Research Issues: Dually Eligible Medicare and Medicaid Beneficiaries, Challenges and Opportunities.

This overview describes the Medicare and Medicaid dually eligible beneficiaries, and it summarizes challenges in obtaining information and designing health care and supportive systems across the continuum of their needs. Some of the challenges include: the complexities of Medicaid eligibility, key structural differences between Medicare and Medicaid, long-standing data limitations, and determining appropriate payment mechanisms and amounts. The overview discusses and highlights changes that are expected to improve the potential for research on dual eligible issues.

Introduction

Individuals who are simultaneously enrolled in both the Medicare and Medicaid programs are commonly referred to as “dually eligible” beneficiaries. Table 1 shows that in 1997, an estimated 6.7 million Medicare beneficiaries received some level of additional benefits through Medicaid “buy-in” at some point during the year. Dually eligible beneficiaries are estimated to represent 17 percent of all Medicare beneficiaries in 1997, and are estimated to account for at least $56.7 billion in expenditures, or 28 percent of total Medicare expenditures. For Medicaid, enrollment and expenditure experience is strikingly similar. Dually eligible beneficiaries are estimated to represent 19 percent of total enrollment and 35 percent of Medicaid expenditures totaling $56 billion, of which more than 50 percent is Federal match to States (Murray and Shatto, 1998; Klemm and Gibson, 1999). In comparison with estimates published in 1995, Medicare dually eligible beneficiary enrollment for 1997 has increased by 700,000 beneficiaries with an additional expenditure of $3.7 billion. Medicaid dually eligible beneficiaries are estimated to have increased by an additional 2 percent of total enrollment, and expenditures have increased by $3 billion (Health Care Financing Administration, 1997).

Table 1

Estimated Medicare and Medicaid Enrollment and Expenditures for Dually Eligible Beneficiaries, 1997

Program	Total	Enrollment		Expenditure
		Non-Dually Eligible Beneficiary	Dually Eligible Beneficiary	Dually Eligible Beneficiary
		Number in Millions		Number in Billions
Medicare1
Total	39.6	232.9	6.7	$56.7
Percent	100	83	17	28
Medicaid3
Total	33.7	427.3	6.4	$56.0
Percent	100	81	19	35.0
Cost
Premium	—	—	—	$4.4
Cost-Sharing	—	—	—	6.6
Other Acute (Drugs)	—	—	—	5.0
Long-Term Care	—	—	—	41.0

Medicare estimates are based on the Medicare Current Beneficiary Survey (MCBS). The MCBS-based estimate of dually eligible beneficiaries represents persons having any form of Medicaid coverage at any point in the year.

Amount represents beneficiaries estimated to be Medicare enrollees without Medicaid eligibility.

Medicaid estimates are based on Federal fiscal year 1997 Health Care Financing Administration 2082 reports actuarially adjusted to represent person years of enrollment and approximates the average monthly or April 1 enrollment.

Amount represents beneficiaries estimated to be Medicaid enrollees without Medicare eligibility.

SOURCE: Clark, W.D. and Hulbert, M.M., Health Care Financing Administration, 1998.

The importance of the dually eligible population is magnified by the fact that the population of Americans 80 years of age or over—those most likely to become dually eligible due to frailties and impairments—is expected to grow by 100 percent for males and 50 percent for females by the year 2025. This will result in even greater health financing expenditures and care challenges (Velkoff and Lawson, 1998). These estimates may illuminate why there appears to be a growing interest by Federal and State governments, foundations, and the broader policy community, in examining the characteristics, needs, utilization, and expenditures for dually eligible beneficiaries.

Definitions of Dual Eligibility

Important factors that State and Federal policymakers face when designing care systems for dually eligible beneficiaries are the statutory definitions of dual eligibility. State Medicaid agencies are required to assist certain low-income Medicare beneficiaries with payment of Medicare premium and cost-sharing provisions. As described by Carpenter (1998), the Medicare buy-in provision has expanded and evolved over time through a series of incremental expansions. For example, a Medicare beneficiary enrolled in Part A and/or Part B, who also receives State support in meeting any Medicare cost-sharing requirements may be considered a dually eligible beneficiary. (For a list of definitions of Medicaid participation in Medicare for dually eligible beneficiaries, refer to the Technical Note.)

