Capitated Payment of Medicaid Long-Term Care for Older Americans: An Analysis of Current Methods

#2001-03 March 2001 Capitated Payment of Medicaid Long-Term Care for Older Americans: An Analysis of Current Methods by Richard Kronick, Ph.D. and Tony Dreyfus, MCP The Public Policy Institute, formed in 1985, is part of the Public Affairs Group of the AARP. One of the missions of the Institute is to foster research and analysis on public policy issues of interest to older Americans. This paper represents part of that effort. The views expressed herein are for information, debate, and discussion, and do not necessarily represent formal policies of the Association. 2001, AARP. Reprinting with permission only. AARP, 601 E Street, N.W., Washington, DC 20049

ACKNOWLEDGMENTS We would like to thank numerous people who contributed to our understanding of capitation for long-term care: Susan Barth of the New York Department of Health; Pam Coleman and Bill Farnsworth of Texas; Tom Hamilton and Tom Lawless of Wisconsin; Bob Hurley of the Medical College of Virginia; Jim Lubitz, Don Sherwood, and Noemi Rudolph of HCFA; Joelyn Malone and Pam Parker of Minnesota; Hunter McKay of the Office of the Assistant Secretary for Planning and Evaluation; Mark Meiners of the University of Maryland Center on Aging; Dan Milne and Reid Reynolds of Colorado; Erin Nagy of the American Public Human Services Association; Foster Northrup of William M. Mercer, Inc.; Chris Perrone of Massachusetts; Paul Saucier of the Muskie Institute, Maine; Jim Verdier of Mathematica. We would also like to thank Janet O'Keeffe and Jane Tilly, who were formerly employed at AARP, and Joyce Dubow, John Luehrs and others presently at AARP for their work in conceptualizing and revising the paper.

TABLE OF CONTENTS Foreword... i Executive Summary...ii Introduction... 1 I. The Growing Interest in Capitating Long-Term Care... 1 Integrating Acute and Long-Term Care Services Under a Capitated Approach... 3 II. Current Payment Methods for Long-term Care... 4 Nursing Home Care... 4 Home and Community-Based Services... 5 III. Setting Capitated Payment Rates for Long-Term Care... 7 Defining the Comparison Group... 7 Calculating Expenditures per Eligible Month... 8 Adjusting the Cost Experience... 8 IV. Policy Considerations in Moving to a Capitated Payment Approach for Long-Term Care...9 Defining the Comparable Population Policy Implications... 10 Facing the Challenge of Serving Additional Older Beneficiaries... 12 Conclusion: Evaluating Whether to Capitate Long-Term Care... 12 Appendix. Case Examples of Capitated Payment Systems for Long-Term Care... 14 The Arizona Long-Term Care System... 14 The Program of All-Inclusive Care for the Elderly... 15 Minnesota Senior Health Options... 17 The Texas Star+Plus Program... 18 Endnotes... 26 References... 29 List of Tables Table 1... 6 Table 2... 6 Table 3... 20 Table 4... 22 Table 5... 23

FOREWORD In recent years there has been a shift in the delivery of long-term care services from institutions to the home and community. The reason is two-fold. First, home and communitybased care has the potential to be more cost effective. Second, long-term care recipients prefer to remain in their homes and avoid institutional settings. Notwithstanding the benefits of home care, programmatic limitations in Medicaid hamper the substitution of home care for institutional care for Medicaid beneficiaries. As the number of individuals over age 85 increases as a percentage of the U.S. population, it is likely that there will be a significant rise in the demand for long-term care services. In an effort to restrain expenditure growth and improve services, state Medicaid programs have shown an increasing interest in developing new methods of long-term care financing. One approach is to capitate long-term care, either on its own or in an integrated system of acute and long-term care. Under a capitated system, states would have greater flexibility to provide long-term care in less costly home and community-based settings, while controlling long-term care expenditures. AARP s Public Policy Institute initiated this study to obtain information about current efforts to capitate long-term care. This report provides a discussion of several important considerations that policymakers need to weigh when deciding whether or not to adopt a capitated approach for financing long-term care. It also describes a variety of efforts currently underway to use capitated long-term care financing. Although this report does not provide a recommendation about the appropriateness of capitating long-term care, it does provide a framework for considering the costs and benefits, thus enhancing the knowledge base that policymakers can turn to when evaluating long-term care financing methods. Rick Kronick is an Associate Professor in the Department of Family and Preventive Medicine at the University of California San Diego. He has assisted state Medicaid programs in the design and implementation of risk-adjusted payment systems for HMOs. Tony Dreyfus is an independent consultant with broad experience in the study of health care financing issues. Stephanie L. Bernell, Ph.D. Senior Policy Advisor AARP Public Policy Institute i

