Skip to main content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Interstate Variation and Progress Toward Balance in Use of and Expenditure for Long-Term Services and Supports in 2009

Publication Date
Rosemary Borck, Victoria Peebles, Dean Miller and Robert Schmitz
Mathematica Policy Research
 
Abstract

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 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 and/or developmental disabilities or other working-age adult disabilities).

DISCLAIMER: The opinions and views expressed in this report are those of the authors. They do not necessarily reflect the views of the Department of Health and Human Services, the contractor or any other funding organization.


MEASURING STATES' PROGRESS IN MAINTAINING AND EXPANDING MEDICAID HOME AND COMMUNITY-BASED SERVICES

This report was prepared under contract #HHSP23320095642WC between the U.S. Department of Health and Human Services (HHS), Office of Disability, Aging and Long-Term Care Policy (DALTCP) and Mathematica Policy Research. For additional information about this subject, you can visit the DALTCP home page at http://aspe.hhs.gov/office_specific/daltcp.cfm or contact the ASPE Project Officers, John Drabek and Pamela Doty, at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, S.W., Washington, D.C. 20201. Their e-mail addresses are: John.Drabek@hhs.gov and Pamela.Doty@hhs.gov.

"

Acronyms

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

Appendix
ACAAffordable Care Act
ACSAmerican Community Survey
ADAAmericans with Disabilities Act
ADRCAging and Disability Resource Center
 
BEABureau of Economic Analysis
(within the U.S. Department of Commerce)
BIPBalancing Incentive Payments
BLSBureau of Labor Statistics
(within the U.S. Department of Labor)
BOEBasis of Eligibility
 
CMSHHS Centers for Medicare and Medicaid Services
 
DRADeficit Reduction Act
 
FFSFee-For-Service
FYFiscal Year
 
HCBSHome and Community-Based Services
HHAHome Health Agency
HHSU.S. Department of Health and Human Services
 
ID/DDIntellectual and/or Developmental Disabilities
ICF/IIDIntermediate Care Facility for People with Intellectual Disabilities  
ICF/MF  Intermediate Care Facility for the Mentally Retarded
ILTCInstitutional Long-Term Care
 
KFFKaiser Family Foundation
 
LTCLong-Term Care
LTSSLong-Term Services and Supports
 
MAXMedicaid Analytic eXtract
MFPMoney Follows the Person
MSISMedicaid Statistical Information System
 
NFNursing Facility
NOAANational Oceanic and Atmospheric Administration
(within the U.S. Department of Commerce)
 
OAAOld Age Assistance
OESOccupational Employment Statistics
 
PACEProgram of All-Inclusive Care for the Elderly
PSMAX Person Summary
 
SEPSingle Entry Point
SSISupplemental Security Income

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 disabilities [ICFs/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 such as intellectual and/or developmental disabilities (ID/DD) compared to 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; for example, those with ID/DD compared to adults whose physical and/or cognitive disabilities began after reach age 18 but before turning 65. Moreover, 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, 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 37 states and the District of Columbia and represent Medicaid service use and expenditures in calendar year 2009.

Key Findings on Variation in Medicaid Long-Term Care System Performance in 2006

  • Across the 38 study states in 2009, about 45 percent of Medicaid LTC spending was for HCBS in 2009, while almost 67 percent of Medicaid LTC users used HCBS. Medicaid spent about $19,500 per user for HCBS, or 48 cents per user of HCBS for every dollar on people in institutional care. However, there is considerable variation, across states and population subgroups.

  • Most states reported modest progress on re-balancing toward HCBS from 2006 to 2009.

  • Taken together, two measures (the percentage of long-term services and supports (LTSS) expenditures for HCBS and the percentage of LTSS users receiving HCBS) identify a few states that appear to have the highest levels of balance in the breadth and depth of their LTSS. These states, which include Alaska, California, Colorado, Vermont, and Washington, ranked highly on both measures for most or all subpopulations of enrollees.

  • Throughout the rankings, however, a number of states achieved a notably higher ranking on one measure than on the other. For these states, alternative measures of the LTSS system provide different perspectives on LTSS utilization and expenditures. For example, for two states with the same percentage of expenditures allocated to HCBS, one may provide limited HCBS to a broad range of users, and the other may provide more expansive services to a small number of HCBS recipients. Thus, assessing multiple measures continues to provide a more complete picture of the role of HCBS in state Medicaid programs than any single measure alone.

  • Subgroup analyses by state suggest that differences in HCBS use and expenditures between aged enrollees and those eligible on the basis of disability remained widespread across the states. As we found in the previous study, several states achieved overall balance by serving a relatively large number of aged people (e.g., the District of Columbia and New York), but most did so by providing more HCBS to younger enrollees with disabilities (e.g., New Hampshire, Vermont, and Wyoming), particularly people with ID/DD, and ranked relatively low for the aged. This suggests that, even in states that rank near the top on overall balance toward HCBS, there may be room for further re-balancing for some services or subpopulations.

  • Subgroup analyses also suggest that HCBS use continues to be most common within the Medicaid ID/DD service system, compared to systems designed for the aged or people with physical disabilities. This differential emphasizes the importance of measuring system performance on multiple dimensions and within different service systems.

  • Looking at population subgroups, about 65 percent of their HCBS LTC spending went for those with ID/DD, compared with 49 percent adults under age 65 with other disabilities, and 30 percent for LTC recipients over 65. About 86 percent of Medicaid enrollees using LTC services for ID/DD received HCBS, compared with 78 percent of those with other disabilities under age 65, and 55 percent for users over 65.

