An Investigation of Interstate Variation in Medicaid Long-Term Care Use and Expenditures Across 40 States in 2006

State long-term care (LTC) financing and delivery systems and, in particular, Medicaid funded LTC have long been criticized for being “institutionally biased.” Shifting the balance in publicly-funded long-term care provision away from institutional care (nursing homes, long-term hospitals, intermediate care facilities for the intellectually disabled) toward greater reliance on home and community-based services has been a federal goal for the past three decades -- a goal often referred to as “re-balancing” state LTC systems. This report explores inter-state variations in LTC expenditure and service use patterns, not only in terms of institutional and non-institutional services, but also by Medicaid LTC users’ age and type of disability (e.g., intellectual/developmental disabilities or other working-age adult disabilities). [86 PDF pages]

"

Acknowledgments

The authors gratefully acknowledge the project's Technical Expert Panel members -- Charlie Lakin, Chuck Milligan, and Susan Reinhard -- for their guidance. We also wish to thank Valerie Cheh, Carol Irvin, and Bob Schmitz for providing helpful comments on earlier drafts of this report, Molly and Jim Cameron for editorial assistance, and Susan Moore for secretarial support.

Acronyms

The following acronyms are mentioned in this report and/or appendices. Also see Appendix A for term descriptions.

Executive Summary

State long-term care (LTC) financing and delivery systems and, in particular, Medicaid funded LTC have long been criticized for being "institutionally biased." Shifting the balance in publicly-funded LTC provision away from institutional care (nursing homes, long-term hospitals, intermediate care facilities for people with intellectual or developmental disabilities [ICF/IID]) toward greater reliance on home and community-based services (HCBS) has been a federal goal for the past three decades -- a goal often referred to as "re-balancing" state LTC systems.

This report explores interstate variations in LTC expenditure and service use patterns, not only in terms of institutional and non-institutional services, but also by Medicaid LTC users' age and type of disability (e.g., intellectual and developmental disabilities [ID/DD] or other adult onset disabilities). Some states have re-oriented more toward HCBS than others. It also well known that greater progress has been made in serving certain subgroups within the LTC population in the community (those with ID/DD) compared to others and that reliance on institutional care remains greatest among the elderly, although here again there are interstate variations. This report seeks to quantify the magnitude of such differences.

Interstate variations in reliance on HCBS compared to institutional care are partly a function of some states having committed more strongly to the goal than others, and having accordingly made greater efforts to "re-balance." However, states also experience differential advantages or handicaps that make re-balancing easier or more difficult for some compared to others. The factors that make re-balancing easier or more difficult vary in malleability; that is, the extent to which state policymakers can exercise control over them. For example, states with colder, snowier climates, states with large areas classified as "rural" or "frontier" because of population density, as well as states with disproportionately high low-income aging populations may find it more difficult to "re-balance" because of the logistical challenges of providing primarily home-delivered services under these circumstances. These particular factors are largely outside a state government's ability to change. In contrast, other factors hypothesized to influence re-balancing toward greater reliance on HCBS are at least somewhat under state control. For example, states can use licensing and Certificate of Need legislation to limit nursing home bed supply and enable expansion of alternative services such as assisted living, other forms of residential care, and home health/home care agencies. States can also choose to offer consumer-directed alternatives to "traditional" modes of service delivery such as agency-delivered personal care services.

In this study, we use data from the Medicaid Analytic eXtract, the American Community Survey (ACS), and a variety of data sources describing state characteristics and policies to quantify interstate variations in Medicaid LTC systems performance, and to explore and begin to test hypotheses about the factors that explain greater or lesser use of HCBS across states and subpopulations. Our findings are based on data from 39 states and the District of Columbia, and represent Medicaid service use and expenditures in calendar year 2006.

I. Background and Objectives

Efforts to transform Medicaid long-term care (LTC) from a predominantly institution-based system to one with more community-based services appear, from a national perspective, to have made substantial progress, particularly over the past decade. Since the Supreme Court's 1999 Olmstead v. L.C. decision affirmed the right of persons with disabilities to receive services in the most integrated setting appropriate for their needs (US Supreme Court 1999), Medicaid home and community-based services (HCBS) use and expenditures have more than doubled (Ng et al. 2009) and the use of nursing homes and intermediate care facilities for people with intellectual disabilities (ICFs/IID) has declined substantially (Alecxih 2006; Wiener et al. 2009; Lakin et al. 2009).¹

This overall success, however, masks wide variation in the levels of success across states and different subgroups. Efforts to re-balance LTC systems from their traditional reliance on institutional care to HCBS have been achieved more widely for some populations (young enrollees with disabilities) than others (people over 65) (Wenzlow et al. 2008) and have varied widely across states (Howes 2010; Kassner et al. 2008; Ng et al. 2009). In this report, we explore what factors are linked to successful state outcomes to help identify where new solutions for the remaining institutionalized populations may lie.²

A. Progress in Measuring and Understanding Long-Term Care System Performance

1. Measuring Progress

The desire to transform Medicaid LTC systems has led to the need for meaningful measures of the extent to which state LTC systems have met the aims set forth by Olmstead. In theory, such measures should capture the degree to which people needing LTC services are being served in the most integrated setting appropriate for their needs. In practice, such refined measures are expensive to develop, in part because of the limitations of available data. Furthermore, such measures initially were not needed because more basic measures could quantify progress and meet policy needs. As states continue balancing their systems, policymakers can now benefit from more refined measures to identify areas for program improvement.

The most commonly used indicators of LTC system performance -- the percentage of LTC spending allocated to HCBS and increases in the number of people receiving HCBS -- typically have relied on two readily available and annually updated sources of state-level data on Medicaid expenditures and HCBS use. These include aggregate spending data by service type reported by states in CMS Form 64 (Eiken et al. 2010), and counts of waiver enrollees reported in CMS Form 372 combined with state survey-based counts of personal care and home health users, as summarized each year by the Kaiser Commission on Medicaid and the Uninsured and University of California/San Francisco (Ng et al. 2009). Although these data convey important information on system performance, they cannot be used to conduct subgroup analyses (except for some waiver populations). This limits the ability to measure the extent to which HCBS have reached the people that need them. For this reason, researchers and policymakers have begun using CMS person-level administrative data in the Medicaid Analytic eXtract (MAX) system to explore who is being served by Medicaid LTC and to better understand system transformation.³ These projects include efforts to measure HCBS and institutional use and spending for various subgroups eligible for the Money Follows the Person (MFP) demonstration (Brown et al. 2009; Irvin and Ballou 2009); the AARP state LTC system scorecard that includes some person-based measures of system performance (Reinhard 2010; Reinhard et al. 2011); and CMS's development of a more comprehensive set of indicators of Medicaid re-balancing, which aims to measure LTC system performance and quality (Urdapilleta 2010). Although some reports have been issued, these measures are under development or are being further refined.

These previous and ongoing efforts to develop state LTC systems performance indicators have focused, on the one hand, on what can best be described as summary "outcome" measures of "re-balancing;" and, on the other hand, on codifying expert opinion concerning desirable systems attributes (i.e., "best practices") and scoring state systems accordingly without, however, conducting research to find out whether (which ones and how many) of these best practices predict or correlate with better scores on the summary outcome measures. The present study differs from these others, first, by developing more varied summary outcome measures, including ones for subpopulations, and, second, by looking for descriptive attributes (both the non-malleable or less malleable state systems characteristics that state government has little or no control over and the "policy" variables that they can influence) that correlate with desirable outcome measures.

2. What Do We Know About the Progress in Long-Term Care System Transformation?

As we noted earlier, LTC systems are becoming more balanced in favor of HCBS (Ng et al. 2009; Doty 2010). However, studies have shown substantial interstate and intrastate variations. Some states -- for example, Alaska, New Mexico, Oregon and Washington -- have been identified as successes, whereas others have received low rankings (Howes 2010; Kassner et al. 2008). Within states, performance indicators suggest that HCBS use is much more common among young disabled beneficiaries than older LTC recipients (Wenzlow et al. 2008), but our understanding of how components of the LTC systems function for people with physical disabilities compared to those for people with ID/DD is quite limited.

Factors that may be related to systems judged more successful (because they have been "re-balanced" in favor of HCBS) include single access points; availability of person-centered services; participant involvement; precipitating events or crises; and effective state leadership, planning, and processes (Eiken 2004). In their review of the literature, Mollica and Reinhard (2005) also identified as critical components the availability of broad HCBS, single global LTC budgeting, standardized assessment tools, transition programs, and quality improvement. In a study of MFP grantees, Irvin and Ballou (2009) found two additional features -- the depth of HCBS experience and coverage of optional state plan personal care -- among more balanced systems in terms of LTC spending. A recent survey of state programs confirmed many of these factors as facilitating re-balancing and also highlighted the broad fiscal challenges states face in maintaining and improving LTC systems in hard economic times (Rose et al. 2010).

Many important questions about LTC system performance remain unanswered. Among states identified as successes, are they providing HCBS to more people or are they providing more services? Who remains without access to appropriate HCBS? Can successful policies implemented in some states work for others? How do fiscal constraints and other state characteristics hinder or facilitate system transformation? For example, we would expect that rural states, in which the distance between service providers and recipients is large, may find it more challenging to provide their clients with LTC in home and community-based settings; hence, the lessons learned in more urban states may not apply. Insights into these questions would be particularly helpful to states as they face budget crises and as some consider cuts rather than expansions of Medicaid.

B. Goals of This Study

In this study, we expand on earlier work in Wenzlow et al. (2008) by using MAX 2006 to gain insight into both interstate and intrastate variations in LTC system performance. The study has two broad aims: (1) to characterize differences in LTC systems within and across states; and (2) explore how state constraints and policies might lead to better or worse LTC system performance.

To characterize the performance of LTC systems in each state, we summarize HCBS and institutional care service use and expenditures to determine whether some states are achieving better balanced systems either by serving more people or spending more per person covered compared to other states. We also explore how balance varies across important Medicaid subgroups -- aged enrollees over 65, enrollees under 65 and eligible for Medicaid on the basis of disability, and two subgroups of enrollees with disabilities -- those with physical disabilities and those with ID/DD.⁴ (See Appendix A for a glossary of terms, including the basis of eligibility [BOE] groups.) These subgroups of enrollees tend to have different demographic characteristics and service needs and often are served by different Medicaid programs.

The second portion of our analyses explores how state constraints and policies are associated with the LTC system performance indicators developed in the first portion of the study. Specifically, we examine how factors that may challenge system transformation -- for example, cost of living, fiscal constraints, and state demographics -- and Medicaid and non-Medicaid policies are linked with LTC balance and other indicators of system performance. Finally, we examine how the association between state constraints, policies, and system performance varies across three enrollee subgroups -- the aged, enrollees with physical disabilities, and enrollees with ID/DD.

