surveys conducted by NCHS, HCFA, and NCHSR. Health expenditures of the elderly population are part of the national health expenditure accounts for the entire population published annually by HCFA.

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, of the total health care costs of the elderly population. They also serve as a basis for assessing the possible consequences of changes in public policy and programs, although this is always a question of the relevance of current and past data to a possible future program that changes the eligibility rules of current or past programs. For example, current cost and utilization data reflect the effects of the current mix of private and public insurance for the elderly. Changes in policy regarding long-term care financing would alter the situation, and it is important in the development of data to consider requirements to provide reasonable estimates of these effects.

Because of the policy uses of these aggregate expenditures data for the elderly population by source of funds and type of expenditures, their publication on a regular basis is important.

Recommendation 8.1: The panel recommends that the Health Care Financing Administration develop and publish timely annual estimates of the national health expenditures by age. These estimates should follow the publication of the estimates of expenditures for the total population by a few months at most. At a minimum, these health expenditures for the population age 65 and over should include estimates by type of expenditure and source of funds and by age (ages 65–74, ages 75–84, and age 85 and over).

National Medical Expenditure Survey

The 1987 National Medical Expenditure Survey follows a series of national medical expenditures surveys, including the 1980 National Medical Care Utilization and Expenditures Survey and the 1977 National Medical Care Expenditures Survey.

Like the NMCES and the NMCUES, the NMES surveys a national probability sample of the civilian noninstitutionalized population. The NMES Household Survey is a year-long panel collecting measures of health status, use of health care services, expenditures and sources of payment, insurance coverage, employment, income and assets, and demographic information. When information planned for collection is the same as that collected for the NMCES and the NMCUES, and the previously used questions were found satisfactory, the same wording was retained. A particular focus is community-based long-term care. Household data are supplemented by surveys of medical and health insurance providers and by data from Medicare administrative files. In planning for future surveys, the panel urges the National Center for Health Services Research and Health Care Technology to continue this policy of providing comparability of data items between surveys.

A major gap in the NMCES and the NMCUES is the lack of data for the institutionalized population. Since at least 85 percent of the institutionalized population is age 65 and over, the omission of expenditures for this population group presented serious gaps in accurately estimating the health care costs of the elderly population. An important feature of the NMES is an Institutional Population Component (IPC), which will survey about 10,000 persons in nursing homes, facilities for the mentally retarded, and psychiatric hospitals and collect data similar to those for the noninstitutionalized household population. The IPC universe includes all persons in these long-term care institutions for any part of 1987. IPC and household data will provide the first composite picture of the nation's use of long-term health care.

The institutional component will permit analysis of long-term care for the nation, including special attention to the increasing number in the age groups 75 and older. Current plans include the incorporation into the data base of three groups of providers of long-term institutional care: nursing homes, facilities for the mentally retarded, and psychiatric hospitals. Each of these types of facilities provides care to substantial numbers of federal beneficiaries, primarily Medicaid recipients.

With this institutional component, the analytic potential of the NMES will encompass national estimates of health services use, expenditures, and insurance coverage for: (1) the entire U.S. civilian population, including the institutionalized; (2) the entire long-term care population, whether residing in institutions or in the community; and (3) institutionalized groups of persons in nursing homes, facilities for the mentally retarded, and psychiatric hospitals.

The current sample design for the National Medical Expenditure Survey, as the panel understands it, will provide reliable estimates for two age groups, 65–74 and 75 and over. However, since the group age 85 and over is the most rapidly growing segment of the population and the group that makes the greatest use of health care services it will be critically important to collect information on their use of health care service.

Recommendation 8.2: The panel recommends that the sample of the elderly population for the household survey for future national medical expenditure surveys be large enough to provide accurate estimates to study utilization and expenditures for medical care for three elderly age groups: 65–74, 75–84, and 85 and over.