  • Several states that have achieved much better than average HCBS coverage for one or more population groups deserve further study so that other states may learn from their experience. Specifically:

    • Overall, Washington, Alaska, Vermont, California, and Colorado had the highest percent of Medicaid LTSS expenditures going for HCBS (75 percent to 58 percent, in declining order). Alaska, California, Washington, Idaho and Iowa had the highest percent of Medicaid LTSS users receiving HCBS (90 percent to 75 percent). New Hampshire, Washington, Indiana, Utah, and Wyoming had the highest per-user spending for HCBS, relative to per-user spending for institutional care (102 percent to 74 percent).
    • In terms of serving the aged, Washington, Alaska, California, New York and the District of Columbia had the highest percent of Medicaid LTSS expenditures going for HCBS (59 percent to 38 percent, in declining order). Alaska, California, Washington, Idaho and Iowa had the highest percent of Medicaid LTSS users receiving HCBS (86 percent to 63 percent). Louisiana, New York, Washington, Indiana, and New Hampshire had the highest per-user spending for HCBS, relative to per-user spending for institutional care (77 percent to 56 percent).
    • In terms of serving those under age 65 with disabilities other than ID/DD, Kansas, Alaska, Colorado, North Carolina, and California had the highest percent of Medicaid LTSS expenditures going for HCBS (75 percent to 63 percent, in declining order). California, North Carolina, Virginia, and Alabama had the highest percent of Medicaid LTSS users receiving HCBS (90 percent to 88 percent). Kansas, Indiana, Ohio, and Texas had the highest per-user spending for HCBS, relative to per-user spending for institutional care (64 percent to 58 percent).
    • In terms of serving those under age 65 with ID/DD, New Hampshire, Alaska, Maryland, Colorado, and Wyoming had the highest percent of Medicaid LTSS expenditures going for HCBS (99 percent to 91 percent, in declining order). New Hampshire, Alaska, Colorado, Maryland, and Kansas had the highest percent of Medicaid LTSS users receiving HCBS (100 percent to 98 percent). New Hampshire, Alaska, Oklahoma, Indiana, and Utah had the highest per-user spending for HCBS, relative to per-user spending for institutional care (144 percent to 57 percent).
  • If small ICFs/IID (having fewer than six beds) were considered to provide HCBS rather than institutional services, then the percent of Medicaid LTSS expenditures for ID/DD that went towards HCBS would increase from 65 percent to 68 percent, the percent of users would increase from 86 percent to 89 percent, and the per-user expenditure of HCBS relative to institutional care would increase from 33 percent to 36 percent.

  • States with relatively high rates of HCBS spending overall did not always have consistently high rates of HCBS spending for all subgroups of enrollees. For example, the relatively high overall rankings for New Hampshire, Vermont, and Wyoming on the expenditure share measure appear to be driven primarily by higher rates of expenditures on enrollees under age 65 with disabilities and lower spending on aged enrollees. In comparison, the high ranks of the District of Columbia and New York appear driven by high rates of HCBS use and per-user spending among the aged.

Key Findings on Associations between State Constraints, Policies, and Long-Term Care System Performance

We examined the correlations between LTSS balance measures and state characteristics and policy variables. The associations found indicate several relationships that appear relevant for understanding variations in HCBS use and expenditures across states.

  • Of the several measures selected to capture exogenous state characteristics, only two were significantly correlated with measures reflecting HCBS penetration in state LTSS systems overall: (1) personal and home care aides per 1,000 elderly and younger persons reporting disability, which was positively associated with HCBS spending and use; and (2) percentage of potential Medicaid eligibles age 75 or older, which was negatively associated with HCBS spending and use.

  • We hypothesize that the relationship between home care workers and HCBS may be the result of several factors. HCBS may expand when there are home care workers available to serve more people in residential settings. Conversely, communities with very high levels of demand for these services may find that there are insufficient community resources, including care workers, available to serve everyone in the community, or the increased demand for these workers may drive an increase in their supply.

  • The three policy variables most consistently related to LTSS systems with higher rates of HCBS use were consumer-direction, percentage of out-of-home placements in facilities with six or fewer residents, and availability of assisted living and residential care units. These factors may be important contextual variables to consider when assessing LTSS balance.

  • Some of the relationships were only significant for some subpopulations of enrollees. Three factors -- total taxable resources, percentage of potential aged Medicaid eligibles, and size of the waiver waiting list for ID/DD HCBS waivers -- were only associated with increased HCBS use for individuals with ID/DD. Other factors -- availability of home health aides, rates of consumer-direction, and availability of assisted living and residential care units -- were only significantly related to HCBS use for individuals who were aged or had physical disabilities. These findings underscore the importance of assessing drivers of variation in HCBS use and expenditures for subpopulations separately, as different factors appear to be relevant for each group.

Given the complex and dynamic environment across and within states for LTSS delivery, it is not feasible to isolate and determine the precise nature of the relationship between a single state policy constraint or factor and HCBS balance in a descriptive analysis. Further analysis is needed to understand the interaction of different factors underlying the relationships identified in this analysis.

Our exploratory analysis of the associations between system performance, state policies, and other factors that might facilitate or hinder Medicaid HCBS expansions suggests that:

  • Two factors over which states have little control -- poor weather conditions and size of the workforce needed to provide adequate HCBS -- are associated with systems less balanced toward HCBS.

  • Three factors that states could alter -- availability of Medicaid consumer-directed services, state plan personal care coverage, and availability of state Supplemental Security Income supplements for people living in the community -- are positively associated with systems more balanced toward HCBS. Note that consumer-direction may promote HCBS use because it has the potential to enlarge the workforce insofar as self-directing program participants are not limited to receiving services from workers recruited into home care agency employment but are permitted to choose to hire other individuals who may be motivated to become paid helpers because of pre-existing personal relationships, as relatives, friends, and neighbors.

  • State policies and constraints are likely to function differently for different subgroups of Medicaid enrollees. Consumer-direction, for example, was significantly associated with HCBS spending for the aged and people with physical disabilities, but not for enrollees with ID/DD. Other factors appear to be related to progress in re-balancing LTC for people with ID/DD, most notably financial resources.

We cannot infer causal relationships from these findings, but rather note that they point the way to possibly fruitful work in the future.

Directions for Future Research

Our findings indicate that alternative system performance indicators provide a more nuanced understanding of LTC system transformation and potentially could lead to different conclusions about program effectiveness and re-balancing efforts across states and subgroups than those based on one or two aggregate measures, such as total Medicaid spending on non-institutional compared to institutional LTC or total numbers of Medicaid beneficiaries receiving HCBS compared to those residing in nursing homes, long-term hospitals, or ICFs/IID. It will be important for future studies to assess state LTC systems on multiple dimensions for distinct target populations. As Medicaid continues to serve more enrollees in the community, it also will be important to monitor the breadth and type of LTC services low-income people need and receive.