EXECUTIVE SUMMARY Background The number of older Americans, who are at greatest risk of needing long-term care (LTC) services, will rise steadily over the next decades. Many individuals will clearly need services provided in a nursing home. However, for many, services could be provided effectively in the home or community, which are by and large preferred settings. Unfortunately, due to limitations within the Medicaid program, there is an institutional bias towards providing services in nursing homes. Without policy changes, the share of state budgets devoted to LTC for older persons mostly nursing home care is expected to increase over the next 20 years from its current level of about 4 percent to about 8 percent. Consequently, state Medicaid programs have shown increasing interest in capitated LTC financing. If LTC capitation provides an incentive to serve people in less costly, preferred home and community-based (HCB) settings, then it has the potential to improve the availability of services, improve consumer satisfaction, and decrease state LTC expenditures. One approach is to capitate LTC on its own, while a second is to capitate an integrated system of acute and long-term care. The few programs that currently capitate LTC for example, those in Arizona and Texas aim to facilitate flexibility in the provision of services, decrease reliance on nursing homes, and control expenditure growth. Most states already use HCB services waiver programs in an attempt to achieve the same goals. However, the advantages of LTC capitation over current approaches to paying for long-term care have not yet been clearly demonstrated. In addition, LTC capitation raises concerns that such financing could lower quality and inappropriately restrict services. Purpose The purpose of this paper is to provide information on long-term care capitation that will aid policymakers in assessing this approach to LTC payment for all types of LTC services. The paper describes and assesses methods that states use to set long-term care capitation rates. It also addresses the benefits and challenges involved in capitating long-term care. Methodology The paper is based on a review of the published LTC literature and official documents concerning current programs of capitated long-term care. The authors also conducted in-depth interviews with more than 20 informants from state and federal government and elsewhere. ii

Discussion States already have considerable experience in two areas that are closely related to capitating LTC: accounting for case mix as part of nursing home payments, and providing HCB services through waiver programs. Almost all methods of setting capitated rates for LTC base the rate on the cost experience of a group of people comparable to those projected to be served in the capitated program. Identifying the comparison group is the most challenging step in the rate-setting process. A key beneficiary characteristic used for capitated LTC rate setting is whether individuals health conditions and functional limitations would make them eligible for nursing home care. States vary greatly in how they assess the need for nursing home care, with some requiring only functional limitations, others requiring nursing needs, and others a combination of both. The level of functional impairment required also varies by state, but a need for assistance in at least two or three activities of daily living is common. To calculate capitated LTC rates for beneficiaries who are nursing home-eligible, states use various blends of the expenditures for nursing home residents and nursing home-eligible beneficiaries living in the community. In most states, expenditures for people in nursing homes are two to three times the expenditures in waiver programs, creating a wide range in which capitated rates could fall. In setting payment rates, states must decide which services will be covered, calculate projected expenditures per eligible month, and adjust the cost experience for additional factors, such as inflation, managed care savings, recent changes in the Medicaid program, and beneficiary cost-sharing. Conclusions States that pursue capitation of LTC must carefully consider several questions. What advantages does capitation offer over established HCB waiver programs? How can states use capitation to extend services to the many older individuals living in the community who could qualify as nursing home eligible without placing greater demands on the state budget too quickly? If states choose to adopt a capitated approach, how will they deal with the substantial state administrative effort that will be required to establish and manage the program and how will they address concerns that capitation might lead to underservice? The central policy question of whether capitating LTC services is a good idea will need to be answered in part with more practical experience, including more years with a large number of enrollees. However, existing programs, such as the Program of All-Inclusive Care for the Elderly and programs in Arizona, Minnesota, and Texas, which are described in the report, already iii

provide some useful information about the kinds of rates and incentives that certain rate setting approaches yield. These capitated LTC programs deserve close observation. States with the broader goal of integrating acute and LTC in a capitated program for persons age 65 and older face significant difficulty, however, because such integration requires Medicare funds to cover the costs of acute care. The federal government prohibits the mandatory enrollment of Medicaid beneficiaries in health maintenance organizations for Medicare-covered acute care services, which prevents states from creating large integrated capitated programs. Additionally, beneficiaries eligible for both Medicare and Medicaid appear reluctant to enroll voluntarily in programs where they will lose their freedom to choose providers. As a result, prospects for the broader goal of capitating LTC with acute care are mixed. It seems likely that wider experiences of separately capitating acute care for older Medicare beneficiaries and capitating LTC for Medicaid beneficiaries is needed before efforts at full integration can proceed on a larger scale. iv