The National Medical Expenditure Survey will provide a rich source of data on the health services utilization and expenditure patterns of the community-residing and institutionalized elderly. Its panel design may permit assessment of the impact of changes in service delivery and payment systems experienced by some of its respondents during the course of a calendar year on their utilization of health services and expenditures for health care. In addition, the planned linkage between the survey data collected during this year-long panel and the Medicare records of those persons age 65 and over in the samples should enhance the value of the utilization and expenditure data collected. However, to capture trends in utilization and expenditure patterns in response to changes in delivery and payment systems over time, a national medical care expenditure survey should be conducted periodically. These data are used for estimating the cost of policy alternatives for health care—policy uses that require more recent data than those currently available. In addition, changes in health care expenditures and utilization stimulate policy questions that affect decisions about public financing and regulation of services.

Recommendation 8.3: The panel recommends that a national medical care expenditure survey be conducted periodically. The periodicity should be determined in relation to policy needs and the timing of other health-related surveys.

The NMES survey should be a joint effort of appropriate federal agencies (including the National Center for Health Services Research and Health Care Technology, the Health Care Financing Administration, and the National Center for Health Statistics) in order to take full advantage of the various types of relevant expertise in those agencies.

Follow-Up Studies of the Elderly in the NMES through Administrative Records

A major gap in understanding the impact of changes in financing of medical care services in 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.

A considerable amount of data about the sampled individuals is also available from administrative records of HCFA, SSA, and IRS. The National Death Index conducted by NCHS also provides information on an individual's death. This rich source of information from administrative records can enhance the usefulness of the survey data collected in the NMES for longitudinal analyses at minimum costs.

Recommendation 8.4: The panel recommends that the National Center for Health Services Research and Health Care Technology identify and follow the population age 55 and over in the 1987 National Medical Expenditure Survey through the linking of administrative records, including Medicare reimbursements from Health Care Financing Administration records, and, to the extent feasible, Medicaid reimbursements from state record systems. In addition, the National Death Index of the Center for Health Statistics should be used to identify the year and cause of death of each sampled person.

In carrying out these record linkages, it would be essential that NCHSR comply with confidentiality restrictions.

Outreach Program for the NMES and Timely Release of NMES Tapes

The analytic potential of the NMES encompasses national estimates of health services use, expenditures, and insurance coverage. These data must be shared in a timely fashion with the researchers in the academic and private sectors as well as researchers in other government agencies, many of whom are directly involved with policy makers.

Several federal agencies, including the National Center for Health Statistics and the Census Bureau, have been successful in assisting outside researchers in the efficient use of their public use data tapes by conducting conferences for data tape users throughout the country.

Recommendation 8.5: The panel recommends that the National Center for Health Services Research and the Health Care Financing Administration: (a) begin planning an outreach program, including the conduct of conferences for data tape users similar to those conducted by federal agencies such as the National Center for Health Statistics and the Census Bureau, to inform and educate the policy and research communities in the efficient use of the forthcoming 1987 NMES data tapes and (b) prepare a schedule for the timely release of the National Medical Expenditure Survey data tapes, prepare these public use data tapes as soon as feasible after the reference period, and make them available to the policy and research communities outside the National Center for Health Services Research.

Timely Data from the Medicare Statistical System

The Medicare Statistical System was designed to provide data to measure and evaluate the operation and effectiveness of the Medicare program. It has also been a major source of information for evaluating many policy questions relating to equity and efficiency of the Medicare program. For example, data on the distribution of Medicare reimbursements for survivors and decedents and by type of service provide useful information on the high use of medical care services in the last year of life. Medical reimbursements per capita by state and county are useful measures of equity. Provider certification data related to population are important measures of the supply of facilities and services and their variation across the country. Geographic variations in surgical procedures among the elderly are important indicators of practice patterns.

The statistical system is a by-product of three administrative record systems that are centrally maintained in the operation of the Medicare program: (1) the Health Insurance Master File, which contains a record of each person who is enrolled in Medicare, (2) the Provider of Service File, which contains information on every hospital, skilled nursing facility, home health agency, independent laboratory, and other institutional provider that has been certified to participate in the program, and (3) the Utilization File, which is based on the Medicare billing information. Since each record in the utilization file contains the beneficiary's claim number and the provider's number, the utilization records can be matched to the enrollment and provider records. This then provides the basis for developing population-based statistics or provider-based statistics.