Several promising policy options -- including Medicaid and non-Medicaid policies -- are associated with LTC system performance, but longitudinal studies will be needed to assess impacts. Of particular interest are which approaches are most cost effective and their applicability to different Medicaid subgroups. As state budgets change over time, also of interest is the extent to which fiscal constraints will limit states' ability to support or maintain HCBS expansions into the future.

I. Background and Objectives

Expanding the role of community care relative to institutional care has been a goal of long-term care (LTC) policy almost before LTC policy can be said to have existed. As Vladeck(1980), among others, has noted, the Social Security Act of 1935 prohibited federal matching payments for Old Age Assistance (OAA) to "inmates of public institutions." Such institutions were understood at the time to include public almshouses, where many poor and frail elders had resided. A precursor to today's Supplemental Security Income (SSI) for the elderly and younger, OAA payments were often referred to as "outdoor relief" in contrast to the "indoor relief" provided by municipal and county-owned and operated almshouses. Thus, OAA was intended to provide elders with an income that would permit them to live in the community, away from the often dreadful conditions of the almshouse. As it happened, the prospect of receiving OAA payments may have encouraged poor elderly people to leave almshouses, but it did little to meet the care needs of those whose functional limitations or disabilities prevented them from caring for themselves. As such people turned increasingly to private institutions as a source of care, they stimulated the growth of what became the private nursing home industry.

These events paralleled substantial increases in United States life expectancy. Since 1930, the life expectancy of a 65-year-old has increased by 5.6 years (to 82.3) for men and by 7.2 years (to 85.0) for women (Arias 2008). Over the same period, the life expectancy of people with developmental disabilities grew by nearly 40 years (Lightfoot 2006). These changes led to increased reliance on institutions to house and provide care for the elderly and those with physical disabilities and/or cognitive impairment. Previously, institutional care, especially at public expense, had been for the indigent elderly and disabled who lacked family to care for them at home, although some religious and charitable institutions served the "genteel poor" (mostly childless spinsters and widows who had outlived their inheritances). The 1950s saw the rise of a new phenomenon, termed "medical indigence," which referred to elderly people who entered private nursing homes and paid out-of-pocket but exhausted their resources and sought public assistance. During the 1950s and 1960s, children born with Down's syndrome or other conditions associated with intellectual, as well as physical developmental, disabilities were routinely institutionalized at birth and were expected to remain in the institutionfor life. Most nursing homes for the elderly were privately owned and operated, but most institutions for the developmentally disabled were state-run and entirely state-funded.

During the 1950s, federal funding for nursing home care was expanded under the Kerr-Mills Act, which included coverage for the medically indigent, as well as the long-time poor elderly (Moore and Smith 2006). In 1965, Kerr-Mills was replaced by Medicaid, which established an individual entitlement to state/federal assistance to pay for "skilled" nursing home care for individuals too poor to pay privately, again including residents who entered nursing homes as private payers and "spent down" to means-tested eligibility after exhausting their personal financial resources. This nursing home coverage was one of five "required" services (along with coverage of hospital care, physician services, x-ray and laboratory services, and "skilled" home health services provided by home health agencies (HHAs) certified to provide Medicare as well as Medicaid coverage) that states were federally mandated to include in their Medicaid state plans. Additional services could be included at state option. In 1972, Medicaid's coverage of institutional long-term care (ILTC) was expanded to encompass, at state option, care in intermediate care facilities. These included a category of nursing homes that were not required to employ registered nurses, but only licensed practical nurses and nurses' aides, and intermediate care facilities for the mentally retarded (ICF/MR) (most of which were state-run institutions). By fiscal year (FY) 1978, expenditures on long-term services and supports (LTSS) -- almost entirely ILTC -- accounted for 40 percent of total annual Medicaid expenditures (HHS 1981).

By the 1970s, policymakers had become increasingly concerned, not only by the growth and cost of ILTC, but also by scandals involving poor quality of care and the abuse, neglect, and mistreatment of residents in some facilities -- both nursing homes for the elderly and state facilities for the developmentally disabled. Congress began to consider allowing and encouraging states to pay for home and community-based alternatives to institutional care. In 1975, non-institutional personal care services were added to the list of optional benefits that states could elect to offer, although few states chose to do so at first. By 1978, only 13 states and the District of Columbia offered "personal care services" at home as an optional Medicaid benefit to low-income elderly and disabled beneficiaries who required help with basic activities of daily living such as bathing, dressing, transferring, eating, and toileting. Although Medicaid coverage required that a physician prescribe these services, the services were provided by unlicensed home care aides. Unlike "skilled" home health services, personal care services could be provided by non-certified agencies or by individually employed "independent providers," rather than by Medicare/Medicaid-certified agencies. New York State alone (mostly New York City) accounted for 75 percent of Medicaid personal care services expenditures. Other states, most notably California, chose to provide in-home personal care services to low-income people with disabilities with federal Title XX grant funding, supplemented by state social services dollars. The cost of these Medicaid and other publicly-funded home care programs was miniscule compared to Medicaid nursing home expenditures (Health Care Financing Administration 1981).

During the late 1970s and early 1980s, the Federal Government sponsored controlled experimental design research and demonstration projects to test the cost effectiveness of home and community-based alternatives to institutional care for the elderly (Kemper et al. 1987). Although these experiments largely failed to show that increased spending on home and community-based services (HCBS) significantly reduced nursing home use, Congress nevertheless amended Medicaid law to permit states to request federal approval to offer HCBS alternatives to institutional care (long-stay hospital, nursing home, or care in institutions for the developmentally disabled) under so-called 1915(c) waivers to people judged to be at high risk for admission to institutions.

During the first 15 years after HCBS waivers became available, spending on Medicaid-financed HCBS remained low compared to spending on ILTC. Federal and state officials were concerned that many elderly and younger physically disabled adults who qualified for nursing home admission were not likely to enter such facilities. Increased access to HCBS under 1915(c) waivers that did not generate offsetting reductions in nursing home expenditures would cause growth in total Medicaid spending on LTSS.