There is a strong -- virtually universal -- consensus among LTC experts and as well as among federal and state Medicaid officials that state systems should encourage use of HCBS over institutional care. The goal is for HCBS to account for at least half of Medicaid LTC expenditures. As evidence of this consensus: in the 2010 Affordable Care Act, Congress legislated the "Balancing Incentive Payments" (BIP) program that allows states that spent less than 50 percent Medicaid LTC expenditures being spent on HCBS as of 2009 to apply to receive a higher federal match rate to make infrastructure improvements intended to increase their LTC spending on HCBS to at least 50 percent (http://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Long...). As of June 2013, CMS has approved 16 states to receive BIP. If the 50 percent spending standard is met, it logically implies (since institutional care costs more per-capita) that more than 50 percent of Medicaid LTC services users would be receiving HCBS rather than institutional care. The "50 percent" benchmark is, however, admittedly arbitrary. It suggests that the appropriate balance is "equality" whereas, in fact, many experts would like HCBS to become the dominant mode of service provision.

Many LTC experts consider the "oldest-old" (those 85 and older) and LTC service users who lack informal caregivers and must rely largely or exclusively on paid help to be those most likely among Medicaid LTC users with high service needs to require institutional care; that is, those for whom available Medicaid HCBS is least likely to be an adequate alternative to long-stay nursing home placement. With respect to individuals with ID/DD, most of whom are adults under 65, a massive shift occurred during the 1980s and 1990s from large state-run institutions, into private (non-state-operated) smaller institutions and group homes. Currently 14 states have no state-operated ID/DD residential care facilities. In the past decade, there has been a further shift toward family support (providing services or individual budgets) to individuals with ID/DD living with parents or other caregivers and toward out-of-home living arrangements (group homes and supported apartments) where fewer than six individuals with ID/DD share a residence. Braddock (2009) found that 75 percent of all Medicaid and other federal/state funding for ID/DD services went toward non-institutional care (that is, services in settings with fewer than seven residents with ID/DD) and 92 percent of all LTC users with ID/DD in out-of-home placements were in settings with six or fewer residents with ID/DD.

Nevertheless, some states that serve some Medicaid users with ID/DD in residential care settings with 7-16 residents (which, at least by some definitions, qualify as "non-institutional" because of their size, nevertheless certify and pay for care in these settings (including room and board) as small Medicaid ICFs/IID -- which has the effect of blurring the boundaries between Medicaid institutional and non-institutional spending and services use for the ID/DD subpopulation. In marked contrast, however, there are no comparable small residential settings serving the elderly and younger physically disabled adults that may be certified and paid under a special category of "small" Medicaid nursing facilities in some states but not in others. There are comparatively few Medicaid-eligible elders or younger adults with physical disabilities residing in assisted living, adult foster care, or other "out-of-home" residential care settings and any services covered in these settings for Medicaid beneficiaries is always classified as HCBS (room and board costs are ineligible for Medicaid coverage).

Our analyses of state constraints, policies, and LTC system characteristics should be viewed as exploratory. We were unable to assess causal impacts.

C. Summary of Data and Methods

We used MAX 2006 Person Summary (PS) files to develop our measures of LTC system performance. MAX PS files contain demographic and enrollment information for each Medicaid enrollee, as well as information on total Medicaid expenditures for services used during the calendar year, by service type. They also contain information on use of and spending on 1915(c) waiver services -- an important vehicle that states use to provide expanded HCBS to select populations.

We defined HCBS to include services covered under Section 1915(c) waivers and personal care, residential care, home health care, adult day care, and private duty nursing services that are mandatory or provided at state option outside of waiver programs. Institutional care includes nursing home care, ICFS/IID care, inpatient psychiatric services for people under age 21, and psychiatric hospital services for those 65 and older. The MAX PS files cannot be used to differentiate between people using institutional care for long periods and those using Medicaid institutional care for acute events. This study's operational definition of ILTC thus broadly includes all care received in institutions, whether or not a person is using them for LTC.⁵ Moreover, as earlier noted, this analysis was unable to differentiate between Medicaid beneficiaries living in and associated Medicaid spending on certain residential care facilities certified as "small" ICFs/IID (those with no more than 16 residents) and similar settings and their residents in other states licensed as "group homes" (and therefore qualifying for Medicaid HCBS reimbursement only for services and not room and board costs also covered in ICFs/IID).

Our analyses were limited to Medicaid enrollees eligible on the basis of disability or age and who were eligible for full Medicaid benefits in 2006. We excluded Program of All-Inclusive Care for the Elderly (PACE) or other managed LTC enrollees because information on their use of services (HCBS or institutional care) often is missing or unreliable in MAX. We also excluded from the analysis 11 states with MAX fee-for-service (FFS) data that are potentially unrepresentative or unreliable, including Arizona, Maine, Massachusetts, Michigan, Minnesota, Montana, New Hampshire, Oregon, Pennsylvania, Rhode Island, and Texas. Finally, we could not differentiate enrollees with physical disabilities from those with ID/DD in the District of Columbia, Washington, and Wisconsin, and excluded these states from our subgroup analyses. A more detailed discussion of the MAX data, analyzed measures, and methods used is in Appendix B. Appendix C lists state-specific data anomalies.

The analysis of state constraints and policies related to LTC provision relied on a wide range of publicly available data sources. When available, we used data from 2005, 2006, or previous years to capture policies in place and state characteristics at the time that services were being used in 2006. We also used the 2007 American Community Survey's (ACS's) income and disability data to construct estimates of the number of people over 65 or with disabilities potentially eligible for Medicaid in each state.⁶ Medicaid programs vary substantially in terms of the populations they cover. We used ACS-based measures of the size of potential Medicaid-eligible populations (assuming national eligibility criteria) to determine the extent to which cross-state differences in LTC utilization and spending result from state coverage policies.

D. Roadmap to This Report

In the following chapters, we characterize Medicaid LTC system performance (Chapter II) and present the results of our exploratory analysis linking state characteristics and policies with system performance indicators (Chapter III). In both chapters, we present analysis summaries for the overall LTC population, as well as for aged enrollees (over 65), those under 65 with physical disabilities, and people with ID/DD. In Chapter IV, we summarize these results and discuss directions for future research.

II. Variation in Long-term CARE System Performance

Policymakers are interested in learning about states that are making progress in re-balancing their systems, and whether these states are able to transform LTC for important subgroups of the disabled population. In this chapter, we present a summary of the variation in LTC balance and other system indicators across states in 2006, overall, based on their Medicaid eligibility (over 65 or under 65 with disabilities), and for those under 65 with disabilities, by whether or not they used long-term services and supports designed for people with ID/DD.

A. Measures Characterizing Long-Term Care System Performance

No one measure adequately captures LTC system performance in terms of the breadth of the population covered, choice of services, and breadth and intensity of services provided. For this reason, we used a combination of measures to capture variation in system performance across states, including:

• Percentage of Medicaid LTC expenditures allocated to HCBS;
• Percentage of LTC users receiving HCBS;
• Ratio of per-recipient spending on HCBS to spending on institutional care; and
• Percentage of people potentially eligible for Medicaid LTC who used Medicaid HCBS.

The first two measures are commonly agreed-upon indicators of the degree to which states have balanced their LTC systems toward HCBS use and spending relative to institutional care. However, available data sources (e.g., state federal financial participation claims on the CMS "64" forms) have provided data only on spending. These data are available for all states through 2009. Our use of 2006 MAX data makes it possible to obtain (albeit for 37 rather than all 51 states, including the District of Columbia) not only spending data but unduplicated counts of service users and to develop additional measures that require data on expenditures for services, numbers of service users, and age and other characteristics of users of various types of LTC services.

The third measure derived from the MAX files captures the extent to which state spending on HCBS per user is similar to state spending on institutional care. Higher ratios of per-user spending on HCBS relative to per-user spending on institutional care are not necessarily indicative of "better" performance; rather, this measure provides additional insight into how a state is allocating financial resources between HCBS and institutional care. Medicaid law pertaining to HCBS waivers requires such services to be "cost effective" relative to institutional care and defines "cost effectiveness" as spending per HCBS waiver participant that on average does not exceed per-user spending on institutional care that waiver participants would otherwise require. Medicaid regulations further specify that states must spend up to the average amount spent per-capita on institutional care if the state's individualized needs assessment process determines that an HCBS waiver eligible individual requires that level of covered services. Medicaid law and regulations pertaining to other HCBS benefits neither require states to cap individual expenditures relative to institutional costs nor require states to cover the costs of all services assessed as necessary for a given individual (in other words, states may set lower coverage limits so long as these limits apply equally to all beneficiaries).

Some advocates of "re-balancing" argue that public program participants residing in the community who qualify for institutional coverage should be entitled to receive HCBS costing at least as much as what Medicaid would spend on their institutional care -- if a professional assessment indicates that they need care costing that much. Others point out, however, that Medicaid institutional care reimbursements cover room and board costs that account for at least one third of the total and that Medicaid law prohibits HCBS benefits to cover any room and board costs. By this standard, average per-capita HCBS spending per-capita ought not to exceed two thirds of average institutional care spending. At the same time, the Medicaid spending per-user statistics do not reflect the full cost of LTC services -- especially institutional care -- because beneficiaries with Social Security pension/disability benefits or other personal income insofar as Medicaid beneficiaries are required to contribute all but a small personal needs allowance to pay for care and Medicaid. Post-eligibility financial contributions toward the cost of nursing home care are substantially greater than for HCBS or for care in ICFs/IID. Thus, rather than making judgments of state performance based on this indicator, we simply examined whether higher ratios of HCBS to institutional spending per LTC user correlate with the agreed-upon indicators of "re-balancing" toward greater reliance on HCBS.

Many individuals who need human assistance with personal care (basic activities of daily living [ADLs] such as bathing, dressing, transferring from bed to chair, moving from room to room, toileting, and eating), and who have a level of need for such assistance similar to that of nursing home residents, nevertheless continue to reside in the community and to rely exclusively on unpaid help from family, friends, and neighbors. In many cases, such persons are not financially eligible for Medicaid and cannot afford to pay out-of-pocket for personal care services. In principle, low-income persons with personal care needs would be expected to seek Medicaid-covered HCBS rather than rely exclusively on informal help which could impose considerable burden on family members. However, a number of factors may prevent low-income individuals with personal care needs from accessing Medicaid-covered HCBS. These factors include being income-qualified but having assets in excess of the Medicaid allowable level, limits that the state has set on the numbers of qualifying Medicaid beneficiaries who can be served under HCBS waivers, and inadequate supply of HCBS providers to meet demand. To measure the extent to which state residents who potentially qualify for HCBS based on their level of income and need for assistance with personal care who are actually receiving Medicaid HCBS, we report the ratio of HCBS users to the numbers of low-income state residents in need of human assistance with personal care tasks as reported in the ACS. This fourth measure provides policy context for other performance indicators and is assessed in our subgroup analyses.

We examined two additional measures and included them in the Appendix D summary tables:

• Share of total Medicaid expenditures for LTC users spent on enrollees using HCBS.⁷

• Percentage of nursing home and ICFS/IID residents who used HCBS prior to their spell of institutional care (Ballou et al. 2011).

The purpose of the first measure -- the share of total Medicaid expenditures on LTC users spent on HCBS users -- is to adjust the LTC spending share measure for any differences between services captured in our definition of HCBS and institutional care. For example, prescription drugs may be included in nursing home payments, whereas our definition of HCBS excludes such services. Although this measure differs from the LTC spending share measure, the general analysis results were relatively consistent across states and subgroups and so are not presented here. Finally, we include in Appendix D tables (those for the aged and people with ID/DD only), measures developed by Ballou et al. (2013) to capture how often institutional residents had used HCBS prior to entering a nursing home or ICFS/IID. These measures indicate HCBS penetration as part of the continuum of care leading to traditional institutionalization.