CAPITATED PAYMENT OF MEDICAID LONG-TERM CARE FOR OLDER AMERICANS: AN ANALYSIS OF CURRENT METHODS INTRODUCTION The number of older Americans, as a percentage of the population, is rising steadily. As a result, over the next several decades, many older persons will need long-term care (LTC) services. While some individuals will need services provided in the nursing home, many others can be served effectively in the home or community, the setting most people prefer. Limitations within the Medicaid program, however, create a bias towards providing services in institutional settings. State programs have shown increasing interest in using capitated financing to pay for long-term care. 1 Based on their experience in using a capitated payment system in acute care settings, states view this approach to financing as a possible means to decrease reliance on nursing homes as well as restrain expenditure growth. If LTC capitation provides an incentive to serve people in less costly, preferred home and community-based (HCB) settings, it may also have the potential to improve the availability of services, increase consumer satisfaction, and decrease state long-term care expenditures. The purpose of this paper is to provide a basic, non-technical framework for considering LTC capitation. Section I provides the policy and program context for the increased interest in LTC capitation. Section II describes how states traditionally pay for long-term care. Section III describes the methods states use to calculate rates for LTC under capitated arrangements, and Section IV discusses several key considerations policymakers will need to weigh in deciding to move to a capitated approach for financing long-term care. The Appendix describes the efforts of several programs and states to develop capitated financing approaches for LTC services. These include the Program of All-Inclusive Care for the Elderly (PACE) and programs in Arizona, Minnesota, and Texas. I. THE GROWING INTEREST IN CAPITATING LONG-TERM CARE Long-Term care is currently provided in two major ways nursing home services and HCB services. Care in a nursing facility is a federally mandated service for state Medicaid programs, whereas states decide whether and how much to cover HCB services. Home and community-based programs have grown significantly in recent years, although current public expenditures for LTC are devoted mostly to nursing home care. Of the $59 billion in national Medicaid expenditures in 1998 for LTC, the vast majority, about 58 percent, was for institutional care 25 percent going to home care and close to 17 percent going to intermediate care facilities for the mentally retarded and developmentally disabled (HCFA, 2000). 1

Currently, approximately as many older persons with limitations in three or more activities of daily living (ADLs), such as bathing, dressing, and toileting, live in the community as live in nursing homes, and many of these individuals would likely qualify as nursing homeeligible. 2 In addition, an equally large group of older persons have fewer but still significant functional impairments; some might also be nursing home-eligible depending on their condition and on the eligibility rules in their state. However, many individuals who meet Medicaid financial eligibility requirements are not receiving nursing home or HCB care because they have not been evaluated for the program or there is insufficient space in the programs to accommodate them. Even many impaired older persons in the community who would not meet Medicaid s financial eligibility requirements would and could benefit greatly from supportive services in their homes or the community. 3 In the absence of HCB services, only a strong preference to avoid the nursing home and the support of family, friends, and some formal services keep many impaired older persons in the community. Much of the interest in capitating the payment for LTC services stems from its potential as a payment mechanism that could facilitate a more appropriate balance of HCB and nursing home services. For example, some older Medicaid eligible persons who need a nursing home level of care prefer to receive services in their own homes, but cannot obtain services because all the state-funded slots designated for these services are filled. A managed care organization receiving capitation that includes payment for both HCB services and nursing home services would not be confronted with this problem. It could provide the enrollee with needed HCB services at a lower cost than that required for nursing home services if it could spend resources creatively on new and more efficient approaches to providing needed services. As much as combining LTC services under a capitated payment system has created hope for improving the availability of services in preferred settings and constraining expenditures, it has also raised concerns. One of these concerns is the incentive created by capitation to stint on necessary services. Under the typical full-risk arrangements of capitated plans, there is an incentive to deliver fewer services in order to decrease expenditures. Indeed, much of the current ill-will toward managed care stems from the concern that health maintenance organizations (HMOs) limit access to needed services in order to increase profits. If payment for LTC services is capitated, then the incentive to under-serve could be extended to a population who are often vulnerable and less able to advocate or make decisions for themselves than persons in managed care health plans. State policymakers also need to understand that capitation will not easily restrain growth in public LTC expenditures. When states implement programs that capitate acute care for younger Medicaid beneficiaries, budget analysts have usually assumed that savings will occur in the next year's budget. In contrast, LTC capitation does not promise substantial short-run savings, in part, because of Medicaid's already fairly economical purchasing of LTC and in part because of substantial unmet needs among beneficiaries for long-term care services. Although HCB services can often be provided to older persons at a lower cost than nursing home services, experience has shown that many individuals who are nursing home-eligible, but who elected not to go into a nursing home, might be attracted to programs that offer services in the community. 2