To control costs, enrollment in HCBS waiver programs was restricted by requiring states to obtain federal approval for a limited number of 1915(c) HCBS waiver "slots." The Executive Office of Management and Budget and the U.S. Department of Health and Human Services (HHS) Centers for Medicare and Medicaid Services (CMS) enforced what was termed the "cold bed" rule, under which states could not be approved for more waiver slots than available institutional beds (Shirk 2006). In addition, to be granted HCBS approval for the number of waiver slots requested, states often were required to submit assurances that they planned to close existing facilities or not expand institutional bed capacity as previously planned. During these years, HCBS programs frequently targeted children and adults with developmental disabilities because states had begun to close or downsize state-run facilities for the developmentally disabled. Mildly and moderately intellectually developmentally disabled adult residents of institutions began to be "de-institutionalized" to small-group homes. When federal special education funding became available and states were mandated to provide special education (1975), admission of all but the most severely disabled children with developmental disabilities into institutions ceased.

In 1994, at the request of the National Governor's Association, the Clinton Administration agreed to abandon the cold bed rule. In 1995, despite the availability of the state plan personal care services optional benefit and 1915(c) HCBS waiver authority, only 17 percent of Medicaid LTSS expenditures were for HCBS. In the mid-1990s, however, the United States economy was doing well, and states became increasingly willing to spend more on HCBS without being concerned about whether savings from reduced institutional care offset growing expenditures for HCBS. In addition, in 1997, Congress passed provisions of the Balanced Budget Act that were intended to put a stop to overuse of the Medicare home health services benefit to finance long episodes and large numbers of aide visits. This put pressure on some states where utilization of Medicare-funded home health services had been especially high to expand Medicaid coverage of in-home aide services. Expanded funding for HCBS was further stimulated by the U.S. Supreme Court's 1999 Olmstead decision that the Americans with Disabilities Act (ADA) required states to offer home and community-based alternatives to ILTC whenever feasible. To encourage states to comply with the ruling and "re-balance" their LTSS systems away from reliance on institutional care toward HCBS, Congress voted funding, and CMS awarded $289 million for Real Choice/Systems Change grants to 39 states between 2001 and 2010.

The Deficit Reduction Act (DRA) of 2005 expanded options for Medicaid coverage of HCBS by allowing states to offer services similar to those provided under 1915(c) waivers under their state plans (that is, without requiring federal "waiver" approval). Moreover, the requirement that HCBS be a cost effective substitute for institutional care was dropped. The DRA established 1915(i) HCBS services, which could be offered to individuals who did not meet "level-of-care"need criteria for coverage of ILTC. The DRA also provided "Money Follows the Person (MFP)" grant funding that gave states a financial incentive (enhanced federal matching funds) to transition nursing home residents back to community living with HCBS. A further DRA provision expanded opportunities for states to offer consumer-directed HCBS that allowed disabled Medicaid beneficiaries and their families to exercise more choice and control over the type and amount of HCBS they received by managing budgets or receiving cash payments -- and also expanded state flexibility to allow spouses and parents of minor children to become paid service providers.

In 2010, the Affordable Care Act (ACA) amended the 1915(i) HCBS state plan optional benefit to require states choosing it to provide the benefit on an entitlement basis to all who meet coverage requirements. States were prohibited from setting a cap on enrollment, which often resulted in waiting lists for HCBS waiver programs. The ACA also extended the MFP grant program and provided enhanced federal matching funds, called "Balancing Incentive Payments (BIP)," to encourage states to increase spending on HCBS. These payments targeted states that were not yet devoting more than one-quarter (more than one-half in some cases) of their Medicaid expenditures on LTSS to HCBS, as long as states agreed to meet the required spending targets by the end of FY 2015. Finally, the ACA also authorized yet another optional state plan benefit, called "Community First Choice," which provided states with a financial incentive (six additional percentage points of federal funding) to offer personal assistance services at home to all beneficiaries meeting "level-of-care" need criteria for nursing home coverage.

Since 1999, Medicaid HCBS use and expenditures have more than doubled (Eiken et al. 2011; KFF 2012), 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; Lakinet al. 2009). This general shift masks wide variation in the levels of re-balancing across states (Howes 2010; Kassner et al. 2008; KFF 2012; Wenzlowet al. 2011). Efforts to re-balance LTSS systems from their traditional reliance on institutional care to HCBS have also been achieved more widely for some populations (enrollees under age 65 with disabilities) than others (people over 65) (Wenzlow et al. 2008, 2011).

In this report, we examine patterns in LTSS use across states and subgroups of enrollees in 2009, just after many states began to experience fiscal constraints and increased demand for services from the national recession. This analysis updates findings from a previous study based on 2006 data and expands upon state-level factors linked to LTSS systems that are more balanced toward HCBS use.

A. Re-Balancing as a Goal

The primary argument for the expansion and enhancement of community-based care has increasingly come to be that recipients greatly prefer it, an argument so widely accepted as to be regarded as self-evident. Farmer's (1996) study of nursing home organization was explicitly based, in part, on the assumption that "no one's first choice of residence was a nursing home." A recent AARP survey described by Keenan (2010) found that 86 percent of respondents aged 45 and older, asked where they would like to live as they aged, agreed or strongly agreed with the statement, "What I'd really like to do is stay in my current residence for as long as possible."

Apart from the well-accepted preference of people over age 65 or people with disabilities for community-based care is the strong likelihood of better health outcomes for people who are successfully supported in the community compared to outcomes likely to have occurred under institutional care. Outbreaks of influenza and norovirus, for example, are recognized threats to people living in institutions. Nursing home-acquired pneumonia is a significant cause of mortality and morbidity to residents (Mills et al. 2009). Although evidence sometimes conflicts, several studies have demonstrated that residents in assisted living and HCBS settings have fewer depressive symptoms and better psychological well-being, and more were generally happy (Pruchno and Rose 2000; Franks 2004; Wodchis 2003).