B. Interstate Differences

Across the 40 study states, there were about 40 million enrollees eligible for full Medicaid services in 2006. About seven percent were aged or eligible for Medicaid on the basis of disability and used any FFS LTC services -- almost 5 percent used HCBS, and 3 percent used institutional care (Table II.1). (See Appendix Table D.1 for state-level detail.)

Although they represent a small share of enrollees, Medicaid spent $77 billion (about 39 percent of total Medicaid expenditures) on LTC services for these enrollees in 2006 (Table II.2). (See Appendix Table D.2 to view the information from Table II.2, ordered alphabetically by state.) About 41 percent of these expenditures were allocated to HCBS, ranging from 73 percent in Alaska to 11 percent in Mississippi, with a median of 38 percent across states.

As reported in our previous study, the percentage of LTC recipients using HCBS exceeded the percentage of expenditures used for HCBS.⁸ Overall, only 41 percent of LTC expenditures in the 40 states were for HCBS whereas 64 percent of LTC users utilized HCBS.⁹ However, we found wide variation across the states -- 87 percent of the LTC recipients in Alaska used HCBS, compared with just 33 percent of those in Indiana.

Per-user expenditures for HCBS ($17,000) were on average less than half of per-user expenditures for institutional care (46 cents on HCBS for every dollar on institutional care). This ratio also varied substantially by state, with Tennessee spending more per user on HCBS than per user on institutional care ($1.11 for every dollar spent per user of institutional care, or $37,500 per user). At the other extreme, Mississippi spent only 19 cents on HCBS for every dollar spent per user of institutional care. Both these states have fewer numbers of Medicaid HCBS users relative to ILTC users than most other states. Note that some states with particularly high housing costs, such as Alaska and New York, show relatively low ratios of HCBS to institutional care spending even though HCBS spending per-capita is higher than in other states. This likely is due to particularly high room and board costs for institutional care in these states.

These data also demonstrate how a single indicator of system performance could be misleading. State rankings differed substantially across the three measures. For example, both California and New Mexico ranked among the top six states in HCBS as a percentage of LTC spending and users. Yet, California spent only $11,300 per user of HCBS, or about 31 cents for every dollar spent for persons in institutional care, compared with $25,700 per HCBS user in New Mexico -- or 83 cents per dollar of institutional care. Compared with California, which provides at least some (but not necessarily costly) HCBS to a very large number of enrollees, New Mexico serves fewer enrollees, but apparently at a level closer to that of its institutionalized population as a whole. Since California and New Mexico have achieved similar outcomes with respect to overall Medicaid LTC spending on HCBS and percentages of LTC users receiving HCBS, it might be asked: is California's pattern of much lower per-capita spending on HCBS (coupled with somewhat higher per-capita spending on ILTC) a more cost effective way to achieve these overall "re-balancing" results? This is not a question we address in this report, but it is one worth asking and attempting to address in future research. States vary in the financial resources they have available to pay for LTC and, whereas the higher federal match rates available to poorer states help, they do not eliminate, these inequalities. Thus, from a state policymaking perspective, it would be extremely helpful to have a better handle on how much spending per user is "enough."

C. Subgroup Differences

Progress toward re-balancing varies substantially among the different types of users of LTC services. HCBS and ILTC service use and expenditures were far more balanced among young enrollees with disabilities than among enrollees age 65 or older. The percentages of total LTC spending accounted for by HCBS ranged from 66 for people under age 21, 37 for people between ages 65 and 74, and 18 for those age 85 and older (Table II.3). HCBS accounted for an average of 26 percent of LTC spending among all enrollees age 65 and older, compared with 56 percent for those under 65.

Compared to estimates for 2002 reported in Wenzlow et al. (2008), the current results suggest that the spending share for HCBS increased by about five percentage points or more in each age group (under 65, 65-74, 75-84, and 85 or older) since 2002 in the 34 states included in both studies.¹⁰ This suggests that gains have been made in the provision of HCBS to people of all ages over the four-year period.

People with ID/DD are primarily under age 65 and thus make up a large portion of enrollees under 65 with disabilities. In 2006, they accounted for 14 percent of LTC users but 31 percent of Medicaid LTC spending and almost half (46 percent) of HCBS spending. More than 85 percent of these enrollees used HCBS (compared to 64 percent overall and 77 of those with physical disabilities) and HCBS accounted for almost 61 percent of the Medicaid LTC spending used for them (compared to 41 percent overall and 46 percent for those with physical disabilities). Per-user spending on HCBS for people with ID/DD was almost $41,000 in 2006, higher than for any other subgroup shown in Table II.3. However, Medicaid spent only 35 cents per user on HCBS for every dollar spent for persons using costly ICFS/IID care. Other enrollee characteristics associated with use of HCBS included Hispanic ethnicity, male gender, and enrollment in only Medicaid (not Medicare) -- all factors associated with age.

1. Long-Term Care System Performance Indicators by Basis of Eligibility

As in the overall population, system performance varied across states within subgroup. Tables II.4-II.6 summarize, for each of the three indicators, the scores and ranks for the overall population, those 65 and older, and those under 65 eligible on the basis of disability. (See Appendix Table D.2, Table D.3 and Table D.4 to view information in Table II.4, Table II.5 and Table II.6, ordered alphabetically by state.) Each portion of the table is ordered by the overall balance of LTC for that measure across all subgroups.

In some cases, states with the highest HCBS spending or use overall, were those with the highest HCBS or use among subpopulations. Alaska and New Mexico were ranked among the top five for HCBS spending as a percentage of overall HCBS spending overall and in each subgroup (aged, all LTC users under age 65, all ID/DD users under age 65, and non-ID/DD LTC users under age 65). The other top three overall scorers on this measure failed to score in the top five for one or more subpopulations. Vermont and Wyoming, top scorers overall and for LTC users under age 65, ranked 15th and 27th respectively with respect to the percentage of HCBS spending relative to ILTC spending among LTC users age 65 and older. Wisconsin's rank in the top quartile (#10) the expenditure share measure (Table II.4) appears to be driven primarily by its high ranking among young enrollees (#9) since it ranks much lower (#21) on this measure among the elderly. (Data for Wisconsin could not be disaggregated for the under 65 subgroups with ID/DD and physically disabilities.) In contrast, the high ranks of California (#6) and New York (#9) on the overall HCBS relative to ILTC spending and service user ratios appear driven by HCBS use among the aged. The District of Columbia ranked 2nd to last overall despite ranking 9thfor aged enrollees. When measured in terms of the HCBS and institutional care expenditure ratio (Table II.6), Kansas ranked in the top ten overall despite ranking 25th for people over 65 with disabilities. Yet Kansas ranked comparatively highly on the percentage of HCBS/LTC spending on people with disabilities under age 65 (#7) and the percentage of LTC users under age 65 receiving HCBS. Kansas' high rank on the HCBS/ ILTC spending ratio per user overall appears to reflect its comparatively high rank on this measure for the elderly population (#10); yet, here again, whereas Kansas scored high in terms of percentage of total LTC spending on HCBS for the elderly, it scored much lower (#20) on the percentage of elderly LTC users receiving HCBS.

2. Long-Term Care System Performance Indicators by System Type (Aged, ID/DD, and non-ID/DD)

Medicaid beneficiaries using LTC are served by distinct "systems" (program administration and service delivery infrastructures depending on whether their disabilities are ID/DD or not. The administrative and service delivery infrastructures for the elderly and younger disabled individuals (particularly adults under age 65) are the same or closely linked in many states. Both the elderly and younger adults with disabilities other than ID/DD receive ILTC in nursing homes. Whereas skilled home health care delivered by home health agencies (HHAs) may be provided to all subgroups, state plan personal care services are provided almost entirely to the elderly and younger adults without ID/DD, although a small percentage of adults and children with ID/DD may receive this benefit if they also have severe physical disabilities. In some states, the elderly and younger adults with disabilities but without ID/DD are served in the same HCBS waiver programs; however, in a number of states there are separate HCBS waiver programs for people with non-ID/DD according to their age (i.e., either under or over age 60 or age 65).¹¹ Insofar as states have created distinct and separate service systems for subgroups of LTC users, it is not surprising that there is variation across "re-balancing" or other system performance measures by subgroup. Table II.7, Table II.8 and Table II.9 display, for each of the three balance measures, the scores and ranks for each subgroup. These tables replicate those shown earlier by BOE except that, for people under 65, performance indicators are displayed separately for people using ID/DD system services and those not using such services.¹² As above, the states in each table are ordered by the overall balance of LTC for that measure across all subgroups. (See Appendix Table D.5 and Table D.6 for a summary of performance indicators by state for people with physical disabilities and those with ID/DD.)

In almost all states, the lowest HCBS spending and user share is evident among the aged and the highest among Medicaid enrollees using ID/DD services. Among people with ID/DD, either the spending or utilization share allocated to HCBS exceeded that for people with physical disabilities in all but five states (Idaho, Iowa, Mississippi, Missouri, and North Carolina).

Although these results suggest that the Medicaid service system is far more balanced for people with ID/DD than for those who are aged or have physical disabilities, we note several caveats. Some enrollees with ID/DD may be using nursing home or other institutional care services, especially in states where ICFs/IID have closed. We were unable to identify such enrollees using MAX data. Analyses of Nursing Home Minimum Data Set assessment data indicate that only 2.4 percent of NF residents have ID/DD as well as other medical conditions and physical disabilities that justify their nursing home placement (Martin et al. 2011). However, other research evidence indicates that Medicaid beneficiaries who are not living in ICFs/IID but are also not living at home with family members or in supported housing (e.g., shared apartments) are living in "group homes" of limited bed size (16 beds or fewer) where the room and board costs are covered by Social Security and/or Supplemental Security Income (SSI) and related state supplemental cash assistance payments and the service costs are covered under HCBS waiver programs. (O'Keeffe et al. 2010). One definitional issue is that some states certify similar facilities as "small" ICFs/IID so that Medicaid reimbursement will also be available for room and board. To the degree that similar small residential settings for people with ID/DD are Medicaid-covered institutions (ICFs/IID) in some states but classified as non-institutional residential setting ineligible for Medicaid reimbursement for the entirety of their costs in other states, the performance indicators presented here will be biased.

Our measures capture Medicaid system performance among Medicaid enrollees only. To test whether states differ in the extent to which they cover other low-income people with disabilities we employed a combination of MAX and ACS data. The ACS measures age and income in a conceptually clear manner, and low-income was defined to include SSI recipients and other individuals with income up to 300 percent of the SSI limit. However, the ACS disability questions are relatively primitive measures of LTC disability, especially when compared with the detailed assessments that are typically performed to satisfy the medical criteria for Medicaid LTC services. Nevertheless, they provide a consistent measure across states, and are a useful gauge of the extent to which Medicaid programs serve broadly-defined groups of low-income people with disabilities.