Thus, states should view LTC capitation as a possible long-run method of restraining expenditures and encouraging the growth of high quality, community based services, but not as a method of saving money in the short term. Integrating Acute and Long-Term Care Services Under a Capitated Approach The potential for LTC capitation to shift spending from nursing facilities to HCB settings, to restrain expenditures, and to meet the needs of a growing population in the future depends on the scope of services included in the capitated payment. In general, a broader range of services included under capitation should create the possibility for the development of more comprehensive and coordinated systems. This is why some programs have implemented a capitation approach that goes beyond LTC services alone. This broader approach integrates acute care covered by Medicare with LTC provided by Medicaid. A concerted system of medical, therapeutic, and supportive services could help many older persons enjoy better health, functioning, and independence than they do under the current separate systems for acute and long-term care. The range of services would be provided whenever possible in the community and whenever necessary in hospitals and nursing homes according to the needs of the individual, not the dictates of the source of coverage. More resources devoted to supportive services, preventive care, and timely primary care might reduce the need for acute hospitalization and reduce or postpone the need for nursing home care. This larger pooling of acute and LTC dollars represents some of the more far-reaching hopes for creating a coordinated approach to care for older persons. The interest in capitating an integrated acute and LTC system stems from states concern that the division of responsibility for acute and long-term care services between Medicare and Medicaid, and the resulting cost shifting between the two programs, may lead to duplicative and avoidable expenses, poor care coordination, and poor outcomes. The current payment arrangements do not foster aggressive preventive care and timely primary care that might improve health status, save money, and reduce use of nursing homes. The capitation of LTC in integrated systems is seen as a method that has the potential to improve both health and long-term care outcomes. In addition, capitation's potential to improve services and restrain expenditure growth has more promise for an integrated package of acute and long-term care than for LTC alone. The financing of long-term care through Medicaid has been fairly economical (Wiener and Stevenson 1998b, p.98), and so LTC capitation offers much less potential for savings than capitation of the acute care fee-for-service system, which grew large and comfortable with third-party insurance and unmanaged specialist, hospital, and high-technology care. States with the broader goal of integrating acute and LTC in a capitated program for persons age 65 and older face significant difficulty, however, because such integration requires Medicare funds to cover the costs of acute care. The federal government prohibits the mandatory enrollment of Medicaid beneficiaries in HMOs for Medicare-covered acute care services, which prevents states from creating large integrated capitated programs. 4 Additionally, beneficiaries eligible for both Medicare and Medicaid appear reluctant to enroll voluntarily in programs where 3

they will lose their freedom to choose providers. As a result, prospects for the broader goal of capitating LTC care so as to integrate it with acute care are mixed. The Medicare program has explored the idea of integrating acute care and LTC through a demonstration program. The national Social Health Maintenance Organization (S/HMO) demonstration program was initiated in the mid-1980s to test the integration of acute care and a prescribed set of LTC services within a capitated managed-care framework. Under the demonstration, the S/HMO provides the full range of Medicare benefits offered by standard HMOs plus additional limited LTC benefits including some nursing home care, and a full range of HCB services such as homemaker services, personal care services, adult day care, and respite care. The initial evaluation found that although S/HMOs offered long-term care services, they did not develop a well-coordinated system of care; in part because many physicians were unaware of the LTC benefit package. The evaluation also found that S/HMO enrollees with functional impairments were less satisfied than either unimpaired S/HMO enrollees or functionally impaired beneficiaries in the traditional Medicare program (Newcomer et al. 1994) The Congress mandated the second-generation S/HMO demonstration in 1990. This newer demonstration embraces similar objectives, but refines the financing methods of the first S/HMOs and improves the benefit design. Specifically, there is a greater emphasis placed on geriatric care, and Medicare capitation rates are risk adjusted for enrollee s reported health status, ADL limitations, and other characteristics. The Health Care Financing Administration (HCFA) chose six organizations to participate in the second-generation program, however only one, Health Plan of Nevada, has become active. II. CURRENT PAYMENT METHODS FOR LONG-TERM CARE Almost all capitated programs use the enrollment and expenditure experience of a comparable group of non-capitated beneficiaries as the foundation for setting payment rates. Thus, it is instructive to review current methods used to pay for LTC as background for understanding the rate setting calculations and policy decisions that will be necessary if states decide to pursue a capitated option for financing long-term care for their older residents. Nursing Home Care Nursing homes provide skilled nursing care, protective oversight, rehabilitative services, and room and board to individuals who, because of their mental or physical condition, require care and services that can be obtained most readily in institutional facilities. States generally pay for nursing home care through a set payment for each day of care provided, averaging $99 per day in 2001. In many states, the per diem Medicaid payments to nursing homes vary according to several factors. First, the per diem can vary by facility, with different nursing homes having rates that depend to some degree on their cost experience. Thus, if a nursing home spends more than the average nursing home in a state, it will get higher-than-average payments in the 4