Finally, of course, evidence suggests that providing services in the community may result in cost savings to state Medicaid programs. Several studies (for example, Kitchener et al. 2006; Kaye et al. 2009; Harrington et al. 2011; Kaye 2012) find that replacement of institutional with community LTSS resulted in cost saving for Medicaid programs. These studies provide only incomplete guidance to states, however, for two reasons. First, the term "HCBS" is much less well defined than is institutional care. A nursing home day is commonly understood to mean round-the-clock care and to include room, meals, and other services. In contrast, HCBS care may include quite different services from state to state, possibly with limitations on the number of hours of care allowed. Second, empirical studies can provide information only on the results of re-balancing that has occurred so far. Given that states and providers surely began re-balancing by focusing on those people most easily supported in the community, rather than, say, a random group of people eligible for institutional care, evidence on re-balancing to date provides little or no evidence on the probable effects of further extending efforts to re-balance care for the elderly and people with disabilities.1 That said, the most reasonable policy for states to adopt, given current research, would appear to include gradual expansion of community-based LTSS, at least until evidence suggests that further expansion was more expensive or led to worse outcomes than institutional care. No evidence yet indicates that we have reached that point.

Despite evidence of cost effectiveness, states facing budgetary crises may be less likely to develop and implement new programs, particularly due to evidence that shifts to HCBS may involve initial cost increases at the beginning of these programs before savings are achieved (Kaye et al. 2009). Given ongoing economic constraints in many states, policymakers have questioned how budget concerns have affected progress toward HCBS re-balancing across states, and within states, for subpopulations of enrollees. Initial feedback indicates that many states have continued to pursue the goal of re-balancing during the recent economic downturn (Walls et al. 2011; Cheek et al. 2012), but limited information is available on how progress has varied across states and for subpopulations of enrollees during this period.


1 New Jersey's Rebalancing Workbook, for example, recognizes that "the least frail client in a nursing facility (NF) could be the frailest client in HCBS, and the frailest client in HCBS could be the least frail client if moved to a NF." (New Jersey Department of Health and Senior Services 2009).

B. Challenges to Re-Balancing

It can hardly be surprising that adoption of HCBS has varied strongly from state to state. For obvious reasons, community-based care is easier to implement in urban areas, where care providers can more easily reach elderly people and people with disabilities in their homes. In contrast, in highly rural states or areas, even people who are largely independent may sometimes receive care in nursing homes simply because community care is impractical in remote localities. Moreover, implementing new modes of care requires the development of new systems of initiating, arranging, and managing care, much of it paid for by third parties. This implementation, difficult in itself, also creates new economic winners and losers who invariably will attempt to influence legislatures and regulatory bodies to alter the process. These influences can be hard to discern and virtually impossible to measure, making it difficult to assess their importance on progress in re-balancing.

Until recently, a major barrier to re-balancing toward greater reliance on HCBS was concern that open-ended "entitlement" funding of HCBS would give rise to a phenomenon referred to as the "woodwork effect" -- a more colorful way to describe what economists more commonly term "induced demand." The argument held that institutional care is so unpleasant that eligible individuals and their families often are willing to incur significant private cost (mainly in the form of unpaid family caregiving) to avoid it. When publicly-funded HCBS becomes available, they no longer have a reason to pay the price of unpaid family caregiving to avoid institutionalization, accessing paid home care instead. Moreover, when Medicaid and other public payers pay family members to provide aide and attendant care, which has increasingly been allowed, demand for HCBS may be further stimulated. Accordingly, many people who qualified for institutional care but would not have sought admission or would have postponed admission for as long as possible will "come out of the woodwork" to claim these far more desirable HCBS benefits. This will result (even if there is substitution of HCBS for nursing home care by some), in far greater demand for publicly-funded LTSS that translates into higher total expenditures -- and increased burden on federal/state budgets -- than would have been the case if institutional care remained the only option.

Although much of the research literature on "woodwork effect" dates back to the 1980s (see, for example, Kemper et al. 1988), academics have continued to debate its magnitude (Grabowski 2006; Kaye 2009). In 2013, a special issue of the Journal of Aging and Social Policy was devoted to the topic. Edited by Frank Caro, it contains articles by Mitchell La Plante, Robert Kane, Steve Eiken, and William Weissert. Although these authors attempt to measure the effect using different data sets and methods, all have generally concluded that the growth in the number of users of Medicaid-funded HCBS has greatly exceeded the reduction in numbers of users of Medicaid-funded nursing home care. They further agree that, whereas increased funding for HCBS apparently had only a modest effect on decreasing nursing home use among the elderly, there is no strong evidence that it has led to either an increase or decrease in total Medicaid LTSS expenditure. Where they disagree is on the importance of doing a better job of targeting HCBS spending so that it will have a greater likelihood of reducing nursing home use.

Absence of a clear link between increased funding for HCBS and increased Medicaid LTSS costs may help explain the diminished concern about "woodwork effect" among federal and state policymakers. There are other possible reasons as well. First, Medicaid means-testing places an upper limit on how many disabled individuals -- especially elderly people with pension income and life savings -- can meet Medicaid's strict financial eligibility test. The allowable asset limit ($2,000 for an individual in nearly all states) is particularly restrictive and has not been raised in nearly three decades. Second, states have a mechanism to control the flow of access to HCBS, if they choose to use it. Although more than half the states provide HCBS (personal care services) as individual entitlement to all who qualify financially and on the basis of disability-related need, the remaining states rely exclusively on 1915(c) waivers to finance HCBS. Medicaid law allows states to cap enrollment into HCBS waivers and establish waiting lists. In 2012, the number of people on waiver waiting lists for elderly/disabled HCBS waiver programs nationally exceeded the total number of approved "slots" by 22 percent (Ng 2013). However, examination of unpublished waiting list data shows that nearly half of all states using 1915(c) waivers were operating their elderly/disabled waiver programs as virtual entitlements because they had requested and obtained CMS approval for enough slots to enroll all who qualified and reported no waiting lists. More than half of those waiting on HCBS elderly/disabled waiver waiting lists in 2012 were concentrated in three southern states (Florida, Louisiana, and Texas).