We compared Medicaid coverage of the elderly and the non-elderly using two ACS disability questions: (1) Does the person have a physical, mental, or emotional condition lasting six months or more resulting either in difficulties conducting ADLs (dressing, bathing, or getting around inside the home)? and (2) Does the person have a physical, mental, or emotional condition lasting six months or more resulting in difficulties in learning, remembering, or concentrating? To compare the elderly to the non-elderly we considered a person disabled if the response was yes to either question. Our calculations based on combined ACS and MAX data suggest that only 15 percent of people under 65 potentially eligible for Medicaid LTC actually used Medicaid HCBS, compared to 22 percent of potential aged eligibles.¹³ (See Appendix Table D.5 and Table D.3 for state-level detail.) On this indicator as well as the others previously discussed, we see considerable interstate variation in where states rank with respect to the percentage of potential LTC users receiving HCBS by subgroup. California ranks highest overall and for the elderly and is second for LTC users under age 65. The other states in the top five overall include New York, Iowa, Alaska, and Vermont. The other states in the top five for the elderly include Alaska, Washington, New York, and Iowa. New York ranks first for the under 65 population, followed by California, Kansas, Iowa, and Vermont.

For the 37 states where we could separate those with ID/DD from other HCBS users under age 65, we tailored the definition of disability more closely to the characteristics of the population. We considered a person to be disabled with ID/DD if they had difficulties in learning, remembering, or concentrating. We considered a person to have a disability other than ID/DD if they had difficulties conducting ADLs, since most people which did not have ID/DD probably had a physical limitation that restricted performance of ADLs.

On this indicator as well as the others previously discussed, we see considerable interstate variation in where states rank with respect to the percentage of potential LTC users receiving HCBS by subgroup. For people with ID/DD, the top five states had very similar scores. They include Wyoming, South Dakota, Iowa, North Dakota, and New York. For people with physical disabilities, the top five states also had very similar scores. They include Kansas, Alaska, New York, California, and Vermont. (See Appendix Table D.5 and Table D.6).

It is not surprising that the states that serve the highest percent of aged potential eligibles also tend to do well in serving those under 65 with physical disabilities, since state service delivery and program administration infrastructure for LTC tend to be quite different for Medicaid beneficiaries with ID/DD and the aged/disabled (those with disabled conditions other than ID/DD). In effect, state LTC systems for these groups are almost entirely separate. In contrast, the infrastructure serving the elderly and people with non-ID/DD are often combined.

D. Summary of Long-Term Care System Performance Findings

In this chapter, we have examined the differences across states and subgroups using four measures: traditionally computed expenditure-based measures of the balance of institutional care and HCBS (percentage of expenditures for HCBS), a utilization-based measure (percentage of LTC users who used HCBS), a relative per-user expenditure ratio (per-user HCBS expenditures to per-user institutional care expenditures), and a measure capturing the percentage of people potentially eligible for LTC who used services. Because expenditures reflect both the amount of use and the cost of services, and because HCBS are typically less costly than institutionalization, aggregate expenditure comparisons mask key differences in utilization. We found that in 2006, about 41 percent of LTC expenditures paid for persons served were for HCBS; in contrast, almost 64 percent of LTC users used HCBS. Medicaid spent about $17,000 per user for HCBS, or about 46 cents for every dollar for persons in institutional care.

Examination of differences across states illustrate that alternative measures of LTC balance provide different perspectives on LTC utilization and expenditures. For example, of two states with the same percentage of expenditures allocated to HCBS, one may provide limited HCBS to a broad range of users whereas another may provide more expansive services to a small number of HCBS recipients. Eight states ranked in the top ten on at least one measure but among the bottom ten on another. In summary, no one perspective provides a complete picture of the role of HCBS in state Medicaid programs.

Subgroup analyses by state suggest that differences between aged enrollees and those eligible on the basis of their disability or as people with ID/DD, were widespread across the states.

Subgroup analyses also suggest that HCBS use appears to be most common within the Medicaid ID/DD service system, compared with those designed for the aged or people with physical disabilities. However, our estimates indicate that a much smaller percentage of people with ID/DD potentially eligible for Medicaid actually are enrolled and using HCBS. This emphasizes the importance of measuring system performance on multiple dimensions and within different service systems.

III. Correlates of Long-term CARE System Performance

Having found substantial differences in the measures of LTC system balance, the important question is: why do such differences exist? If the differences are due to factors outside of a state's control, they suggest that either new approaches must be developed to address these challenges or that the extent of re-balancing will, under the best of circumstances, be limited in some states or populations. If they are related to factors within the state's control, they could indicate where low-scoring states may focus future efforts. In this chapter, we explore the relationship between potential influencing factors and LTC system performance.

We stress that this cross-sectional analysis is exploratory. The results do not indicate whether there are any causal relationships between state factors and the balance of LTC. However, we expect our results to point to directions for future work.

A. State Constraints and Policy Variables

We differentiate two types of state characteristics likely to affect the role of HCBS in LTC systems: (1) factors over which states have little or no control over such as climate or little or no control in the near term, such as the cost of living; and (2) factors that states could alter, such as their LTC policies, such as whether or not to offer HCBS as an "entitlement" (e.g., state plan personal care services or HCBS waiver services without capped enrollment that requires waiting lists) or policies those that affect the supply of LTC (such as Certificate of Need requirements that restrict nursing home bed supply).

1. Exogenous Factors

Policymakers have always known that states face unique circumstances that make it important to allow for state differences in implementing health care programs. Re-balancing also faces regional challenges, which may result in slow development of HCBS in some communities. Previous studies have cited tight fiscal constraints (Howes 2010; Smith et al. 2009) and access to adequate housing (Denny-Brown and Lipson 2009; Siebenaler 2005) as challenges states face in their efforts to shift their LTC systems from heavy reliance on institutional care to community settings.

Some states face unusual obstacles to re-balancing toward HCBS because of factors beyond policymakers' control. Other states may benefit from atypically favorable conditions. For example, North Dakota is a rural state with a cold, snowy climate, which has also experienced substantial outmigration, especially among the young, over the past several decades. Arguably, the logistics and economics of delivering HCBS to the elderly and disabled living on rural farms without access to much informal support from younger, healthier relatives and neighbors, inevitably limits the extent to which the state's LTC system can be re-balanced toward greater reliance on HCBS. On the other hand, demography (a comparatively younger population), geography, climate, and settlement patterns in Alaska have long favored HCBS. Because many villages (many of which are Native American communities) are largely inaccessible by modern transportation except by air, elderly and disabled individuals may remain in the community longer because their relatives, friends, and neighbors would have great difficulty visiting them if they were placed in specialized residential care facilities that exist only in far away urban centers. A different kind of advantage is enjoyed by states that have large immigrant populations available for low paid, less skilled work; typically, immigrants are drawn to settle in particular geographic areas for reasons unrelated to home care work opportunities. However, once there, immigrant workers become available to work in home care.

To understand how some of these factors may be related to the relative success of re-balancing the LTC system, we investigated the following factors:

• The High Cost of Living in the Community. High costs may make it very difficult for the elderly poor to maintain their residence, whereas admission to a nursing home can relieve those financial burdens.

• Community Financial Resources. A high level of local financial resources may make it feasible for the community to support programs that subsidize utility bills and other living costs, making it less expensive for an individual to remain in the community.

• Environmental Factors. Extreme weather conditions may make it unsafe to live alone or difficult to travel, encouraging more nursing home placements.

• Limited State Resources. States with very limited financial resources may find it difficult to identify resources for use in designing community-based programs or be unwilling to risk developing a new program that may add to Medicaid program costs.

• High Demand for Services. Communities with a high proportion of elderly residents may be more likely to be at the forefront of HCBS because meeting those elderly needs is seen to be a high priority. On the other hand, states with unusually high percentages of low-income elderly, especially in the age 85 and older cohort, may be hard pressed to meet the associated demand for Medicaid-covered LTC services.

• Ability to Provide Care. States with relatively few home care workers or labor shortages may be reluctant to introduce programs that might strain already overtaxed labor markets when nursing homes can serve more residents with fewer workers. There is evidence that Medicaid program coverage rules, in some states, during certain periods of time, authorized more HCBS based on individualized professional needs assessments than home care agency providers were able to deliver because they could not recruit and retain sufficient numbers of frontline workers (Benjamin and Fennell 2007).

Some may question whether or the extent to which some of the factors listed above are truly outside of state control. For example, there is considerable debate over whether and how state policymakers could reduce or eliminate home care worker shortages. Some experts say they could do so by providing home care workers who provide Medicaid HCBS with better pay and benefits or by giving the job enhanced status via training and credentialing requirements. Some also argue that unionization of home care workers (extensive in some states but sparse or non-existent in most) will result in improved pay, benefits, and more training for home care workers. State policies may affect the ability of home care workers to unionize. It has also been suggested that state policymakers can alleviate home care worker shortages by adopting policies that allow Medicaid beneficiaries to hire individual aides, including family members, friends, and neighbors (often referred to as offering options for "consumer-directed" services). The argument is that consumer-directed services options expand the labor pool because that pool is no longer restricted to individuals who are interested in becoming employees of home care agencies.

Measuring these factors and finding data that can support an analysis of their relationships to re-balancing efforts is challenging. Table III.1 lists the factors we were able to measure for this study. The table also lists the measures we used as indicators of constraints, their sources, and their hypothesized relationships with the degree of HCBS provision in a state.

2. State Policy Variables

Numerous Medicaid policies could potentially increase the use of HCBS. States have the options to provide personal care and expanded home health services under their Medicaid state plans and to waive certain Medicaid regulations to cover HCBS for select subpopulations under 1915(c) waivers. Other state policies, such as nursing home regulations and SSI supplements that support independent living, may influence the use of HCBS (Irvin and Ballou 2010, Ng et al. 2009). Under the Deficit Reduction Act of 2005, states have even more options to provide HCBS via state plans through 1915(i) and 1915(j) waivers, although few changes had been implemented by 2006.

To understand how some of these policies may be related to the relative success of re-balancing the LTC system, we investigated the following factors:

• Consumer Direction Options. Consumer direction of personal care services has been shown to improve client satisfaction with services. States that adopt this option may have more residents interested in using HCBS.

• Financial and Functional Eligibility Options. States may develop lenient financial or functional eligibility rules to encourage the use of HCBS. Although lenient rules may increase overall spending on and use of HCBS, they also may result in lower spending per HCBS user if the resulting population using Medicaid LTC has fewer service needs.

• State Plan Coverage. States may offer personal care services under their state plans, eliminating the need for the individual to be covered by a waiver program to receive HCBS, where enrollment can be limited.

• Residential Care Coverage. States that support residential placements other than traditional institutions, such as assisted living facilities, may have more enrollees who can avail themselves of HCBS.

• Number of Waiver Program Enrollees. States that set a relatively high level for HCBS waiver enrollment will have fewer people on waiting lists and provide more HCBS.

• SSI Supplements to Support Independent Living. States that supplement federal SSI payments for people living in the community at a higher level than those in Medicaid facilities may encourage the disabled poor to remain in the community.

• Institutional Supply Policies. States that limit the number of institutions or are actively closing institutions (such as recent trends to reduce the number of ICFs/IID) will increase their need to use HCBS.

• Nursing Home Policies. States that enable nursing homes to hold rooms and receive payment for those enrollees who take short leaves either to hospitals or home may encourage nursing home rather than HCBS use. By allowing people short home stays, residents and families may feel less need to enroll in an HCBS program.