In the past, HCFA has produced a variety of reports, including annual Medicare Program Statistics, Health Care Financing Review, Health Care Financing Notes, Health Care Financing Grants and Contracts Reports series, Medicare Reimbursements by State and County Facilities certified under the Medicare Program, and Enrollees under the program. The latest published data are the Annual Medicare Program Statistics for 1984.

Recommendation 8.6: The panel recommends that the Health Care Financing Administration devote more resources, including budget and staff, to the timely release, publication, and analysis of data from the Medicare Statistical System, including national and geographic data on enrollees, providers, and reimbursements.

Improved Access to Medicare Data

The volume of data potentially available from the MSS is large and especially useful for evaluating different aspects of the Medicare programs as noted above. The panel commends the Health Care Financing Administration for its efforts to develop useful files such as the Medicare Automated Retrieval System and the Medicare Provider Analysis and Review (MEDPAR). The MEDPAR Public Use File is a national sample of bills for short-stay hospital inpatient services for 20 percent of the Medicare beneficiaries selected according to predetermined digits of the health insurance claim number. The elements of the bill (SSA-1453) contained in the file are: age, sex, Medicare status code; length of stay, discharge status; total and Medicare-covered charges; principal diagnosis in ICD-9-CM code and DRG code. The file has been maintained annually since 1980.

Despite the obvious attractiveness of the Medicare files for analytic purposes, it must be noted that these files were established primarily to assist with administration and monitoring of the Medicare program. In order to make the Medicare administrative data more accessible and less costly for research use, a new file has been designed—the Medicare Automated Data Retrieval System. The MADRS is intended to reorganize and merge Medicare Part A and Part B claims files to shorten search time. Beginning with the 1982 data year, this file will contain all Medicare claims data and patient provider identifiers. The claims records in the Medicare files will be sorted first by year of service rendered, next by geographic region of residence of beneficiaries, and then by the health insurance number of the beneficiaries. It will be possible to create a longitudinal file for cohort analysis by combining data in the annual files (Office of Technology Assessment, 1985a, Appendix E:199; Lichtenstein et al., no date; National Research Council, 1986).

The MADRS file will enable researchers to identify groups of special interest and analyze them by age, sex, and/or admitting diagnosis, for example, and examine the care they have received over time. The development of the Medicare Automated Data Retrieval System is a positive step toward facilitating the analysis of Medicare data, thus gaining a better understanding of health services utilization trends among the elderly.

Recommendation 8.7: The panel recommends that the Health Care Financing Administration develop files designed 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.

Making data from administrative records available to researchers would be expected to result in information useful to both program agencies and policy makers. The Health Care Financing Administration should develop new approaches to improving access by nonfederal users, such as interns and postdoctoral fellows. More use should also be made of the Intergovernmental Personnel Act of 1970, which provides for agreements between federal agencies and state agencies for assignment or exchange of personnel for a specified period. Such exchanges are usually beneficial to both agencies.

Recommendation 8.8: The panel recommends that the Health Care Financing Administration complete the development of the Medicare Automated Data Retrieval System and maintain it on a current basis.

The Health Care Financing Administration Data System for Capitation

An alternative to fee-for-service reimbursement types of insurance is payment on a per capita basis without regard to the volume or type of service. HMOs have been the major systems that charge a fixed monthly fee (capitation fee) to cover all services except for small copayments. With the passage of the Tax Equity and Fiscal Responsibility Act (TEFRA) in 1982 and the issuance of regulations to implement the HMO provisions of TEFRA in January 1984, the number of HMOs participating in Medicare is expected to grow, as is the number of Medicare enrollees in HMOs. There are currently over 1 million Medicare beneficiaries in HMOs (i.e., prepayment for services). Historically, the Medicare Statistical System has provided summary data on beneficiary demographics and extensive data on the use and costs of Medicare benefits on both the beneficiary level and on the level of institutional providers (hospitals, skilled nursing facilities, home health agencies, hospital outpatient departments). The data have been used for program administration, monitoring, and evaluation. However, the use and cost data are by and large derived from claims for payment of service. Capitated payment systems such as HMOs are paid an overall capitation amount by HCFA so there is no transaction record to describe services rendered and payment made. As more beneficiaries leave the fee-for-service sector, the information gap on use of services will grow.