following year. This facility-specific approach provides more money to nursing homes that provide better amenities. Often, these are homes with a high proportion of private pay residents. Second, the per diem can vary according to the case mix or level of need of the residents, with higher per diem rates paid for the care of residents who are evaluated as needing greater levels of services. Higher payments to nursing homes with a more severe case mix is a key technique used by some states to encourage nursing homes to assume responsibility for older persons with greater needs. Case-mix adjustment can also reduce the likelihood that nursing homes will offer substandard care to older persons with above-average needs. States that do not use case-mix reimbursement systems often have other systems of payment that recognize to some degree variation in the level of care needed. For example, a state may pay one rate for a skilled level of care and another rate for lower levels of care. 5 Home and Community-Based Services Home and community-based care programs provide a wide range of services, including personal care to assist with bathing, dressing, toileting and other ADLs; adult day health care; respite care; homemaking; and nursing or rehabilitative care, such as assistance with medications, monitoring of health status, and physical therapy. Medicaid typically pays for these services on a fee-for-service basis. States can provide HCB services through (1) the regular Medicaid program, (2) programs funded solely by states, or (3) programs established through waivers of Medicaid regulations. Under the regular Medicaid program, all states must cover home health care for persons age 21 and older who meet nursing home functional eligibility criteria; some cover limited amounts of personal care and other HCB services as well. In solely state-funded programs, the state can extend eligibility to people who do not qualify for Medicaid; total spending for such programs was only $1.2 billion in 1997 (Wiener and Stevenson 1998b, p.89). Approximately 110,000 persons are served in waiver programs exclusively for older persons, and an additional 200,000 are served in waiver programs that care for both older and disabled persons. Expenditures on these 310,000 persons in 1997 were approximately $2 billion dollars a small fraction of the over $30 billion in Medicaid expenditures for all long-term care (i.e., nursing homes and HCB) for older persons. Home and community-based waiver programs represent an important method for states to develop alternatives to institutional care. With a Section 1915(c) waiver, a state can expand eligibility for Medicaid LTC services in the community by applying a more liberal methodology in determining financial eligibility. The waiver also allows the state to set a strict limit on the number of individuals it will serve and the amount of dollars it will spend per participant. The HCB waiver programs stand in sharp contrast to acute care. States have little control over the number of acute care service users or the cost per acute care user in fee-for-service programs. In the waiver programs, states can control both the number of program participants and the cost per person. The ability to control the overall size of the waiver programs has reduced states' 5

concerns that the availability of LTC services outside of nursing homes will lead to an uncontrolled increase in demands for service. Table 1 provides summary data on waiver programs and Table 2 presents more detailed state data on waiver programs that serve older persons. Table 1. Numbers Served and Expenditures in Home and Community-Based Waiver Programs Waiver Type Number Served Expenditures (million $) Annual Per Capita Cost of Waiver Program Annual Per Capita Cost of Institutional Alternative Per Capita Waiver Cost/ Per Capita Institutional Alternative Older 110,521 $493 $8,450 $22,260 38% Older and Disabled 196,134 $1,499 $11,440 $23,170 49% Mental Retardation 175,901 $4,638 $35,290 $87,230 40% Other 33,257 $367 $19,690 $40,540 49% Total 515,813 $6,997 $19,460 $45,940 42% Source: Authors' analysis of American Public Human Services Association,1915(c) Database, 1997. Averages are weighted by the number served. Table 2. Number Served and Expenditures for Older Participants of Waiver Program Annual Per Capita Cost of Waiver Program 6 Annual Per Capita Cost of Institutional Alternative Per Capita Waiver Cost/ Per Capita Institutional Alternative Number State Served Alaska 411 $23,796 $62,171 38% Arkansas 7,771 $4,677 $13,037 36% California 8,004 $9,478 $20,121 47% Connecticut 13,026 $14,136 $36,737 38% Florida 4,476 $18,374 $25,765 71% Illinois 16,448 $4,685 $20,853 22% Iowa 2,874 $11,818 $14,703 80% Kansas 5,662 $9,359 $16,741 56% Massachusetts 3,174 $4,137 $26,586 16% Minnesota 6,923 $8,537 $21,970 39% Missouri 18,946 $5,805 $17,165 34% Ohio 19,666 $7,762 $23,246 33% Pennsylvania 1,298 $14,758 $28,881 51% Rhode Island 600 $4,446 $34,000 13% Utah 542 $9,680 $19,886 49% Average 7,321 $8,450 $22,260 38% Source: Authors analysis of American Public Human Services Association 1915(c) Database, 1997. Averages are weighted by number served.

A notable feature of the HCB waiver program for older individuals has been that their average annual expenditures of about $8,500 per capita are only about 40 percent of the costs of institutional care. Similarly, the HCB waiver program expenditures for older and disabled individuals are about 50 percent of the costs of institutional care. Although federal waiver rules limit waiver expenditures to be no higher than institutional costs, actual waiver expenditures are much lower. This level of expenditure is probably achieved, in part, through protocols for state case managers and contracts with counties or other agencies that administer the waiver program, and in part because the LTC needs of community-based enrollees are perhaps less than those of nursing home residents. III. SETTING CAPITATED PAYMENT RATES FOR LONG-TERM CARE Setting a capitated payment rate involves several steps, all of which incorporate the enrollment and expenditure experience of a comparison group. Federal Medicaid regulations stipulate that states may not spend more in the new capitated program than the amount the state would spend on beneficiaries through traditional fee-for-service payments. Even without such federal regulation, budget-conscious state legislators and Medicaid administrators would generally not want to spend more on a capitated program than they would on non-capitated payments. Defining the Comparison Group The initial and most challenging step in rate setting is to define a group that is comparable to the group that will be served by the capitated program the comparison group. The closeness of the final capitation rate to what otherwise would have been spent will largely depend upon how well the comparison group approximates the group in the capitated program. All capitated LTC programs to date either cover nursing home-eligible persons only or have a special rate for them in programs that cover non-nursing home eligible persons as well. To define the comparison group, ideally, a state would have data on all nursing home eligible individuals (in other words, persons who would meet the nursing home eligibility criteria if they were evaluated). However, states have data only for nursing home eligible persons who are currently in nursing homes and persons who are in HCB waiver programs. States therefore use these groups' expenditures to estimate the expenditures for all beneficiaries who are nursing home-eligible. In fee-for-service, there are two main sources of variation in the amount of LTC expenditure: level of need and site of care. Comparing beneficiaries with the same level of need, expenditures for those in nursing homes are greater, on average, than for similar beneficiaries in the community. On average, however, beneficiaries in nursing homes have greater needs than nursing home eligible beneficiaries who remain in the community. Therefore, when setting rates for capitated programs, states need to consider both program enrollees' level of need as well as the likely site of care. If most program enrollees would likely receive services in the community 7