• Payment Policies. Policies that encourage the supply of HCBS in a state -- such as higher rates for such services -- may increase the number of HCBS providers who also could provide care to Medicaid recipients. However, policies that pay nursing homes more may encourage the growth of that industry, thus increasing the use of nursing home services.

Table III.2 lists the state policies that may affect HCBS use, how we measured them, and how they may be related to Medicaid HCBS use.

B. Factors and Policies Associated with Long-Term Care System Performance

To shed light on the relationship between the state factors and the LTC balance, we summarized outcomes for the top and bottom ten states based on their LTC system performance scores and measured the association between the factors and each LTC balance measures. We tested the associations for statistical significance to help differentiate state policies and factors very likely to be related to the balance measures. We again stress that these associations do not imply causation. For brevity, we present the detailed information using the first measure -- the association between the state factors and the percentage of LTC expenditures for HCBS in that state -- and summarize the results for the other measures.

1. Associations between State Factors and Long-Term Care System Performance Indicators

Few state factors are significantly associated with the percentage of LTC expenditures for HCBS (Table III.3). We found two statistically significant associations. One was the average winter precipitation in the state between 1971 and 2000, which was negatively correlated with spending on HCBS -- that is, the more that it rained or snowed in winter months, the lower the re-balancing measures. The second was for our measure of workforce availability: personal and home care aides per 1,000 elderly or persons with a disability. This association suggests that in states where there is greater availability of personal and home health aides, there is a greater level of HCBS expenditures, as would be expected. Note that this workforce measure was available only for 2009, three years after our balance indicator was measured, and so is likely to be capturing demand for this type of personnel.

The associations between constraints and balance were similar in direction and significance for the other balance measures, except that when balance was measured as the percentage of LTC users or potential users receiving HCBS, only the workforce measure (not the winter precipitation measure) was significantly different from zero. Taken together, these results suggest that the exogenous factors we examined are not substantially linked to levels of LTC system balance across states for the LTC Medicaid population as a whole.

2. Associations between State Policy and Supply-Side Variables and Long-Term Care System Performance Indicators

For policy measures, we found that about half of the measures were related to our LTC system performance indicators (Table III.4). However, we found that none of the measured policy and supply-side factors showed the same relationship across all four indicators. Three policy measures and one supply-side factor, however, showed a consistent pattern using both the spending share measure and percentage of LTC recipients using HCBS. These three policy measures were: (1) the availability of consumer direction options; (2) coverage of state plan personal care; and (3) the availability of higher state SSI supplements for people living in the community than for those using Medicaid facility care. These measures were associated with HCBS use at the state level. In addition, the number of nursing home beds per elderly in the state in 2003 was associated with lower levels of balance.

3. Subgroup Differences

Policymakers also are interested in the progress of particular subgroups. Because the elderly recipients generally make up a large portion of those in the LTC system, they will dominate the overall results. One question, however, is whether state factors are linked to LTC system performance for those who are under age 65 and have physical disabilities, or for those with ID/DD.

The results of our analyses suggest that state factors may function differently for recipients with ID/DD than for other Medicaid LTC recipients, as can be seen in Table III.5.

• Although local financial resources were not associated with HCBS re-balancing in the overall population, per-capita personal income and taxable resources per-capita were positively associated with HCBS expenditures as a percentage of all LTC expenditures for people with ID/DD.

• Consumer direction was significantly related to the balance of expenditures for all subgroups except people with ID/DD.

• Having personal care state plan services as well as residential services covered under Medicaid was associated with LTC spending balance overall and for the aged, but not for people with disabilities or ID/DD.

C. Summary of Findings on the Relationship Among the State Factors, Policy Variables, and Long-Term Care System Performance

Although exploratory, the associations between LTC balance measures and state factors and policy variables presented in this chapter indicate several areas that may warrant further research. First, of seven measures selected to capture state characteristics, only two were significantly correlated with measures reflecting HCBS penetration in state LTC systems overall: (1) average winter precipitation, which was negatively associated with LTC expenditure balance; and (2) personal and home care aides per 1,000 persons who are elderly or have disabilities, which was positively associated with both HCBS use and spending.

We hypothesize that winter precipitation may hinder enrollee or provider transport, making it more challenging for states in which winter snow or poor weather conditions occur to provide care in some community settings. We also hypothesize that worker shortages may reduce HCBS expansions because nursing home care requires fewer staff. This suggests that weather conditions and workforce availability would be important contextual variables to consider when developing more refined measures of LTC system transformation progress.

When examining policy variables, we note that the different measures of LTC system performance sometimes produced different results, highlighting the need to develop measures that carefully reflect states' progress, as CMS is doing. The three policy variables most consistently related to systems more balanced toward HCBS were consumer direction, coverage of personal care, and SSI supplements for people living in the community. However, having personal care was associated with HCBS penetration only for aged enrollees, not for enrollees under 65 with a physical disability and people with ID/DD. Other factors, such as financial resources appear to be related to LTC balance for those with ID/DD.

IV. Summary of Results and Directions for Future Research

This study examined Medicaid FFS HCBS use and spending across 39 states and the District of Columbia using 2006 MAX data. The study expanded on LTC balance analyses presented in Wenzlow et al. (2008), which was based on 2002 data, to summarize additional long-term balance measures, assess balance for additional populations (people with ID/DD), and explore associations between state factors and policies and LTC balance.

A. Summary of Results

Important Dimensions of Long-Term Care System Performance. HCBS spending as a percentage of LTC spending is the most commonly used measure of LTC system transformation. In this study, we examined differences across states and subgroups between this traditionally computed expenditure-based measure and several additional systems performance measures that may be relevant to a discussion of the relative balance between HCBS and ILTC: a utilization-based measure (percentage of LTC users who used HCBS), a relative per-user expenditure ratio (per-user HCBS spending to per-user institutional care spending), and a measure capturing whether Medicaid HCBS are reaching individuals that may need them (percentage of potential Medicaid LTC eligibles who used HCBS). Examination of differences across states on these measures illustrates that alternative indicators of LTC system performance provide different insights and an interpretive context for cross-state comparisons of the percent of total Medicaid LTC spending on HCBS.

There is high overlap between the highest scoring states on two "balance" measures: the percent LTC spending on HCBS overall and percent LTC users receiving HCBS overall (eight states are in the top ten on both these measures). However, there is considerably less overlap between the top scoring states on these two indictors and those in the top ten with respect to ratio of HCBS to ILTC spending per user overall. Only four of the ten states with the highest ratios of per-user HCBS to ILTC spending are in the top ten with respect to the percent of LTC spending on HCBS and only two are among the ten states with the highest percent of potential LTC users receiving HCBS.

Of the states that scored highest (top five) with respect to potential LTC users receiving HCBS, two scored in the top five on the percent of Medicaid LTC spending going toward HCBS and an additional two scored in the top ten. Similarly, of the states that scored among the top five in providing HCBS to potential LTC users, all but one were among the top ten highest scorers with respect to percent of LTC users receiving HCBS and the one state not among the top ten ranked number 11.

The patterns evident in states rankings on a range of performance indicators tell a story about how state LTC policies differ with respect to priorities and trade-offs. Some states choose to provide HCBS to large numbers of eligible and potentially eligible LTC users but they spend comparatively low amounts per HCBS user. Indeed, they may have comparatively fewer nursing home residents but spend a great deal more per user on ILTC than per user of HCBS. This pattern has certain logic if state policymakers have reason to believe that the comparatively small number of ILTC users are, on average, much more severely disabled than HCBS users. In contrast, some states have comparatively large numbers of institutional residents (perhaps because they over-invested in nursing home bed capacity compared to other states several decades ago) and provide comparatively fewer LTC users with HCBS but HCBS benefits are comparatively generous (closer to the average amount spent per user on ILTC). Again, there is a certain logic to this approach if HCBS is being targeted toward beneficiaries who are considered to be at high risk of nursing home admission and in need of generous benefits in order for HCBS to serve as an effective substitute for nursing home care.

Subgroup Differences in Balance. Our review of LTC balance measures by subgroup indicated that differences between aged and disabled enrollees, as well as people with ID/DD, were widespread across the states. Although our previous study based on 2002 MAX files (Wenzlow, Schmitz, and Shepperson,2008) concluded that states with LTC systems most balanced toward HCBS among all LTC users were those with services most balanced among the aged, we found more exceptions to this pattern in 2006.

The subgroup analyses indicate that it has proved much more difficult for states to overcome the "institutional bias" with respect to LTC spending and services for the elderly than for LTC users under age 65, especially those with ID/DD. Only a handful of states among the 40 included in this study stand out as having been much more successful than others in re-balancing their LTC systems for the elderly. These states are Alaska, New Mexico, California, and Washington. These four states all spend 46 percent or more of their Medicaid LTC dollars for the elderly on HCBS. In addition, they all provide HCBS to 63 percent or more of elderly LTC users. With the exception of Alaska (an exceptionally high cost state), all of these states spend less per elderly ILTC user than the national average. New Mexico and Washington spend slightly more per elderly HCBS user than the national average California is unique in also spending less per elderly HCBS user than the national average. California also has a ratio of spending per elderly HCBS user to spending per elderly ILTC user lower than the national average, whereas the ratio of per-user spending on HCBS for the elderly relative to per-user spending on ILTC is higher than the national average in the other three states. Washington -- but especially California and Alaska -- all serve far higher percentages of potential elderly HCBS users than the national average, whereas this is not the case for New Mexico. Of these four states, California appears to have arrived at the most cost effective formula for re-balancing Medicaid LTC spending and use patterns among the elderly toward HCBS while also providing HCBS to more potential users.

The states that stand out as having maximized spending for and use of HCBS among the under 65 Medicaid LTC population are, in addition to Alaska and New Mexico (the states that rank highly on re-balancing measures for all populations), Vermont, Colorado, Hawaii, and Wyoming. These states all spend more than 90 percent of their Medicaid LTC dollars for the ID/DD subgroup on HCBS and all of them. Alaska, Colorado, and Vermont have fewer than 2 percent of LTC users with ID/DD in institutional care. Alaska and Vermont now spend more per user on HCBS than on ILTC (although this indicator may no longer mean much since these two states have virtually eliminated ILTC for people with ID/DD). All of these states -- with the exception of Colorado -- spend well above the national average on HCBS per user with ID/DD. Colorado spends slightly above the national average on HCBS per user with ID/DD. Wyoming provides HCBS to 20.8 percent of potential users with ID/DD (compared to a national average of 8.7 percent); Hawaii, Colorado, Vermont, and New Mexico all exceed the national average with respect to providing HCBS to potential users with ID/DD. However, Alaska falls below the national average in providing HCBS to potential users with ID/DD. It should perhaps also be noted that quite a few states (California, Connecticut, North Dakota, Nebraska, Iowa, Maryland, New York, and South Dakota) provide HCBS to percentages of potentially eligible users with ID/DD well above the national average, but do not perform as well as Alaska, Hawaii, Colorado, Vermont and New Mexico with respect to the percent of spending on HCBS or percent of users receiving HCBS measures for people with ID/DD. Looking at the pattern of rankings across indicators, Colorado, among the 40 states in our study, appears to be doing the best job of "re-balancing" its LTC system for people with ID/DD toward reliance on HCBS rather than ILTC.