Waiting lists for HCBS, particularly those that require enrollees to wait a year or more, admittedly exist in a state of tension with the ADA and Olmstead. The Olmstead decision itself explicitly recognized that "a waiting list that moved at a reasonable pace not controlled by the state's endeavors to keep its institutions fully populated" would meet the standards of the ADA if it existed within the context of "a comprehensive, effectively working plan for placing qualified persons with mental disabilities in less restrictive settings." Neither the Supreme Court nor lower courts adjudicating similar cases since have taken the position that states cannot control HCBS access by establishing waiting lists, and they have left the definition of "reasonable pace" vague. Courts typically have not questioned waiting lists in states that have a so-called "Olmstead Plan."2

Although initial policy research on the "woodwork effect" focused almost exclusively on the elderly (probably because of concerns about the potential cost impact of an aging population), states have experienced far greater difficulty meeting demand for HCBS waiver services for individuals with intellectual and developmental disabilities (ID/DD). Medicaid's means-test does not serve to restrict eligibility among adults with ID/DD as it does for the elderly, because almost all adults with ID/DD severe enough to qualify for HCBS as an alternative to institutional care are receiving SSI, which gives them automatic Medicaid eligibility.

Mitchell LaPlante (2013) asserts that the existence of a "woodwork effect" in terms of growth of users of Medicaid-funded services for people with ID/DD is clear because there are now many times more Medicaid beneficiaries with ID/DD receiving HCBS via waiver programs than there ever were residents in ICFs/IID: "In 1992, there were about 145,000 persons in Medicaid-funded ICFs/IID, and 62,000 received community LTSS. From 1992 to 2009, there were 50,000 fewer people in ICFs/IID, but greater than 500,000 more persons being served in the community. This would appear to be a classic woodwork scenario…". Waiting lists for HCBS waiver services are far more common and wait time is longer for people with ID/DD than for the elderly and physically disabled. In 2012, the number of those waiting for HCBS ID/DD waiver services exceeded existing enrollment by 58 percent, and the average time spent on the waiting list, nationally, was 47 months (Ng 2012). Nevertheless, more than a quarter of states (14) reported that they had no individuals waiting for HCBS ID/DD waiver services.

Overall, in 2011 and 2012, even though the United States had yet to fully recover from the Great Recession and its negative impact on state budgets, about 20 percent of states were fully able to meet the demand for HCBS among all Medicaid LTSS subgroups (for example, elderly, adults with physical disabilities, children and adults with developmental disabilities, and others with special needs, such as those with HIV/AIDS or traumatic brain injury). What explains why some states have no or very short waiting lists for HCBS waiver programs, whereas others have very long waiting lists? One recent study of waiver waiting lists for HCBS waivers serving physically disabled people in Iowa casts doubt on whether maintaining first-come, first-served waiting lists is an effective way to control Medicaid LTSS costs. The study found that, in Iowa, long waiting periods for high-risk beneficiaries increased nursing home use and costs enough to fully offset savings from delaying access to HCBS (Peterson et al. 2013).3 These questions clearly deserve further research. In this report, we explore only whether a statistically significant relationship exists between HCBS waiver waiting lists and "re-balancing" toward higher levels of spending on HCBS compared to institutional care.


2 A ruling of the Third Circuit Court, for example, stated that "When such a plan exists, a remedy that would force the agency to abandon or alter its long-term compliance efforts could sacrifice widespread compliance for immediate, individualized relief. Imposing such a remedy would be penny-wise and pound foolish." (Pennsylvania Protection and Advocacy v. Pennsylvania Department of Public Welfare, 2005).

3 The authors also found that, among high-risk cases, estimated LTC costs over the two years following application were higher for those facing long wait lists than for those facing shorter wait lists, although this result was not statistically significant.

C. State Actions to Increase Use of Home and Community-Based Care

States, of course, cannot directly control the balance of institutional and community care for elderly and Medicaid enrollees with disabilities. Balance is instead the result of each state's Medicaid LTSS eligibility and payment policies, regulation, and the extent of communication and coordination with providers and enrollees. Medicaid policy naturally becomes more community-friendly when it covers personal and home health care without strict limits on the quantity of services and when it is covered under the state plan, rather than under waivers that can limit the number of enrollees who can be served or the communities where it is offered. In addition, states must set personal needs allowances at a level that is realistic for a person living in the community. Even generous coverage of community LTSS will fail if people cannot retain enough income to maintain themselves in a home or apartment. Wenzlow et al. (2011) found that availability of Medicaid consumer-directed services, state plan personal care coverage, and availability of state SSI supplements for people living in the community were each positively associated with greater levels of community care.

States can control, and take steps to promote, the availability of residential (facility-based) care settings other than nursing homes and ICFs/IID. Historically, the Medicaid definition of "home and community-based services" has included coverage of assistive services, but not room and board provided in residential care facilities other than those defined as "institutions" where Medicaid reimbursement does cover room and board costs (hospitals, NFs, and ICFs/IID). In the early years after Medicaid was enacted, the numbers and bed capacity in such residential care facilities appear to have declined because Medicaid coverage and reimbursement rules greatly encouraged providers of residential LTSS to meet Medicaid institutional "conditions of participation" and become certified for the more generous Medicaid payment rates. Initially, Medicaid-covered only "skilled" nursing homes. In 1972, however, coverage was extended to "intermediate care" nursing homes (where aides could be supervised exclusively by licensed practical nurses rather than registered nurses) and facilities formerly labeled "state schools" for people with ID/DD could be redefined as medical institutions and become eligible for Medicaid reimbursement as ICFs/MR.

In the late 1970s and early 1980s, studies comparing the cost effectiveness of HCBS as a substitute for institutional care using experimental design methods sometimes found that diverting low-income elderly seeking nursing home admissions into alternative residential care settings (variously termed personal care homes, domiciliary care, board and care, or adult foster care homes) could be cost effective (Doty 2000). Before the 1990s, these alternative residential settings for the elderly catered primarily to those who qualified for SSI cash assistance.