rather than in a nursing home, then payment close to the level of waiver expenditures is likely appropriate. Conversely, if substantial numbers of enrollees would likely be in nursing homes in the absence of the capitated program, then payment closer to the nursing home level is appropriate. In most states, per capita expenditures for people in nursing homes are two to three times expenditures in waiver programs, creating a wide range in which the rate could fall. In capitated LTC, adjusting the rate to take into account the needs of the beneficiaries could play a key role in focusing more services on older persons with greater needs. Effective risk adjustment for LTC could do much to allay concerns among payers and plans that adverse selection of beneficiaries into plans could lead to large overpayments or underpayments. Medicare's experience with voluntary enrollment in HMOs for acute care has shown that overpayments can result when healthier-than-average beneficiaries enroll disproportionately. Plans that design attractive systems have reason to fear drawing an adverse selection of frail beneficiaries who will need a greater-than-average level of services. The goal of moving care from institutions to the community inevitably complicates comparability. In most states, there are many Medicaid beneficiaries in the community, but not in waiver programs, who would be nursing home eligible if their levels of need were evaluated. A program to serve more nursing home eligible people in the community can attract many people who are not as impaired as those actually in nursing homes. Calculating Expenditures per Eligible Month After deciding which services capitated plans will provide, the state must identify the state s expenditures for these services for the comparison group. These expenditures must then be divided by the length of time that beneficiaries are eligible to receive services thus calculating expenditures per eligible month. 6 Adjusting the Cost Experience Rate setters must adjust the cost experience to account for several additional factors. Inflation. The cost experience for the comparison group needs to be adjusted for any anticipated differences between the costs for the comparison group and the group for which the capitated rate will be paid. For example, if services to the comparison group were delivered in 1998 and the services for the capitated group are delivered in 2000, then rate setters would typically increase the expenditures made for the comparison group by a trend factor to adjust for anticipated inflation. Determining the trend factor is often more art than science, and is an area with substantial flexibility. The trend factor can rest on a projection of the medical component of the consumer price index, reflecting the expected increase in unit prices in the health care industry. Alternatively, it can be an estimate of the recent trend in Medicaid expenditures per eligible month. 8

Managed care savings. A second type of adjustment is commonly made to reflect savings on fee-for-service expenditures that the state believes should be achieved through capitation. In setting rates for acute care, states commonly discount fee-for-service expenditures by 5 percent, but smaller or larger discounts can also be used. Program changes. Adjustments are also made to reflect recent changes to the Medicaid program. For example, if the Medicaid program in 1999 had increased nursing home payments by 10 percent or agreed to pay for a new type of community-based service, the estimated effects of these changes on expenditures per eligible month would be incorporated into the rates. Beneficiary share of cost. If the state is using actual expenditures in claims to calculate the costs of the comparison population, then the claims should already reflect the small contribution made by the beneficiary. However, if the state is using the nursing home rate to calculate its expenditures, it will have to subtract the beneficiaries average contribution from the rate to more accurately estimate state costs (PACE 1996, p.6-7). IV. POLICY CONSIDERATIONS IN MOVING TO A CAPITATED PAYMENT APPROACH FOR LONG-TERM CARE The central policy question of whether capitating LTC services is a good idea will need to be answered in part with more practical experience, including more years with a large numbers of enrollees. However, existing efforts, such as the PACE program and programs in Arizona, Minnesota, and Texas, which are described in the Appendix, already provide some useful information about the kinds of rates and incentives that certain rate setting approaches yield. These programs to capitate LTC deserve close observation. The new payment systems have various positive features. For example, the single rate for nursing home eligible beneficiaries regardless of where they reside, creates a strong incentive to provide services in the community. Minnesota's payment system shields plans from the full liability for nursing home care, but still encourages plans to provide care in the community through partial liability for nursing home care. Given the wide variation in need among beneficiaries who are nursing home eligible, it is difficult to set a single rate that is accurate when enrollment is voluntary. States face no shortage of areas in which they can learn how to better serve older persons. For new programs to create real improvements in services for older persons, they must do far more than just capitate or just bring together Medicare and Medicaid money. Sustained learning and development efforts are needed both on the side of financing and administration and on the side of innovation in service delivery. States and health plans will have to learn more about a range of topics, such as the different mixes of professionals who can serve older persons who have different levels of frailty, alternative methods of communicating about new programs, and various ways to develop a role for consumer direction of services. More effective coordination 9