New statutory mandates require States to adopt changes in programs and operations, budgetary resources, and information systems. Unfortunately, the multiple levels of State buy-in, and variation among States in actual enrollment patterns, and differences in how States operationalize these eligibility differences within their broader programs leads to tremendous confusion for consumers, providers, regulators, researchers, and policymakers. There has not been a consistent method for defining, reporting, and tracking the various categories of dual eligibility. However, the category of eligibility has significant implications for beneficiary's service delivery options and cost and utilization patterns. For instance, in this issue of the Health Care Financing Review, Parente and Evans describe significant differences in Medicare cost and utilization between those enrolled in the Qualified Medicare Beneficiary (QMB) program and those eligible but not enrolled in the program.

Many of the articles in this issue discuss the difficulties encountered with the lack of precise definitions. The authors who use the MCBS as the basis for their analyses utilize the buy-in data element that contains all categories in which a State participates with cost-sharing (Liu, Long, and Aragon, 1998; Parente and Evans, 1998; Ettner, 1998). Others authors choose to define dually eligible beneficiaries as those for whom States are providing full Medicaid benefits (McCall and Korb, 1998; Saucier et al., 1998). The article by Parente and Evans, examines the differences between all levels of QMB eligibility in the aggregate and other Medicare beneficiaries using reported information from the MCBS. With greater specification of benefit level in the future, more information on important subgroup differences may yield additional explanations of differences in the use and costs of service between dually eligible and Medicare-only beneficiaries. The article by Ettner examines differences in inpatient utilization and expenditures for dually eligible beneficiaries who were hospitalized in 1990 with a primary psychiatric diagnosis. Medicare data was used in this study, but Medicaid information was not available.

Who are the Dually Eligible?

The population of dually eligible beneficiaries is comprised of diverse subgroups. It includes: physically disabled, non-elderly beneficiaries; cognitively impaired, non-elderly disabled Supplemental Security Income (SSI) recipients who become Medicare-eligible after a 24-month waiting period; frail elderly Medicare beneficiaries who have such high medical costs that they “spend-down” their income and assets to become Medicaid eligible; and elderly beneficiaries who are low-income, but are not necessarily frail. Both non-elderly disabled beneficiaries, and Medicare's oldest old-those 85 years of age or over—are more likely to be dually eligible (Murray and Shatto, 1998; Firshein, 1999).

Murray and Shatto present results from the MCBS that show dually eligible beneficiaries are a particularly vulnerable population. By definition, they are poor, are more likely than Medicare-only beneficiaries to be from a minority population, unmarried, live alone, institutionalized, and nearly two-thirds did not graduate from high school. Dually eligible beneficiaries are also more likely to report significant and chronic health problems such as diabetes and heart disease. Significant service delivery and financing challenges arise in serving this population because they are also more likely to suffer from cognitive impairment, mental disorders, and limitations in their ability to perform daily activities, i.e., eating, walking, dressing, etc. (Murray and Shatto, 1998).

Given the health status and demographic characteristics of the dually eligible population, it is not surprising they account for a disproportionate share of both Medicare and Medicaid expenditures. As further illustrated by Murray and Shatto (1998) average Medicare expenditures for dually eligible beneficiaries are 2.4 times those for non-dually eligible Medicare beneficiaries. The analysis of community-dwelling versus institutionalized costs shows that 66 percent of Medicare funding for facility residents was for dually eligible beneficiaries. Twenty-four percent of Medicare dollars spent on community residents were for dually eligible beneficiaries (Murray and Shatto, 1998).

In their analysis of linked Medicare/Medicaid data files from four New England States, Saucier, Bezanson, Booth, Bratesman, Fralich, Gilden, Goldstein, O'Connor, Perrone, and Willrich found that dually eligible beneficiaries used a disproportionate amount of both Medicare and Medicaid resources. They also point out that this utilization pattern was largely driven by the significant subset of those using long-term care (LTC) services.

Dually eligible beneficiaries require a comprehensive set of acute and LTC services. However, they may often be confronted by problems of access, a lack of continuity of care, limited administrative coordination between Medicare and Medicaid, an institutional bias affecting service, and confusion with coverage and payment. While they are most in need of an effective, coordinated system of care, they encounter a fragmented and confusing array of services. Care for dually eligible beneficiaries is fragmented along numerous dimensions including primary versus specialty medical care, acute hospitalization versus chronic care in institutions, medical versus social services, and institutional versus community-based care (Polokoff, 1999). It is further complicated and exacerbated by the separate funding, coverage, and service delivery systems of the Medicare and Medicaid programs. The bifurcation of public financing only reinforces the tendencies of many providers and payers to concentrate on their own particular interests as opposed to the broader interests of the patient and may lead to cost-shifting and movement of patients to justify revenue needs rather than the dictates of beneficiary health and welfare.