For LTC users under age 65 with physical disabilities, Alaska and New Mexico, again rank in the top five with respect to LTC spending on HCBS and LTC users receiving HCBS. However, the other states that appear to do the best job of promoting access to HCBS over ILTC for the non-elderly with physical disabilities are Kansas, Missouri, and Vermont for the percent of LTC spending on HCBS measure. North Carolina, California, and Kansas perform best on the percent of LTC users receiving HCBS measure. In addition to Alaska (but not New Mexico), the states that rank highest in providing HCBS to potential LTC users under age 65 with physical disabilities are Kansas, New York, California, and Vermont. Three states that rank among the top five with respect to the ratio of per-user spending on HCBS to per-user spending also rank in the top five on one or more of the other performance indicators (Missouri, Kansas, and Vermont) whereas two do not (Tennessee and Indiana). However, California, Kansas, and Vermont rank, respectively, below, similar to, or only slightly above the national average in per-user spending on HCBS for the under 65 population with physical disabilities. In Kansas, Vermont, and Missouri, the ratios of HCBS to ILTC spending per user for this population is comparatively high because these states spend considerably under the national average on ILTC spending per user for the ILTC population. Taking all systems performance measures into consideration, Kansas seems to be doing the best overall job of re-orienting its Medicaid LTC system away from ILTC toward HCBS for low-income persons under age 65 with physical disabilities.

The only "state policy" variable measured in this study that predicted high ranking with respect to the percent of LTC spending dedicated to HCBS for the under 65 population with physical disabilities was availability of consumer direction. Kansas has a long-standing tradition of using Independent Living Centers to facilitate consumer-directed personal care services (funded by HCBS waivers). All of the other states that ranked in the top five on performance indicators for this subpopulation (except the ratio of HCBS spending per user to ILTC spending per user) are states where consumer direction is widely available and, in California, Vermont, Alaska, and Missouri, the use of consumer-directed services, especially among the subgroup of Medicaid LTC beneficiaries under 65 with physical disabilities, was much more prevalent in 2006 than in the nation as a whole (Sciegaj and Selkow 2011). We identified the most significant differences in measures of LTC balance by population age group and service delivery system. A comparison of our results with those reported in Wenzlow et al. (2008) for 34 states included in both studies indicated that HCBS spending as a share of all LTC spending increased by at least five percentage points in all age groups between 2002 and 2006 for those enrolled in Medicaid. That is, system transformations over those years appear to have increased use of HCBS across all ages. Our analysis also shows that, although Medicaid systems appear to be least balanced toward HCBS for the aged and most balanced for enrollees with ID/DD, a relatively small share of those with ID/DD potentially eligible for Medicaid LTC services actually receive them. Monitoring the larger population with ID/DD and their needs will be critical for better understanding Medicaid system performance.

Correlates of Long-Term Care System Performance. We conducted an exploratory analysis of the bivariate association between constraints and LTC policies and three indicators of system performance. Our results suggest that two types of state factors are associated with systems less balanced toward HCBS: (1) poor weather conditions that may make it more challenging to serve enrollees with LTC needs in their homes; and (2) the size of the workforce needed to provide adequate HCBS. We also found that availability of consumer-directed services and personal care coverage were positively associated with HCBS use and expenditures, but not for those enrollees with ID/DD. Other factors, not subject to much state control appear to be related to the progress in re-balancing LTC for this population, most notably the availability of resources (per-capita income and availability of taxable resources). Finally, state SSI supplements for people living in the community were associated with more balanced systems among the aged and people with ID/DD. We cannot infer causal relationships from these findings, but rather note that they point the way toward possibly fruitful work in the future.

Our analyses of factors that predicted higher ratios of LTC spending on HCBS for the ID/DD population was particularly weak and failed to yield much useful information with respect to state policy variables, except for a negative relationship between nursing home bed supply and higher percentages of LTC spending for the ID/DD population on HCBS. Future research will need to explore indicators that are more specific to the ID/DD service delivery system, which, as earlier noted, tends to be quite separate from the services system for the elderly and younger adults (under age 65) with physical disabilities. One promising avenue is to look at the percentages of publicly-funded ID/DD beneficiaries receiving services in any institutions with 16 or more beds, in state-run ICFs/IID, and in residential settings with six or fewer residents. It is noteworthy, that of 37 states in our study for which it was possible to measure the percentage of total Medicaid LTC spending for the ID/DD population spent on HCBS, five had closed all of their public ICFs/IID prior to 2006. Four of these states were among the top five highest ranking states with respect to the percent of LTC spending on HCBS for the ID/DD population. Colorado (ranking #3) was the only state that had not closed all of its public ICFs/IID and Colorado had only 2.3 percent of persons with ID/DD receiving public benefits residing in public ICFs/IID (compared to a national average of 8 percent. Moreover, Colorado had no beneficiaries with ID/DD residing in private ICFs/IID, whereas nationally 14.4 percent of all beneficiaries with ID/DD resided in either public or private ICFs/IID in 2006 (RTCL/UMN 2012). In all these states 90 percent or more of public beneficiaries with ID/DD in out-of-home placements resided in settings with fewer than six or more residents with ID/DD. In contrast, the states that ranked lowest in terms of percent Medicaid LTC spending on HCBS for the ID/DD population (Mississippi, Louisiana, Illinois, Arkansas, and North Carolina) all had percentages of beneficiaries with ID/DD residing in state-run institutions in excess of the national average. For example, Mississippi had 40.3 percent of state residents with ID/DD receiving public benefits in state-run ICFs/IID (all with 16 or more beds) in 2006 (RTCL/UMN 2012).

B. Directions for Future Research

The exploratory findings presented here suggest several directions for future research.

Level of Need and the Distribution of Care Received. The utilization and spending per-user patterns reported here suggest substantial differences across states in the populations served and/or service levels provided. Looking at the distribution of spending for both HCBS and institutional care services would provide insight into whether low-spending (or high-spending) states are providing the same level of care to all of their enrollees or serving a wide range of needs. To further our understanding of whether LTC systems are meeting the requirements set forth by Olmstead even further, LTC balance analyses should move toward examining the needs of enrollees, appropriate settings that can support those needs, and whether services received are indeed provided in the most integrated appropriate settings.

Addressing the Continuum of Care in Measures of Long-Term Care Balance. As in past studies, we differentiated HCBS from institutional care to study LTC balance. However, HCBS includes a range of residential settings, such as assisted living, and institutional care can include smaller ICFs/IID more similar to group homes than traditional state institutions. Future research will need to address the true continuum of LTC settings to better understand Medicaid LTC system transformation.

Environmental Barriers to HCBS Use and System Transformation. Although our analyses were exploratory, we identified significant associations between LTC balance and winter precipitation and availability of care providers. This suggests that environmental barriers may need to be taken into account when measuring progress toward system transformation.

Constraints as Mediators of Long-Term Care Policy. Our preliminary analysis identified significant bivariate relationships between state-level contextual factors, such as constraints and policies, and LTC system performance. These cross-sectional comparisons could be supplemented usefully with a more extensive study of the multivariate relationships across measures. Of particular interest is how the association between policies and balance may differ across groups of states experiencing similar fiscal, environmental, and demographic characteristics. A longitudinal study assessing the effects of select policies would be an important extension of this work.

References

Alecxih, Lisa. "Nursing Home Use by the ‘Oldest Old' Sharply Declines." Report to the National Press Club, Lewin: November 21, 2006. Retrieved from http://www.lewin.com/content/publications/NursingHomeUseTrendsPaperRev.pdf on July 7, 2010.

Ballou, Jeffrey, Valerie Cheh, Dean Miller, and Audra Wenzlow. "Medicaid-Financed Institutional Stays: Characteristics of Nursing Home and ICFS/IID Residents and Their Patterns of Care." Report for the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Cambridge, MA: Mathematica Policy Research, 2013.

Benjamin, A.E. and Mary L. Fennell. "Putting the Consumer First: An Introduction and Overview." Health Services Research, Volume 42, Issue 1p2, February 2007, pp 353-361.

Braddock, David, "State of the States in Intellectual Disabilities/Developmental Disabilties: 2009." Coleman Institute, University of Colorado, available at http://sos.arielmis.net/.

Brown, Randall, Carol Irvin, Debra Lipson, Sam Simon, and Audra Wenzlow. "Research Design Report for the Evaluation of the Money Follows the Person (MFP) Grant Program." Report for the Centers for Medicare and Medicaid Services. Cambridge, MA: Mathematica Policy Research, October 3, 2008. Retrieved from http://www.mathematica-mpr.com/publications/pdfs/MFP_designrpt.pdf on November 28, 2009.

Bruininks, R., S.Y. Byun, K. Alba, K.C. Lakin, S. Larson, R.W. Prouty, and A. Webster. "Residential Services for Persons with Developmental Disabilities: Status and Trends Through 2006." Minneapolis, MN: University of Minnesota, Research and Training Center on Community Living, Institute on Community Integration, August 2007. Retrieved from http://rtc.umn.edu/docs/risp2006.pdf on July 19, 2010.

Bureau of Economic Analysis (BEA), U.S. Department of Commerce; Internal Revenue Service and Office of Economic Policy, U.S. Department of Treasury. "Total Taxable Resources." September 30, 2008. Retrieved from http://www.treasury.gov/resource-center/economic-policy/taxable-resources/Documents/2008est.pdf on January 4, 2011.

Bureau of Economic Analysis and Bureau of the Census, U.S. Department of Commerce. "Table 1. Per Capita Personal Income, Personal Income, and Population, by State and Region, 2005–2006." Retrieved from http://www.bea.gov/newsreleases/regional/spi/2007/xls/spi0307.xls on July 7, 2010.

Bureau of Labor Statistics (BLS), U.S. Department of Labor. Occupational Employment Statistics (OES). Home Health Aides (SOC Code 311011) and Personal and Home Care Aides (SOC Code 399021). Retrieved from http://www.bls.gov/oes/ on July 20, 2010.

Bureau of the Census, U.S. Department of Commerce. "Historical Census of Housing Tables: Homeownership by Selected Demographic and Housing Characteristics." Last revised on December 2, 2004. Retrieved from

Denny-Brown, Noelle, and Debra J. Lipson. "Early Implementation Experiences of State MFP Programs." Reports from the Field: The National Evaluation of the Money Follows the Person Demonstration Grant Program, no. 3. Princeton, NJ: Mathematica Policy Research, November 2009. Available at https://www.cms.gov/CommunityServices/Downloads/MFPReportNo3Nov09.pdf.

Doty, Pamela. "The Evolving Balance of Formal and Information, Institutional and Non-Institutional Long-Term Care for Older Americans: A Thirty-Year Perspective." In Advancing Home and Community-Based Services: Transforming Policies, Programs, and Service Delivery in Long-Term Care, Public Policy and Aging Report, vol. 20, no. 1, Winter/Spring, 2010. Retrieved from http://www.hcbs.org/files/193/9616/PP&AR.pdf on December 15, 2010.