of supportive services such as older housing, meals, and transportation could also play an important role in improving services. With all these difficulties, states that seek a better array of services from the Medicare and Medicaid dollars available are tackling an important challenge with a great potential benefit for older persons. The remainder of this section discusses in more detail several important issues that policymakers will need to consider as they decide whether and how to pursue a capitated approach to financing long-term care. The key rate setting difficulty of defining the comparable population is addressed first, followed by an examination of whether states can use capitation to extend services to the many nursing home-eligible people in the community without placing great demands on the state budget too quickly. Defining the Comparable Population Policy Implications The rate setting challenge of identifying a group of people who are comparable to those who will be served in a capitated program has been met in various ways, especially for the key group of those who are eligible for nursing home services. As discussed in Section III, when setting payment amounts for these Medicaid beneficiaries, states must decide whether to pay the high amount spent on nursing home residents, the much lower amount spent on eligible beneficiaries in community-based waiver programs, or some amount in between. Some existing capitation programs, such as PACE, have been paying at the high end, either at nursing home levels or at a blend of nursing home and waiver expenditures weighted strongly toward the nursing home. While these rates might be appropriate for the beneficiaries who typically enroll in PACE, they are much higher than most states are currently paying for the average beneficiary in an HCB waiver program. If states were to adopt the PACE rate setting methods for a capitated LTC program that covered all nursing home-eligible beneficiaries in the community, overall expenditures would increase significantly. Other programs have generally adopted alternative approaches to determining capitation rates. Arizona blends nursing home and community-based expenditures using a target mix that slightly increases the community-based share over the previous year. The single rate appears to have stimulated contractors to increase community-based care, which probably has attracted additional enrollees more than it has diverted people from nursing homes. Texas, by contrast, pays different rates for beneficiaries living in nursing homes than for eligible beneficiaries living in the community, with the rate for nursing home residents based on nursing home costs and the rate for community residents based on the expenditures in the waiver program. This approach may encourage a modest reorientation away from the nursing home and toward community-based services. Minnesota bases its rates for beneficiaries in the community on waiver expenditures, but offers incentives to provide care in the community by requiring plans to cover the first six months of nursing home costs. A bonus paid for residents in the community who have moved 10

out of a nursing home also helps create incentives to shift services to the community. So far, however, few community-based beneficiaries have enrolled, and very few long-term nursing home residents have moved to the community. All of the current rate setting techniques assume that non-capitated spending is at the right level and that the problem is finding the appropriate comparison group for the capitated program. But this assumption is certainly open to question. In some states, excess nursing home capacity may have led to levels of expenditure that were higher than appropriate, at least for those in nursing homes. On the other hand, in some states, low reimbursement rates and limits on covered services may have caused spending to fall below what might be desirable. In the short run, budget constraints and federal rules are likely to keep capitation rates fairly close to a fee-for-service equivalent. In the long run, data from plan experience may guide rate setters to depart substantially from the fee-for-service equivalent. An important limitation of all these rate setting techniques appears to be the use of a single rate for nursing home-eligible beneficiaries living in the community. Nursing home eligibility alone is not a narrow enough category for use in widespread rate setting because this category includes people with very different levels of need. Other factors are needed to make better predictions about resource needs, including more detailed information on functional status, cognitive status, illness, living situation, and family supports. Given the wide range of need among those receiving or eligible for nursing home care, a single rate could encourage adverse selection when beneficiaries can choose among plans (or between capitated plans and fee-for-service). In the same vein, each plan would be eager to enroll people with lesser needs, particularly those who are the least impaired in the community. Without some form of risk adjustment, plans might seek the easiest of community cases and avoid those who have greater LTC needs. As a result, a single rate might have limited effects in encouraging plans to bring new resources to the community, or to develop new methods of caring for more impaired people in the community. To accomplish the goal of shifting resources through capitation, plans must have both the motivation and the resources to provide services in the community that will help keep people out of nursing homes. Covered services should include nursing home care and community care and the capitation rate should include nursing home costs and community service costs. The challenge for rate setting is that the population served must also include both nursing home residents and community residents, two groups that include people with very different levels of need. Government and other funders of research should support larger-scale efforts to establish reliable risk adjustment methods for LTC outside of institutions. The federal government has been encouraged to improve risk adjustment of Medicare payment for frail older persons (MedPAC 1999, p.92), but separate work is needed in this area for long-term care. Existing methods of evaluating case mix in nursing homes, rehabilitation hospitals, and home health care 11