Administrative Complexity

The basic dynamics at work in the service delivery environment for dually eligible beneficiaries, the complementary relationship of Medicare and Medicaid, and changes over time, in both volume and type of service, are explained further in articles by Liu et al., 1998; McCall and Korb, 1998; and Saucier et al., 1998. These and other authors point out that administrative efficiencies may be achieved through improved coordination of the two programs (Feder, 1997; Bullen, Perrone, and Parker, 1998; Nemour, 1997). In addition, dually eligible beneficiaries may also be eligible for, and participate in, a variety of services outside both programs. As discussed by Saucier et al., many States sponsor State-only funded services that are designed to serve Medicare beneficiaries at thresholds enabling participation above Medicaid eligibility requirements. Pharmacy benefit programs and home and community-based services funded by the State and the Older Americans Act of 1965 also are important services for dually eligible beneficiaries and complement both Medicare and Medicaid benefits.

Administrative barriers to program coordination within States also exist. Many States have separate agencies serving dually eligible beneficiaries with mental illness, developmental disabilities, and elder services. These agencies may provide home and community-based section 1915C Medicaid services and/or State funded programs for the elderly (Kassner, 1998). However, data from agencies often may have limited detail and be reported on aggregated line items in State Medicaid data systems. There are analogous data limitations within the Medicare program. The hospital inpatient clinical data abstracts used by peer review organizations (PROs), the minimum data set for skilled nursing facilities, and home health data are not part of HCFA's enrollment or claims data bases. Interestingly, both the minimum data set, which is being collected, and home health data, which is proposed to be collected, are reported by States under Federal arrangements. These State and Federal data systems hold information which, if combined with Medicare and Medicaid claims, may offer vast research potential.

A challenge in the future for both Medicare and Medicaid health plan contractors will be to better arrange and/or provide services not presently included in their contracts. There were more than 200,000 dually eligible beneficiaries in Medicare+Choice plans nationwide during 1997, but many of these individuals had to obtain Medicaid through FFS arrangements. Conversely, in those States that enroll dually eligible beneficiaries in Medicaid contracted managed care organizations, many of these organizations may not have contracts for Medicare so that beneficiaries must obtain Medicare services on a FFS basis. Some beneficiaries are believed to be enrolled in separate Medicare and Medicaid health plans. This administrative complexity and confusion among beneficiaries and providers is one of the most significant issues that requires attention in the future as research and demonstration activities attempt to provide greater knowledge and lead to alternative solutions for an aging America.

Changes in Program Services

Oregon has been a leader among the States in the development of a single State agency devoted to the organization and delivery of all LTC services (Coleman, 1998). The State determined early-on that nursing homes should be placements of last resort. Subsequently, it has developed levels of State-licensed services receiving Federal matching funds through Medicaid that demonstrate how States can move LTC delivery systems weighty institutional stays and costs, to a system in which in-home supportive services, foster care, assisted living, and other options are available in lieu of institutionalization. In 1980, the number of Medicaid residents in Oregon nursing homes totaled about 8,000. By 1996, institutional residents had dropped to 7,100. The number of facilities declined during this period. Meanwhile, the population of Oregonians 85 years of age or over increased by more than 40 percent (Coleman, 1998). In 1998, of a total of 15,000 Medicaid beneficiaries, about 5,000 lived in adult foster homes and another 1,000 lived in assisted living. While reducing the rate of cost increases for nursing facilities, this State, and others, increased expenditures for home and community-based services to approximately 50 percent of all of Medicaid LTC expenditures. National spending for home and community-based services more than doubled between FFYs 1992 and 1997 from $5.8 to $13.5 billion (Burwell,1998) representing 24 percent of total Medicaid spending on dually eligible beneficiaries. Oregon and 22 other States that spent more than one-quarter of their Medicaid LTC budget on home and community-based services. States may achieve substantial progress under existing program waiver authorities, should they seek to shift from institutional to community-based LTC systems, without Medicaid section 1115 demonstration waivers. In the four New England States, as described by Saucier et al., note that even though data are not exactly comparable with Burwell's, the authors report that 55 percent of the average per person per month expenditure went to institutional LTC settings and 8 percent went to community LTC services. In general, the use of merged Medicare and Medicaid enrollment and claims files by these authors represents a significant and exciting step that demonstrates the ability of researchers to understand the characteristics of dually eligible populations by combining information from both programs.