Data will be needed from HMOs to monitor the care received by beneficiaries of public programs and to gather information required for setting and evaluating capitation rates. The panel recognizes that, in some cases, HMOs may have to establish new data systems to obtain such information. HCFA will need data from HMOs to address a number of issues, differences in patterns of care by plan type, access to specialty services, and monitoring the appropriateness of payment formulas for HMOs. Some of these issues are common to the fee-for-service sector, but others, such as biased selection, are unique to capitation.

Recommendation 8.9: The panel recommends that the Health Care Financing Administration develop a data system for information on Medicare beneficiaries in capitated systems that is beneficiary-based, able to accommodate different types of capitated plans, reflect differences in services offered and in cost sharing, and utilizes uniform and consistent data definitions and formats among different types of plans.

This recommendation complements the more general Recommendation 9.1 to modify national health data systems to reflect changing patterns and sources of service delivery.

Disability-Medicare Linked File

The long-term effects of disability are an important component of functional limitations in the older population, medical care, utilization, and expenditures. The Social Security Administration, which administers the Disability Insurance program, maintains a Continuous Disability History Sample, a file stratified by state and including from 5 to 20 percent of each state's newly disabled individuals who have been determined to be eligible for future benefits awarded to the disabled population. The disabled are eligible for benefits under the Medicare program two years after the disability insurance award is made. In 1983, Medicare per capita expenditures for the disabled ($1,900) at ages under 65 (excluding persons covered under the End Stage Renal Disease program) were higher than for the elderly ($1,724). The Medicare disabled, as a group, accounted for about $5.5 billion of a total $57.4 billion, for the entire Medicare program (Health Care Financing Administration, 1985a). As the disabled population ages, they will constitute a significant subgroup of the elderly population requiring considerable medical care outlays. The Medicare experience of the disabled population under age 65 should be analyzed as a basis for forecasting their medical care utilization patterns and future Medicare outlays when they become 65 and older.

HCFA has developed a file detailing the Medicare experience for 1977–1981 for the cohort of persons becoming entitled to disability benefits in 1972. Utilization and expenditures for Medicare-covered services can thus be related to the diagnosis or type of disability that justified the disability award. Analyses of the linked file is now under way. When it is completed, it can serve as a baseline for a more current study, using the population entitled to disability benefits in 1980 linked to 1982–1986 data.

Recommendation 8.10: The panel recommends that studies of the Continuous Disability History Sample linked to Medicare files be fully supported jointly by the Health Care Financing Administration and the Social Security Administration and that a public use tape be prepared for this linked file with identifiers deleted as necessary to comply with confidentiality requirements.

Data for Policy Analysis of the Prospective Payment System

In 1983, HCFA introduced the prospective payment system (PPS) for reimbursing hospitals treating Medicare patients. Each discharged patient is classified into one of 468 diagnoses called diagnosis-related groups based on the information on the hospital bill. The hospital is paid the fixed predetermined amount for that DRG.

Implementation of the prospective payment system has resulted in shortening the average length of stay for Medicare patients. Studies are under way to determine whether patients discharged under PPS were not yet ready for self-care and, if so, where they obtained needed care. The HCFA hospital bill includes items for patients discharged to home under care of organized home health services, discharged to skilled nursing home or to intermediate nursing facility, in addition to the items on the Uniform Hospital Discharge Data Set (UHDDS): routine discharge, left against medical advice, discharged to another short-term hospital, discharged to a long-term care institution, died, and not stated.

The original Uniform Hospital Discharge Data Set was promulgated by the secretary of the Department of Health, Education, and Welfare in 1974. The additional detail for ''Disposition of Patient" has been made part of the Uniform Bill required for each hospital discharge by HCFA.