Eiken, Steve. "Promising Practices in Long-Term Care Systems Reform: Common Factors of Systems Change." Report for the Centers for Medicare and Medicaid Services. Washington, DC: Medstat, November 9, 2004. Retrieved from https://www.cms.gov/PromisingPractices/Downloads/commonfactors.pdf on December 13, 2010.

Eiken, Steve, Kate Sredl, Brian Burwell, and Lisa Gold. "Medicaid Long-Term Care Expenditures in FY 2009." Report prepared by Thomson Reuters, August 17, 2010.

Federal Housing Finance Agency. "Table 713. Single-Family Housing Price Indexes by State." Housing Price Index, 4th quarter, 2008. Retrieved from http://www.census.gov/compendia/statab/cats/prices/consumer_price_indexes_cost_of_living_index.html on July 7, 2010.

Harrington, C., and T. Ng. "Medicaid LTC and HCBS: Trends in Programs and Policies. 2011." Available at http://www.pascenter.org, accessed January 25, 2012.

Houser, Ari, Wendy Fox-Grage, and Mary Jo Gibson. "Across the States: Profiles of Long-Term Care and Independent Living." Washington, DC: AARP, 2009. Retrieved from http://assets.aarp.org/rgcenter/il/d19105_2008_ats_1.pdf on December 8, 2010.

Howes, Candace. "The Best and Worst State Practices in Medicaid Long-Term Care." Direct Care Alliance Policy Brief, No. 3, April 2010. Retrieved from http://www.hcbs.org/files/180/8982/HowesMedicaidPolicyBrief.pdf on November 18, 2010.

Intrator, Orna, Mark Schleinitz, David C. Grabowski, Jacqueline Zinn, and Vincent Mor. "Maintaining Continuity of Care for Nursing Home Residents: Effect of States' Medicaid Bed-Hold Policies and Reimbursement Rates." Health Services Research, vol. 44, no. 1, February 2009.

Irvin, Carol V., and Jeffrey Ballou. "The Starting Point: The Balance of State Long-Term Care Systems Before the Implementation of the Money Follows the Person Demonstration." Reports from the Field: The National Evaluation of the Money Follows the Person Demonstration Grant Program, no. 1. Princeton, NJ: Mathematica Policy Research, January 2009. Retrieved from https://www.cms.gov/CommunityServices/Downloads/MFPfieldrpt4.pdf on July 10, 2010.

Irvin, Carol V., Debra Lipson, Sam Simon, Audra Wenzlow, and Jeffrey Ballou. "Money Follows the Person 2009 Annual Evaluation Report." Final report submitted to Centers for Medicare and Medicaid Services, Division of Advocacy and Special Initiatives Disabled and Elderly Health Programs Group. Princeton, NJ: Mathematica Policy Research, September 2010.

Kaiser Commission on Medicaid and the Uninsured. "Recent Growth in Medicaid Home and Community-based Services Waivers." April 2004. Washington, DC: Henry J. Kaiser Foundation. Retrieved from http://www.avalerehealth.net/research/docs/medicaid_hcbs.pdf on January 24, 2012.

Kassner, E., S. Reinhard, W. Fox-Grage, A. Houser, and J. Accius with B. Coleman and D. Milne. "A Balancing Act: State Long-Term Care Reform." Washington, DC: AARP Public Policy Institute, July 2008. Retrieved from http://assets.aarp.org/rgcenter/il/2008_10_ltc.pdf on December 14, 2010.

Kitchener, M., T. Ng, C. Harrington, and M. O'Malley. "Medicaid Home and Community Based Service Programs: Data Update." Kaiser Family Foundation, Issue Paper, December 2007. Table 9. Retrieved from http://www.kff.org/medicaid/upload/7720.pdf on July 20, 2010.

Lakin, K. Charlie, Sheryl Larson, Pat Salmi, and Naomi Scott. "Residential Services for Persons with Developmental Disabilities: Status and Trends Through 2008." Minneapolis, MN: University of Minnesota, Research and Training Center on Community Living, Institute on Community Integration, 2009. Retrieved from http://rtc.umn.edu/docs/risp2008.pdf on December 9, 2010.

Martin, Linda, Brant G. Fries, John P. Hirdes, and May James. "Using the RUG-III classification system for understanding the resource intensity of persons with intellectual disability residing in nursing homes." Journal of Intellectual Disabilities, June 2011, vol. 15, no. 2: 131-141.

Mollica, Robert L., and Susan C. Reinhard. "Rebalancing State Long-Term Care Systems." Ethics, Law, and Aging Review, vol. 11, 2005, pp. 23-41.

Mollica, R., K. Sims-Kastelein, and J. O'Keeffe. "Residential Care and Assisted Living Compendium: 2007." Report U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Washington, DC: November 30, 2007. Retrieved from http://aspe.hhs.gov/daltcp/reports/2007/07alcom1.htm#table1-7 on July 19, 2010.

National Oceanic and Atmospheric Administration (NOAA). "Area-Weighted Seasonal Normals, 1971-2000 (and Previous Periods)." In State, Regional, and National Seasonal Temperature and Precipitation, Historical Climatography Series No. 4-3, August 2002. Retrieved from http://cdo.ncdc.noaa.gov/climatenormals/hcs/HCS_43.pdf on July 20, 2010.

Ng, T., C. Harrington, and M. O'Malley Watts. "Medicaid Home and Community-Based Service Programs: Data Update." Kaiser Family Foundation, Issue Paper. Available at http://www.kff.org/medicaid/upload/7720-03.pdf on December 2009.

Ng, T., A. Wong, and C. Harrington. "Home and Community-Based Services: Introduction to Olmstead Lawsuits and Olmstead Plans." San Francisco, CA: PAS Center at the University of California/San Francisco. Updated August 2011. Retrieved from http://www.pascenter.org/olmstead/ on January 24, 2011.

O'Keeffe, Janet, Paul Saucier, Beth Jackson, Robin Cooper, Ernest McKenney, Suzanne Crisp, and Charles Moseley. "Understanding Medicaid Home and Community-Based Services: A Primer, 2010 Edition." Chapter 5: Providing Medicaid Services in Community Residential Settings. Washington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Available at http://aspe.hhs.gov/daltcp/reports/2010/primer10.htm#chap5.

Reinhard, Susan, Enid Kassner, Ari Houser, and Robert Mollica. "Raising Expectations: A State Scorecard on Long-Term Services and Supports for Older Adults, People with Physical Disabilities, and Family Caregivers." AARP Public Policy Institute. Available at http://www.aarp.org/relationships/caregiving/info-09-2011/ltss-scorecard... on September 2011.

Reinhard, Susan. "State LTSS Scorecards." Presentation at the CMS Balancing of Long-Term Care: Long-Term Support System Research and Data Summit, AARP Public Policy Institute, Baltimore, MD, October 28, 2010.

Research and Training Center on Community Living (RTCL), University of Michigan (UMN). National Residential Information Services Project, State Tables for 2006. Available at http://www.rtc.umn.edu/risp/build/index.asp, accessed on January 30, 2012.

Rose, Miriam, Farida K. Ejaz, Linda S. Noelker, and Melissa Castora-Binkley. "Recent Findings on Home and Community-Based Services Across States." In Advancing Home and Community-Based Services: Transforming Policies, Programs, and Service Delivery in Long-Term Care, Public Policy and Aging Report, vol. 20, no. 1, Winter/Spring, 2010. Retrieved from http://www.hcbs.org/files/193/9616/PP&AR.pdf on December 15, 2010.

Sciegaj, Mark, and Isaac Selkow. 2011. "Growth and Prevalence of Particpant Direction: Findings from a National Survey of Publicly-funded Participant Directed Services Programs." 2011 Financial Management Services Conference sponsored by the National Resource Center for Participant-Directed Services. Retrieved from http://web.bc.edu/libtools/details.php?entryid=340&page=1&topics=&types=... on June 14, 2013.

Siebenaler, K., J. O'Keeffe, D. Brown, and C. O'Keeffe. "Nursing Facility Transition Initiatives of the Fiscal Year 2001 and 2002 Grantees: Progress and Challenges, Final Report." Research Triangle Park, NC: RTI International, 2005. Available at http://www.hcbs.org/files/74/3656/NFT_final_web.pdf.

Smith, Vernon K., Kathleen Gifford, Eileen Ellis, Robin Rudowitz, Molly O'Malley Watts, and Caryn Marks. "The Crunch Continues: Medicaid Spending, Coverage and Policy in the Midst of a Recession. Results from a 50-State Medicaid Budget Survey for State Fiscal years 2009 and 2010." Kaiser Commission on Medicaid and the Uninsured, September 2009. Available at http://www.kff.org/medicaid/upload/7985.pdf.

Social Security Administration (SSA). "State Assistance Programs for SSI Recipients, January 2006." Office of Research, Evaluation, and Statistics, September 2006. Retrieved from http://www.ssa.gov/policy/docs/progdesc/ssi_st_asst/2006/ssi_st_asst06.pdf on March 18, 2011.

Thomson Reuters. "Medicaid Expenditures for Long-Term Services and Supports: 2011 Update." Cambridge, MA: Thomson Reuters, October 11, 2011. Retrieved from http://www.avalerehealth.net/research/docs/medicaid_hcbs.pdf on January 24, 2012.

Urdapilleta, Oswaldo. "National Balancing Indicator Project (NBIP) Overview." Presentation at the CMS Balancing of Long-Term Care: Long-Term Support System Research and Data Summit, IMPAQ International, Baltimore, MD, October 28, 2010.

U.S. Supreme Court. Olmstead v. L. C. (98-536) 527 U.S. 581. Available at http://www.law.cornell.edu/supct/pdf/98-536P.ZO, 1999.

Wenzlow, Audra T., Robert Schmitz, and Kathy Shepperson. "A Profile of Medicaid Institutional and Community-Based Long-Term Care Service Use and Expenditures Among the Aged and Disabled Using MAX 2002." Report for the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Cambridge, MA: Mathematica Policy Research, January 2008. Retrieved from http://www.mathematica-mpr.com/publications/PDFs/medicaidprofile.pdf on December 8, 2010. Available at http://aspe.hhs.gov/daltcp/reports/2008/profileMAX.htm.

Wiener, Joshua M., Wayne L. Anderson, and David Brown. "Why Are Nursing Home Utilization Rates Declining?" Research Triangle Park, NC: RTI International, 2009. Retrieved from https://www.hcbs.org/files/160/7990/SCGNursing.pdf on December 8, 2010.

Notes

1. The Federal Government's commitment to encouraging and assisting state re-balancing efforts can be traced back at least as far as the enactment of the Medicaid 1915(c) HCBS legislative authority in 1981. Nevertheless, "re-balancing" toward HCBS took place only gradually over the following two decades. Considerably more progress was made "de-institutionalizing" and "diverting" from institutional placement children and adults with developmental disabilities as compared to the aged/disabled, especially the elderly in need of considerable hands on human assistance with personal care tasks and/or severe dementia. In 1992, HCBS accounted for only 15 percent of all Medicaid spending on LTC services (Kaiser Commission on Medicaid and the Uninsured 2004). In 1994, the Clinton Administration adopted policy changes that made it easier for states to obtain federal approval to expand the numbers of Medicaid beneficiaries served under 1915(c) HCBS waivers without having to show that they had or planned to reduce institutional bed supply by corresponding numbers (a requirement referred to as the "cold bed rule" that was particularly difficult to meet with respect to nursing home beds since these facilities, unlike ICFs/IID whose residents had intellectual or developmental disabilities (ID/DD), were not predominantly state facilities serving Medicaid beneficiaries exclusively). From 1995 through 2009, the percentage of total Medicaid spending on LTC going toward HCBS has increased by 1-3 percent annually (Thomson Reuters, 2011).