Changes in Data

Medicare and Medicaid data are improving over time. In spite of the limitations that will always be apparent in using limited administrative data in attempting to understand beneficiary health status, disease diagnosis, and levels of impairment, data related in some way to payment is believed to be more accurate than some other forms of administrative data. Increasingly, more information is mandated for collection in order to determine payment, whether for hospital, physician services, nursing facilities, or home health care. The Balanced Budget Act (BBA) of 1997 has resulted in significant changes in both Medicare and Medicaid data systems. In addition to moving most remaining cost-based Medicare payment systems to prospective payment, managed care organizations also must meet new standards of encounter reporting, beginning with hospital inpatient information. Payments to Medicare+Choice contractors will be adjusted, based on beneficiary prior hospitalization experience, as reported in these encounters starting in 2000. The BBA required all States to adapt the electronic reporting of the Medicaid 2082 information in the format of the Medicaid Statistical Information System (MSIS), previously a voluntary option of States. It also requires encounter reporting based on the format of MSIS. Given the change in scope of MSIS, HCFA determined that the data elements for the system should be revised prior to the 1999 implementation of MSIS by all States. In addition to collecting more detailed information about buy-in, more information about eligibility for, and participation in, waiver programs is included (Buchanan, 1998). Although some States may be expected to lag behind others in the accuracy of submitted claims, never before have researchers and policymakers had the benefit of a national data base of Medicaid administrative information for research studies. Even with the variation among State programs and use of managed care (with mixed adherence to encounter reporting requirements), this national resource should be available for researchers interested in examining Medicaid and Medicaid/Medicare interactions, particularly with respect to the use and expenditures of dually eligible beneficiaries. Key building blocks for both Medicare and Medicaid are their universal use of the UB-92 inpatient hospitalization form and the HCFA Form 1500 for physician and ambulatory services and procedures.

Administrative data sets also hold promise for merging claims with assessment data such as activities of daily living (ADLs) and instrumental activities of daily living (IADLs) information. Although data sets with functional information are not population-based and, therefore, may include biases for subgroups of the total eligible population, they offer researchers opportunities to link this information with claims data for analytic studies. Of course, the use of claims information also is limited by the lack of information available about individuals who do not utilize and/or underutilize services and the resulting absence of claim information. Three articles in this issue examine aspects of health status and the functional impairment of dually eligible beneficiaries. McCall and Korb analyze risk adjustment payment methodologies that use either diagnosis or level of impairment based on their evaluation of Medicaid program information in Arizona and New Mexico. Pope, Adamache, Khandker, and Walsh (1998) use administrative data on function in their study to compare the predictive power of various risk adjustment methods using MCBS data. Liu, Long, and Aragon also used the MCBS in examining differences between dually eligible beneficiaries and those only with Medicare enrollment, controlling for health and functional status. Their results appear to confirm important differences between the subpopulation of beneficiaries that have dual eligibility from Medicare-only beneficiaries. These studies add to other research in this area by Gruenberg, Kaganova, and Hornbrook (1996).

States appear increasingly interested in merging their home and community-based waiver assessment data with claims information, as discussed by Saucier et al. As researcher interest in comparing function-based and diagnosis-based risk adjustment increases, a variety of Federal and State administrative data may have potential use for supplementing or comparing with survey-based analyses such as the MCBS and the Health Outcome Survey being implemented by Medicare+Choice plans. Use of the nursing home minimum data set, home health data, and, perhaps, clinical abstract data used by PROs in combination with State Medicaid eligibility and claims information could be considered for use in the future as these data become available.

Conclusion

As policy alternatives are considered for the future of the Medicare program and as States continue to develop Medicaid reforms, it will be important to assess the impact that changes in one program have on the other. Changes in Medicare payment and proposals for premium support are likely to have substantial impact on Medicaid resource use. Similarly, continued Medicaid managed care initiatives and increasing use of home and community-based services may be expected to have an effect on Medicare expenditures.

Administrative data systems are evolving towards Medicare/Medicaid compatibility. Researchers are beginning to explore more deeply how beneficiaries use both programs, as research published in this issue of the Review demonstrates and further illustrated by Komisar, Feder, and Gilden (1999). This area of research may be expected to yield valuable information of assistance to policymakers over time as data quality continues to improve due to legislative mandates and further development of payment systems that rely on data accuracy. A particular topic of interest and great difficulty related to Medicaid is the potential impact of medically needy programs on preventing declines into full Medicaid benefit coverage for LTC services. In Medicare, it will be helpful to assess the value of QMB-only, SLMB, and other State buy-in participation in avoiding or deferring the onset of Medicaid eligibility.