The UHDDS was reviewed without change in 1980 (U.S. Department of Health and Human Services, 1980c) and by the Health Information Policy Council in 1984 (Federal Register, July 31, 1985:31038-9). The council review served to clarify some categories and definitions but did not add more detailed categories to "Disposition of Patient." More detailed information, similar to that on the Medicare billing form, could be useful in studying length of hospital stay in conjunction with diagnostic information (including multiple diagnoses) and severity of illness. Diagnostic information is available in hospital bill reports, which provide for up to five diagnoses for each discharge, and in the National Hospital Discharge Survey, which provides for seven. Severity of illness information is not available currently for analysis.

Recommendation 8.11: The panel recommends that the National Committee on Vital and Health Statistics reconsider the Disposition of Patient items on the Uniform Hospital Discharge Data Set with reference to the changing data needs resulting from implementation of the prospective payment system.

Medicaid Data

The Medicaid program is unlike the Medicare program, in which data are available on the individual elderly or disabled and his or her use of medical care services. Medicaid is administered by the states, and HCFA does not receive any person-level data on Medicaid eligibles, recipients, or payments made for their medical services. The lack of detailed and uniform administrative data has limited evaluation of the program at the national level.

The Medicaid Tape-To-Tape Project was initiated to expand HCFA's ability to collect data to analyze the Medicaid program. The main data base consists of 100-percent data from five participating states (California, Georgia, Michigan, New York, and Tennessee) in uniform codes and formats. These states cover about one-third of the national Medicaid population. States send HCFA their Medicaid Management Information System (MMIS) tapes, which are edited into a comparable format for analysis. Uniform files are produced for each participating state and year. Separated files are maintained for enrollment, claims, and provider data. Claims, provider, and reimbursements can be linked to the Medicaid enrollee who received the service and to the provider who furnished it. The 1980–1982 data from the five participating states have been collected and uniform files completed; 1983–1984 data from participating states are being collected at this time.

The tapes prepared by HCFA contain utilization and expenditure information on all Medicaid enrollees in the five participating states. The tapes include four different person identifiers on each enrollee, making it possible to link the data with other data sets and national surveys. Experimental studies conducted inside and outside the Health Care Financing Administration—some of which have involved elderly utilization patterns—have yielded high match rates. In addition, for that portion of elderly Medicaid users who are also enrolled in Medicare in these five states (a subset of the "dually eligible" elderly), it has also been found that their Medicaid and Medicare records can be reliably linked, with match rates as high as the high 70s to the low 90s (personal communication, David Baugh, Health Care Financing Administration). HCFA staff are currently working on linking the Medicaid tape-to-tape records to the Medicare files.

Recommendation 8.12: The panel recommends that the Medicaid Tape-to-Tape Project be continued, and that the Health Care Financing Administration continue to conduct studies on utilization patterns and expenditures of the elderly using this data base and create sample files and public use tapes for use by outside researchers.

HCFA has undertaken a project to modernize the agency's information system called Project to Redesign Information System Management (PRISM) (Health Care Financing Administration, 1985b). The first stage, development of the system's design concept, was completed in April 1985. Implementation of the entire system is projected for installation by the end of fiscal 1989. Among its goals are increased support of the Medicare/Medicaid Statistical Systems. Completion of PRISM will facilitate implementation of the panel's recommendation.

Medicaid Eligibility Quality Control System

The Medicaid Eligibility Quality Control system (MEQC) was designed to ensure that public funds are spent only on behalf of people who are eligible under federal and state law. It is concerned with identifying ineligible people enrolled in Medicaid and with payments made in error to providers on behalf of those persons. State-level samples are drawn monthly from the Medicaid population in both civilian and institutionalized settings, using the Medicaid case as the sampling unit. The sample cases are checked for errors.

In 1982, the federal agencies responsible for the AFDC, Medicaid, and Food Stamps programs completed a seven-year effort to design a single form: the Integrated Quality Control System (IQCS) form for use in all programs. Although the medical claims may not be useful because they are added together for the entire case and are not collected on a person-by-person basis, there are other valuable data on the form, such as demographics, detailed income and assets, employment, occupation, spend-down amounts, insurance coverage, utilization, diagnoses, and types of services. The IQCS forms are used extensively by the research units in the AFDC and Food Stamps programs, but HCFA does not use them for research.