The 1990s saw the rise of the "assisted living" industry. A handful of states (Oregon and Washington in particular) were active in promoting and vocal in advocating to other states the use of assisted living and small adult foster care homes for Medicaid beneficiaries as an alternative to nursing home placement. Although assisted living residential care, especially facilities offering primarily private apartments and other amenities, caters primarily to private payers, the number of facilities accepting Medicaid residents and the actual number of Medicaid residents has increased. ASPE research in the early 2000s found that about one-third of all residential elder care beds were in facilities other than certified NFs (Spillman et al. 2002). The ASPE-sponsored 2011 Residential Care Survey found that more than 730,000 Americans lived in residential care facilities in 2010 (Park-Lee et al. 2011). Other data sources, such as the National Health and Aging Trends Survey, that define residential elder care even more broadly estimate the numbers and percentage of elderly residing in non-nursing home residential care settings as opposed to "at home" in ordinary housing not designed for elder care to be even higher.

Wenzlow et al. (2011) found that greater availability per 1,000 elderly population of beds in non-nursing home elder care facilities ("assisted living" broadly-defined) was associated with "re-balancing" toward HCBS for the elderly. For younger adults with developmental disabilities, the initial trend toward "de-institutionalization" involved closing down large state institutions for this population and transferring residents to smaller "group homes." The difference between "institutional" and HCBS settings was often blurred because states could choose to establish small ICFs/IID with 15 or fewer beds that were still eligible for coverage of room and board costs but operated under special Medicaid regulations. These facilities could be state-run as well. Increasingly, most states transitioned former residents of large ICFs/IID into small fewer than 16 bed "group homes" under private for-profit or non-profit auspices that were eligible for Medicaid reimbursement of HCBS only. "De-institutionalization" through the downsizing and closure of large ICFs/IID and their replacement by smaller, less costly residential settings with 15 or fewer beds was associated with a substantial increase in the numbers of individuals with ID/DD receiving Medicaid-funded services. This resulted in rapid "re-balancing" toward HCBS for Medicaid recipients with ID/DD receiving LTSS. Since the Supreme Court's 1999 Olmstead ruling, advocacy for people with ID/DD has increasingly promoted HCBS delivered "at home" to individuals residing with family members in the family home or to individuals in supported living arrangements (persons with ID/DD residing in homes apartments with 1-2 unrelated roommates and a paid helper) or in small-group homes with no more than six residents (Smith et al. 2007). According to Charlie Lakin and colleagues at the University of Minnesota, the number of residents in homes with six or fewer people with ID/DD increasing from 20,400 people in 1977 to 321,500 people in 2010. By 2010, 11 states had no state-operated residential facilities for 16 or more people with ID/DD (Lakin 2011).

On January 16, 2014, CMS issued a final rule defining "community care" that restricts Medicaid HCBS spending to residential care settings that are "non-institutional" in character (HHS 2014). In other words, facilities that are not certified as institutions eligible for Medicaid reimbursement for room and board as well as services may nevertheless be determined to be also ineligible to be HCBS providers because they are considered de facto "institutions" that are overly restrictive and do not offer residents adequate privacy.

Wenzlow et al. (2011) had no measure of non-home/non-ICF/IID residential care use among the Medicaid population with ID/DD receiving LTSS, but this report does include such a measure. We do not hypothesize a relationship between availability of non-nursing home residential care settings and beds and re-balancing toward HCBS for the population of Medicaid LTSS users under age 65. ASPE's 2011 Residential Care Survey did find some residents under age 65, and compared to elderly residents, they were disproportionately more likely than private payers to be on Medicaid (Green et al. 2013). Still, there is little evidence of a movement toward placement of younger adults with physical disabilities in alternative residential care settings rather than in nursing homes or "at home" in their own or family homes. Advocates for younger adults with physical disabilities oppose age/disability segregated residential settings for their constituency, and they consider assisted living to be such a form of segregated housing. Because of this opposition, the MFP grant program explicitly denies the enhanced federal matching payments available to states that help nursing home residents who transition back to the community (most of whom are adults under age 65) if they go to live in assisted living facilities (Reinhard 2012). Advocates for younger adults with spinal cord injuries and other physical disabilities hope to promote accessible public housing in scattered sites so that such people can be integrated into the community and reside as neighbors alongside people of all ages with and without disabilities (see for example, http://wheelsofprogress.org/scatter.html). Nevertheless, lack of affordable housing that is also accessible to people with severe physical disabilities is one of the most often cited barriers to transitioning nursing home residents back to community living.

Re-balancing can also be encouraged through regulation and incentives aimed at nursing homes and ICFs/IID. Some states, for example, have set numerical goals for reductions in institutional beds (Kane et al. 2008). Others allow facilities to place beds on layaway status for several years to ease the path toward eventual delicensure. Measures like these, appropriately used, can ensure that institutional care is provided when in the best interest of the recipient, but not simply because it is the default approach to providing LTSS.

Finally, re-balancing can be accelerated by focusing on system accessibility and efforts to estimate the importance and availability of informal caregivers. The literature suggests that having "a single entry point" (SEP) or "no wrong door" model for LTSS allows participants to more easily navigate the complex array of services available, but states vary in the services these SEPs provide and the populations they serve, which makes predicting the overall effects of an SEP model on re-balancing difficult (Kassner et al. 2010; Reinhard et al. 2011; Mollica and Gillespie 2003). Finally, because research suggests that shifting the balance toward HCBS affects families and caregivers by shifting some work from paid providers to informal caregivers (Feinberg and Newman 2004; Rozario and Palley 2008), it is also important to assess the level of support for these informal caregivers across states.

D. Measuring and Understanding LTSS System Performance

This report examines patterns in LTSS use across states and subgroups of enrollees in 2009, just after many states began to experience fiscal constraints and increased demand for services from the national recession. The analysis updates findings from a previous study based on 2006 data and expands on state-level factors linked to LTSS systems that exhibit greater HCBS use.4 In addition, the ACA of 2010 expanded states' options for offering HCBS through Medicaid.5 Analyses based on data from 2009 provide a baseline of HCBS balance across and within states for future comparison to state policy and program changes that resulted from options made available in 2010.