These topics may be among the highest priorities on a long list of potential studies that should be performed on merged Medicare and Medicaid information when 1999 data become available nationally, for the first time, and as data quality improves.

Qualified Medicare Beneficiaries (QMBs) Without Other Medicaid (QMB Only)—These individuals are entitled to Medicare Part A, have income of 100 percent Federal poverty level (FPL) or less and resources that do not exceed twice the limit for SSI eligibility, and are not otherwise eligible for full Medicaid. Medicaid pays their Medicare Part A premiums, if any, Medicare Part B premiums, and, to the extent consistent with the Medicaid State plan, Medicare deductibles and coinsurance for Medicare services provided by Medicare providers. Federal financial participation (FFP) equals the Federal medical assistance percentage (FMAP).

QMBs With Full Medicaid (QMB Plus)—These individuals are entitled to Medicare Part A, have income of 100 percent FPL or less and resources that do not exceed twice the limit for SSI eligibility, and are eligible for full Medicaid benefits. Medicaid pays their Medicare Part A premiums, if any, Medicare Part B premiums, and, to the extent consistent with the Medicaid State plan, Medicare deductibles and coinsurance, and provides full Medicaid benefits. FFP equals FMAP.

Specified Low-Income Medicare Beneficiaries (SLMBs) Without Other Medicaid (SLMB Only)—These individuals are entitled to Medicare Part A, have income of 100-120 percent FPL and resources that do not exceed twice the limit for SSI eligibility, and are not otherwise eligible for Medicaid. Medicaid pays their Medicare Part B premiums only. FFP equals FMAP.

SLMBs With Full Medicaid (SLMB Plus)—These individuals are entitled to Medicare Part A, have income of 100-120 percent FPL and resources that do not in exceed twice the limit for SSI eligibility, and are eligible for full Medicaid benefits. Medicaid pays their Medicare Part B premiums and provides full Medicaid benefits. FFP equals FMAP.

Qualified Disabled and Working Individuals (QDWIs)—These individuals lost their Medicare Part A benefits due to their return to work. They are eligible to purchase Medicare Part A benefits, have income of 200 percent FPL or less and resources that do not exceed twice the limit for SSI eligibility, and are not otherwise eligible for Medicaid. Medicaid pays the Medicare Part A premiums only. FFP equals FMAP.

Qualifying Individuals (a) (QI-1s)—This group is effective 1/1/98-12/31/02. There is an annual cap on the amount of money available, which may limit the number of individuals in the group. These individuals are entitled to Medicare Part A, have income of 120-135 percent FPL, resources that do not exceed twice the limit for SSI eligibility, and are not otherwise eligible for Medicaid. Medicaid pays their Medicare Part B premiums only. FFP equals FMAP at 100 percent.

Qualifying Individuals (b) (QI-2s)—This group is effective 1/1/98-12/31/02. There is an annual cap on the amount of money available, which may limit the number of individuals in the group. These individuals are entitled to Medicare Part A, have income of 135-175 percent FPL, resources that do not exceed twice the limit for SSI eligibility, and are not otherwise eligible for Medicaid. Medicaid pays only a portion of their part B premiums ($1.07 in 1998). FFP equals FMAP at 100 percent.

Medicaid-Only Dually Eligible Beneficiaries (Non QMB, SLMB, QDWI, QI-1, or QI-2)—These individuals are entitled to Medicare Part A and/or Part B and are eligible for full Medicaid benefits. They are not eligible for Medicaid as a QMB, SLMB, QDWI, QI-1, or QI-2. Typically, these individuals need to spend down to qualify for Medicaid or fall into a Medicaid eligibility poverty group that exceeds the limits previously stated. Medicaid provides full Medicaid benefits and pays for Medicaid services provided by Medicaid providers, but Medicaid will only pay for services also covered by Medicare if the Medicaid payment rate is higher than the amount paid by Medicare, and, within this limit, will only pay to the extent necessary to pay the beneficiary's Medicare cost-sharing liability. Payment by Medicaid of Medicare Part B premiums is a State option; however, States may not receive FFP for Medicaid services also covered by Medicare Part B for certain individuals who could have been covered under Medicare Part B had they been enrolled. FFP equals FMAP.

Technical Note: Summary of Eight Existing Buy-in Eligibility Categories

More information can be found on the website of HCFA (http://www.hcfa.gov/medicaid/obs4.htm and hcfa.gov/medicaid/bbahmpg.htm