The panel recognizes that the MEQC system may have potential for research purposes. A national data base on Medicaid cases could be constructed from the re-review sample (a subsample of the IQC sample), which might be less costly than other alternatives, such as sample surveys of individuals and their associated claims or obtaining the entire claims files. Many issues need to be resolved, and despite its interesting prospects, the panel is not making a recommendation on the MEQC system. The National Research Council's Panel on Quality Control of Family Assistance Programs recently completed a study of other aspects of monitoring and analytic needs of the MEQC system (Kramer, 1988).

Data Resources Required to Study the Medicaid Spend-down Phenomenon

A small group of the elderly have medical expenses that exceed the coverage provided by Medicare and any private insurance they may have. These expenses are chiefly incurred for nursing home services. The costs are high, may continue for years, and are rarely covered by private insurance. Although many elderly people think nursing home services are covered by Medicare, they are not. The Medicaid program is the principal source of public financing for nursing home care, paying for services provided to the indigent and the "medically needy"—those whose income and assets fall below a legally defined level.

Many elderly persons deplete both income and assets in meeting their medical expenses. When they have reached the "medically needy" level, they become eligible for the Medicaid program in the District of Columbia and in the 30 states that have programs for the medically needy. Medicaid then covers all their medical expenses, with few exceptions. Income from pensions or Social Security benefits paid to retired wage-earners who need nursing home care may support not only the retired person but also the spouse and other dependents. When the source of support for a family must spend-down to required levels for nursing home care, the family may be left with insufficient income for survival. To revise the legislation for eligibility for Medicaid to eliminate family hardship, data will be required on how often spend-down occurs, the amount of out-of-pocket expenses paid before the Medicaid program takes over, and the effect of the spend-down on other family members.

There are two sources of data for persons in nursing homes. One is the National Nursing Home Survey. In 1985, this survey included an admissions component that collected data for a sample of admissions and was designed to produce estimates on the spend-down issue.

A second source of data is the Medicaid Quality Control sample. The sample collects information (among other items) on how eligibility was established for Medicaid. Information on persons involved in the spend-down can be compiled for the 5 percent of Medicaid beneficiaries who are in nursing homes and are selected in the Quality Control sample of 400,000 persons nationwide.

Information about the effect of spend-down on the family may be obtained from two ongoing surveys. The Survey of Income and Program Participation in its health care module collects information on insurance coverage, including Medicare and Medicaid for the sample household. It should be possible to examine the detailed data on income and assets in relation to health insurance coverage during the 2 1/2 year period in which the same panel remains in the sample. Changes in coverage and assets could be related. The value of this data source would be greatly enhanced if the recommendation in Chapter 10 to increase the size of the sample of persons of age 65 and over is implemented. Sample augmentation would double the number of those ages 75–84, and 85 and over—the ages in which nursing home admissions are highest.

The NHIS Supplement on Aging, and the subsequent follow-up through the Longitudinal Study on Aging, will also develop information on the family and the individual admitted to the nursing home, if panel recommendation 7.3 on a biennial follow-up is implemented.

The panel reaffirms the need for the expansion of the SIPP sample of the elderly, and for continuation of the Longitudinal Study on Aging in order to study the spend-down phenomenon. In addition, data from the Medicaid quality control data base and from the Nursing Home Survey should be analyzed for information on this problem.

Private Insurance Data

Although at least 25 companies now offer free-standing long-term care insurance, insurers are reluctant to offer and market long-term care policies aggressively. Insurance industry representatives point to concerns about adverse selection, insurance-induced demand, pricing difficulties, and lack of consumer education as barriers to product development. Some insurers have expressed fear that the open-ended liability that can result from long-term care policies could be financially devastating to their companies. There is also concern about the long lag time between purchase of policies and payment of substantial long-term costs and that nonmedical, personal services such as homemaker care and respite care are not insurable.