   In 1999, the Supreme Court issued its landmark ruling in the Olmstead v L.C. case that the Americans with Disabilities Act required states generally (the plaintiff State of Georgia in particular) to make all reasonable efforts to meet the LTC needs of citizens receiving or at risk of requiring publicly-funded institutional care in the community instead (Ng, Wong, and Harrington 2011). President Bush launched a cross-departmental "New Freedom" Initiative to ensure federal programmatic and regulatory compliance with the ruling, and Congress funded a Real Choice/Systems Change grant program to help states develop the infrastructure to comply with the spirit as well as the letter of the Olmstead decision. Between FY 2001 and FY 2010, the Centers for Medicare and Medicaid Services (CMS) awarded almost $289 million in Real Choice/Systems Change grants to help states develop the infrastructure to expand Medicaid beneficiaries' access to HCBS alternatives to institutional LTC (ILTC) (http://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Long...).

2. In this report, the use of the word "states" encompasses the 50 states and the District of Columbia.

3. This study's predecessor summarized the strengths and limitations of MAX data for studying LTC (Wenzlow et al. 2008), finding that although the MAX 2002 data were still incomplete for some states, and service-specific information on HCBS was not yet reliable, MAX can be a useful tool in gaining a better understanding of which populations are receiving HCBS.

4. Some Medicaid enrollees (an unknown number) have disabilities but are not identified as eligible on the basis of disability. We expect this number to be small. However, to the degree that such individuals exist in our study states and differ from persons in our sample, the results presented here will be biased.

5. The PS files do not contain information on the timing or length of institutional stays. MAX claims, which were not available for this study, are needed for such analyses.

6. We used 2007 rather than 2006 ACS because aged enrollees may be underestimated in earlier rounds of the survey's data.

7. For people dually eligible for Medicare and Medicaid, Medicare covers inpatient and other acute services. Because institutionalized enrollees are more likely to be dually eligible for Medicare and Medicaid (Wenzlow et al. 2008), spending on people using HCBS as a percentage of total Medicaid spending for LTC users should be interpreted with caution.

8. Note that in any given year, an individual can receive both HCBS and institutional care.

9. In the 34 states included in both the present study and Wenzlow et al. (2008), HCBS as a percentage of LTC expenditures increased from about 34-40 percent and use of service increased from 59 percent to 64 percent between 2002 and 2006. The share of expenditures increased in all states except Idaho, but the rate of use decreased in a handful of states.

10. Breakdowns for those under 65 by age were not included in Wenzlow et al. (2008).

11. Unique Medicaid programs and services also are available for people with mental illness. However, many people with mental illness use health care services for short durations rather than for LTC, and we were unable to identify long-term mental health care in MAX uniquely. In this study, we thus group people with mental illness by age with enrollees who are aged or have physical disabilities.

12. We identified but did not separately report on the approximately 20,000 individuals over 65 using either ID/DD waiver or ICFS/IID services in the 37 states.

13. The ACS disability questions are relatively primitive measures of LTC disability, especially when compared with the detailed assessments that are typically performed to satisfy the medical criteria for Medicaid LTC services. Nevertheless, they provide a consistent measure across states, and are a useful gauge of the extent to which Medicaid programs serve broadly-defined groups of low-income people with disabilities

14. Section 1915(c) services are identified by program type codes 6 and 7 in MAX. Section 1915(c) (program type 7) of the Social Security Act applies to Medicaid enrollees who otherwise would require Medicaid-covered hospital, nursing facility, or ICF/IID care. Section 1915(d) (program type 6) applies specifically to individuals over age 65 requiring such a level of care. Most states do not differentiate between the two program types in MSIS and report all waiver services under one or the other program code. As suggested in MAX documentation, we sum expenditures reported under the two program codes for our analysis.

15. Expenditures for any institutional or community-based LTC services provided under managed care are subsumed into managed care premiums. Services covered under managed care (including any for LTC) generally cannot be identified in MAX as they are reported in "encounter records," which are known to be incomplete in MSIS and MAX.

16. In this report, the use of the word "states" encompasses the 50 states and the District of Columbia.

17. We excluded people reported to be eligible only for family planning services, unqualified aliens eligible only for emergency services, and restricted-benefit duals receiving coverage only for Medicare premiums and cost sharing.

18. We used 2007 rather than 2006 ACS data for the study because estimates of aged enrollees were unreliable in earlier years of the survey.

19. The ACS also has information about conditions that substantially limit a person's ability to walk, climb stairs, reach, lift, or carry and whether a physical mental, or emotional condition lasting six months or more makes it difficult for a person to go outside alone to shop or visit a doctor's office. However, most states measure functional eligibility based on difficulties with ADLs. Therefore we used only the ADL measure in our definition of functional eligibility.

Appendices

Appendix A. Glossary of Terms

This glossary summarizes the operational definitions of terms used in this report. For more general definitions of Medicaid terms, see Schneider et al. (2002).

Age: Age is defined as of December 31, 2006.

Adult (BOE Group): A BOE group that includes pregnant women and caretaker relatives in families with dependent children. (Adults who are eligible for Medicaid due to disability are coded as disabled.)

Aged (BOE Group): A BOE group that includes enrollees age 65 or older who qualify for Medicaid due to their age. Because some states code all people over 65 as aged, enrollees older than 65 but categorized in another BOE group in MAX were recoded as aged for this study.

Basis of Eligibility (BOE): Eligibility grouping that traditionally has been used by CMS to classify enrollees as children, adults, aged, or disabled.

Child (BOE Group): A BOE group that includes persons under age 18 or under age 21 in states electing to cover older children. (Children who are eligible for Medicaid due to disability are coded as disabled.)

Disabled (BOE Group): A BOE group that includes persons of any age (including children) who are unable to engage in substantial gainful activity by reason of any medically determinable physical or mental impairment that can be expected to result in death or that has lasted or can be expected to last for a continuous period of not less than 12 months. Because disabled people over 65 are often but not always categorized as aged, all disabled people over 65 were recoded as aged in this study.

Fee-For-Service (FFS): A payment mechanism in which payment is made for each utilized service. FFS services exclude services provided under capitated arrangements.

Home and Community-Based Services (HCBS): Services covered under Section 1915(c) waivers and personal care, residential care, home health care, adult day care, and private duty nursing services that are mandatory or provided at state option.

Home Health: Services provided at a patient's place of residence (typically a patient's home), in compliance with a physician's written plan of care that is reviewed every 62 days. These include nursing services, as defined in the State Nurse Practice Act, home health aide services, physical therapy, occupational therapy or speech pathology, and audiology services that are provided by a HHA or a facility licensed by the state to provide these medical rehabilitation services.

Institutional Long-Term Care (ILTC): Nursing facility services, services provided in ICFs/IID, mental hospital services for people over age 65, and inpatient psychiatric facility services for individuals under age 21.

Intermediate Care Facility for People with Intellectual Disabilities (ICFs/IID): ICFs/IID are Medicaid-financed facilities for the care of individuals with ID/DD. These institutions are an optional Medicaid benefit that states may choose to offer; they are required to have four or more beds and offer treatment or rehabilitative services to people with ID/DD.

Managed Care: Payment mechanism used to manage health care, including services provided by health maintenance organizations (HMOs), PACE, prepaid health plans, and primary care case management plans. Services provided under managed care plans are not included in the measures summarized in this report.

Personal Care: Personal services, such as bathing and toileting, sometimes expanded to include light housekeeping furnished to an individual who is not an inpatient or a resident of a group home, assisted living facility, or long-term facility, such as a hospital, nursing facility, ICF/IID, or institution for mental disease. Personal care services are those that individuals typically would accomplish themselves if they did not have a disability.

Private Duty Nursing: Services, except those for mental health or substance abuse treatment, provided by registered nurses or licensed practical nurses under direction of a physician to recipients in their own homes, hospitals, or nursing facilities, as specified by the state.

Residential Care: Although room and board services provided in residential care facilities are not covered by Medicaid, other components of residential care -- for example, personal care, 24-hour services, and chore services -- can be covered. Residential care includes group, family, or individual home residential care; cluster residential care; and therapeutic residential care services, assisted living, supported living, and night supervision.

Program of All-Inclusive Care for the Elderly (PACE): A managed care plan that coordinates both acute and LTC for eligible enrollees (those 55 and older, living in a PACE area, and otherwise eligible for nursing home care). A capitated payment mechanism is used for PACE plan enrollees. As a result, service-specific information is not available for services provided under PACE or other managed plans.

Restricted-Benefit Enrollees: Enrollees who receive limited Medicaid coverage, including unqualified aliens eligible only for emergency benefits, Qualified Medicare Beneficiaries, and people eligible for only family planning services. Some enrollees may be eligible for a restricted set of services but are coded as full-benefit enrollees -- for example, those eligible for prescription drug coverage and Medicare cost sharing only.

Waiver: Services provided under Section 1915(c) of the Social Security Act that enable states to provide Medicaid-financed community-based LTC for people who otherwise would require Medicaid-covered hospital care, nursing facility care, or care in an ICF/IID. These programs can be designed to target individuals in specific age groups and with specific conditions, and the services can be restricted to certain areas of the state. (Other types of Medicaid waivers -- for example, 1115 waivers that cover population subgroups not generally covered under Medicaid, or those that fundamentally change service delivery -- are not discussed in this report.)

Appendix B. Data and Methods

The indicators of LTC system performance presented in this report are based on data from the 2006 MAX PS files. In addition, we used the ACS and a variety of publicly available data sources to develop indicators of state constraints and policies associated with Medicaid LTC. In this appendix, we describe the MAX and ACS data, their strengths and limitations, and the methods we used to develop variables and conduct our analysis.

Appendix C. State Long-term CARE Data Anomalies

MAX data contain a variety of anomalies, many of which are specific to individual states. The anomalies most likely to impact this report's analyses of 2006 MAX are listed in Table C.1, by state. A full list of anomalies is available from the CMS website at https://www.cms.gov/MedicaidDataSourcesGenInfo/07_MAXGeneralInformation.asp.

As a result of these anomalies, we excluded 11 states from all analyses: Maine, because complete 2006 data were not available; Arizona, because it provides most long-term services via managed care arrangements, whereas this study analyzes services provided on an FFS basis; Minnesota, because the state was transitioning many of its LTC recipients to managed care during the study period; New Hampshire, because its expenditure data for HCBS relative to ILTC were believed to be unreliable when compared with data from CMS Form 64; and Massachusetts, Michigan, Montana, Oregon, Pennsylvania, Rhode Island, and Texas because their HCBS data potentially were unreliable. In addition, we excluded the District of Columbia, Washington, and Wisconsin from analyses of populations with physical disabilities or ID/DD because waiver data used to identify these populations were incomplete in these states.

Appendix D. Supplementary Data Tables