present a strong basis from which to work, but these methods need to be extended to reliably assess health and functional status among people with more modest levels of need. Facing the Challenge of Serving Additional Older Beneficiaries States will face an additional challenge of providing services for the many Medicaid beneficiaries living in the community who could meet the eligibility criteria for nursing home services, but who are not currently receiving waiver services. In some cases, waiver programs have no open slots and so have not been accepting new beneficiaries. In other cases, eligible beneficiaries have simply not applied for services. For states that have substantial numbers of nursing home-eligible Medicaid beneficiaries living in the community and not served by the waiver program, creating a statewide capitated program without slot limits could lead to substantial increases in persons served and in expenditures. It is also possible that the additional people being served would have fewer needs on average than those already being served in the waiver, so that adjustment of rates by functional status would be needed to avoid overpayment. States face a related challenge in serving the many community-based residents who have needs for supportive services but who are not nursing home eligible. Most capitated LTC programs enroll only nursing home eligible beneficiaries, but a few cover non-eligible beneficiaries, as well. The Texas and Minnesota programs allow non-nursing home eligible beneficiaries to enroll, but pay a single rate for all beneficiaries who are not nursing home eligible. This single rate discourages contractors from developing systems of care that would be attractive to those beneficiaries who are at risk of becoming eligible for nursing home services, but who have not yet passed the threshold. Good measures of need can help states to focus services on those who most need them and to adjust capitation rates so that providers face the appropriate incentives. Conclusion: Evaluating Whether to Capitate Long-Term Care As noted in the Introduction, capitated LTC and the HCB waiver programs share important goals: meeting the needs of beneficiaries through a flexible array of community-based services, and delaying or preventing nursing home placement. Before embarking on an ambitious program to capitate LTC services, states should think carefully about what advantages capitation would bring over providing HCB services on a fee-for-service basis. Programs that use capitation to integrate acute and long-term care have clear potential advantages, but are likely to grow very slowly. The argument for capitating only LTC services is less compelling. Capitated LTC may outperform traditional HCB service programs if the capitated organization has more resources available than the traditional case manager or can use the available resources more creatively. Case managers of HCB services typically do not control the delivery of home health care services, while capitated long-term care programs typically do, and may be more effective in using them to support older persons in the community. Under fee for 12

service, Medicaid regulations may limit the effectiveness of public case managers, or service dollars may have been captured by politically influential agencies that are not creative in maintaining older persons in the community. These potential advantages of capitation should be weighed against some disadvantages, including the substantial state administrative effort to establish and manage a capitated program and the concerns that capitation might lead to underservice. To realize the potential of shifting dollars from the nursing home to community-based services, capitated programs will have to overcome the political influence of the nursing home industry and substantive problems of implementation. Finally, creating a capitated program will require managed care organizations to become highly adept at LTC case management, much like the case managers who now work for Medicaid, county governments or area agencies on aging. Capitating LTC may lead to improvements in service organization and effectiveness. But capitation also brings the risk of diminished quality. Without appropriate reimbursement rates that take into account the beneficiary s level of need, capitation may do little to improve quality or restrain growth in LTC expenditures. 13

APPENDIX. CASE EXAMPLES OF CAPITATED PAYMENT SYSTEMS FOR LONG- TERM CARE The following sections describe the efforts of several states and programs to develop capitated financing approaches for LTC services. Tables 3 and 4, which follow these descriptions, summarize the major elements of the programs and their rate setting methods. The Arizona Long Term Care System 7 The Arizona Long Term Care System (ALTCS) was developed as part of the Arizona Health Care Cost Containment System, which was established in 1982. ALTCS began in 1989 and serves both older and younger persons with disabilities who are eligible for Medicaid covered nursing home care. ALTCS contracts for long-term care and Medicaid covered acute care with private contractors in most counties but uses county-operated LTC programs in the two most populous counties, Maricopa and Pima, which contain Phoenix and Tucson. As of June 2000, ALTCS enrolled just under 29,000 elderly and physically disabled (EPD) individuals, the largest capitated LTC program in the country and the only one that serves more than a few thousand beneficiaries. 8 (The program also serves 10,000 enrollees with developmental disabilities, but services and rates for this group, which are overseen by the Arizona Department of Economic Security, are not discussed here.) Among EPD beneficiaries, 82 percent are also eligible for Medicare, which pays for their acute care services, primarily on a fee-for-service basis. The federal government allows ALTCS to limit its coverage of copayments and deductibles to services rendered by providers in the ALTCS network. Beneficiaries thus have a strong financial incentive to join a Medicare HMO but are not required to do so. The Arizona program provides an example of how a single rate paid for nursing home eligible beneficiaries whether they are in a nursing home or in the community can create a strong incentive for plans to provide services that help people stay in the community. Contractors are paid the same rate for each enrolled beneficiary within a county. The rate is calculated in two basic steps. First, ALTCS determines the expected LTC expenditures per month for nursing home residents and for community-based residents. The expected monthly expenditures for nursing home and community-based residents are determined separately by county, primarily by calculating historical costs in the most recent base period and trending these forward. LTC expenditures for nursing home residents are much higher than those for HCB care: in a sample calculation, the state shows nursing home expenditures at $2,800 per month, with long-term care expenditures for community-based residents at $775 per month. Second, ALTCS annually computes a blend of the nursing HCB expenditures. The blend's proportion of community-based expenditures is a target set by ALTCS staff, who typically assume that the contractor can slightly increase the historical share of community-based placements. In addition, contractors are paid for acute care based on historical expenditures 14