Insurers also find that actuarial estimates and premium determinations for long-term care policies are difficult to make. Minimal data are available on costs and utilization of long-term care insurance, and very little of the data reflects actual experience. Some are concerned that problems could occur if only high-risk individuals are attracted to long-term care insurance. At present, there is no reliable actuarial model applicable to a long-term care policy that would differentiate the high-risk purchaser from the low-risk one and allow for a variable rate scale.

Private insurance companies are beginning to gain experience with long-term care insurance. The panel recognizes the need for improved data on utilization of covered services, costs, risk management, marketing, and the impact of long-term care coverage.

Cost-of-Illness Data

Cost-of-illness data play an important role in decision making regarding the allocation of resources in the health sector. Illnesses, such as Alzheimer's disease, that primarily affect the elderly will require more resources in the future with the growing number of elderly persons who may be at risk. Alzheimer's disease, affecting an estimated 1.5 million Americans, has become a major priority for federal research organizations (e.g., the National Institute on Aging, the National Institute of Neurological and Communicative Disorders and Stroke, and the National Institute of Mental Health). Many other diseases and impairments that affect the elderly include heart disease, cancer, arthritis, stroke, as well as visual and hearing impairments.

In addition to data on prevalence, incidence, and use of medical care and long-term care services for the elderly suffering from these and other conditions, the costs of these conditions are needed by policy makers, health planners, and researchers to set priorities, make program policy decisions, and prepare and deliver congressional testimony to support program policy decisions and agency budgets.

The economic costs of illness represent the monetary burden on society of illness and premature death. They represent foregone alternatives and are measured in terms of the direct and indirect costs. Direct costs are the value of resources that could be allocated to other uses in the absence of disease, and indirect costs are the value of lost output because of cessation or reduction of productivity due to morbidity and mortality. Morbidity costs are lost wages for people unable to work due to illness and disability and an imputed value for those persons too sick to perform their usual housekeeping services. Mortality costs are the present value of future earnings lost for people who die prematurely, employing discounting to convert a stream of future earnings into present values.

Total economic costs of illness in 1980 amounted to $455 billion based on a 4 percent discount rate of the value of productivity foregone in succeeding years as a result of premature mortality in that year. The elderly (persons age 65 and over) comprised 11.3 percent of the total population in 1980 and 18.2 percent of the total economic costs (Rice et al., 1985).

The rankings by major diagnostic category of the economic costs of illness vary substantially by age. For the population under age 65, the medical condition that ranks highest in economic costs is "injury and poisoning," costing $78 billion, accounting for 21 percent of the total for this age group, and reflecting the relatively high value of lost productivity for the large number of premature deaths at younger ages from this cause. Diseases of the circulatory system rank second in economic costs for persons under age 65, representing 15 percent of the total. For the elderly, the economic costs of diseases of circulatory system far outrank all other diseases, amounting to $29 billion, or 35 percent of the total. In second place are neoplasms, constituting 11 percent of total economic costs for the elderly.

These cost-of-illness estimates are for the major diagnostic categories and are not disaggregated to specific diseases. There have been more than 200 separate cost-of-illness studies in the last 20 years (Hu and Sandifer, 1981). Some of these are national in scope, but most are limited to a selected population or geographic area, and all but a few are restricted to one or a few disease categories. Varying methodologies are used so that the costs of different diseases cannot be compared.

Estimation of the costs of illness depend to a great extent on the data available and on the methodologies used. Although the U.S. Public Health Service has developed guidelines for estimation of the costs of illness (Hodgson and Meiners, 1982), few studies follow them rigorously. For the further development of cost-of-illness studies, data will be available from several surveys including the 1987 National Medical Care Expenditure Survey, the 1985 National Nursing Home Survey, the annual National Hospital Discharge Survey, and the annual National Health Interview Survey. Additional sources of data are the Medicare, Medicaid, and other public program administrative records.

Recommendation 8.13: The panel recommends that the National Center for Health Services Research, the National Center for Health Statistics, and the Health Care Financing Administration continue to collect the detailed data necessary to estimate the economic costs of illnesses, especially those affecting the elderly population, and that the National Center for Health Services Research support cost-of-illness studies using available guidelines for uniform methodology.

Copyright © National Academy of Sciences.
Bookshelf ID: NBK217723


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.8M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...