4 In this report, the word "states" is meant to include the 50 states and the District of Columbia.

5 The ACA gave states new options and financial incentives for providing HCBS to Medicaid enrollees. New options include the Community First Choice program (which allows states to cover personal care and other services for eligible individuals through their Medicaid state plans) and the BIP (which provides incentives for eligible states to re-balance their LTC system toward HCBS). The ACA also expanded the scope of covered services and eligibility requirements for Section 1915(i) programs and extended the MFP program with additional funding. For more information on states' adoption of new HCBS options, see Walker (2010); KFF (2011); and U.S. Government Accountability Office (2012).

1. Measuring Progress

The most commonly used indicators of LTSS system balance -- the percentage of LTSS spending allocated to HCBS and the proportion of LTSS users receiving HCBS -- 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. 2011), 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 the University of California, San Francisco (KFF 2012). 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 all people who need them. For this reason, researchers and policymakers also use person-level administrative data in the CMS Medicaid Analytic eXtract (MAX) system to explore who is being served by Medicaid LTSS and to better understand system transformation.6 These projects include efforts to measure HCBS and institutional use and spending for various groups eligible for the MFP demonstration (Brown et al. 2008; Irvin and Ballou2010; Lester et al. 2013) and the AARP efforts to develop a state LTSS system scorecard that includes some person-based measures of system performance (Reinhard et al. 2011). We use MAX data in this analysis so that we can examine overall system performance, as well as performance for subgroups of Medicaid enrollees.


6 This study's predecessors summarized the strengths and limitations of MAX data for studying LTC (Wenzlow et al. 2008, 2011), finding that, although the MAX data were incomplete for some states, and service-specific information on HCBS was not reliable in all states, MAX provides useful information on which populations are receiving HCBS and how their use of HCBS compares to use of institutional care.

2. What Do We Know About the Progress in LTSS System Transformation?

As noted earlier, LTSS systems have increasingly emphasized and relied on HCBS (Doty 2010; KFF 2012). However, studies have shown substantial interstate and intrastate variations in this progress. Some states -- for example, Alaska, New Mexico, Oregon, and Washington -- have been identified as making these transitions successfully, whereas others have traditionally received low rankings (Wenzlow et al. 2011). Within states, performance indicators suggest that HCBS use is more common among working-age Medicaid enrollees with disabilities than among aged LTSS recipients (Wenzlow et al. 2011). Moreover, the predecessor to this study found that, within the population of individuals under65 with disabilities, those individuals with ID/DD had particularly high rates of HCBS use, higher than rates for individuals under 65 with other disabilities (Wenzlow et al. 2011).

Factors that might be related to balanced LTSS systems 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 LTSS budgeting, standardized assessment tools, transition programs, and quality improvement. In a study of MFP grantees, Irvin and Ballou(2010) found two additional features -- the depth of HCBS experience and coverage of optional state plan personal care -- among systems that were more balanced in terms of LTSS spending. Ruttner and Irvin (2013) found that states offering personal care services through state plans, as opposed to through waivers alone, spend a higher median share of their LTSS expenditures on HCBS. A 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 LTSS systems in hard economic times (Rose et al. 2010). In this study's predecessor, two factors over which states have little control -- poor weather conditions and the size of the workforce needed to provide adequate HCBS -- were associated with systems that were less balanced toward HCBS. Conversely, three factors that states can alter -- availability of Medicaid consumer-directed services, state plan personal care coverage, and availability of state SSI supplements for people living in the community -- were positively associated with systems that were more balanced toward HCBS (Wenzlow et al. 2011). Finally, the importance of state characteristics associated with rates of re-balancing toward HCBS may vary across different age groups of LTSS users (Miller 2011). Greater state investment in HCBS and reduced nursing home capacity were associated with re-balancing for aged individuals, but rates of institutionalization of working-age adults were more closely associated with state sociodemographiccharacteristics and chronic disease prevalence.

Many important questions about LTSS system performance remain unanswered. Are states identified as successes 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 LTSS in home and community-based settings; hence, the lessons learned in more urban states may not apply. Finally, what effect have a weakened economy and constrained state budgets had on progress toward re-balancing in recent years? Insights into these questions would be particularly helpful to states as they face budget crises and as some consider cuts to Medicaid rather than expansions of it.

E. Goals of This Study

This study expands on earlier work in Wenzlowet al. (2011), using MAX 2009 to assess patterns of both interstate and intrastate variations in LTSS system performance. The study has two broad aims: (1) to update information on differences in LTSS systems within and across states that were identified in the previous report; and (2) to explore how state constraints and policies are related to the balance of LTSS systems.

To characterize the performance of LTSS systems in each state, we summarize HCBS and institutional care service use and expenditures to determine whether some states are achieving more balanced systems either by serving more people or spending more per person covered compared to other states. We also explore how balance varies for 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.7 (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. Finally, we compare state-level results from 2009 with results on these same measures in 2006 to assess progress toward re-balancing during this period of budget constraints in many states.

The second set of analyses explores how state constraints and policies are associated with the LTSS system performance indicators assessed 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 state policies are linked with LTSS balance. 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. In this analysis, we examine the continued relationship between factors that Wenzlowet al. (2011) found to be relevant, as well as new measures of state policy and program features for which data have become available since the previous study was conducted.

Because we could not assess differences in need for care or its appropriateness, we do not assume that a more balanced system always will reflect state success in providing HCBS to populations that need them. Although we rank states by the characteristics of their systems, our aim is to gain insight into how policies and state factors are related to LTSS system performance indicators so as to better understand "high" scores.

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


7 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.

F. Summary of Data and Methods

We used MAX 2009 Person Summary (PS) files to develop measures of LTSS 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 users of and spending on Section 1915(c) waiver services -- an important vehicle that most states use to provide 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, ICF/IID care, inpatient psychiatric services for people under age 21, and psychiatric hospital services for those 65 and older.8 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 includes all care received in the selected institutions, whether or not a person is using them for LTSS.9

Analyses were limited to Medicaid enrollees eligible on the basis of disability or age and who were eligible for full Medicaid benefits in 2009.10 We excluded enrollees in the Program of All-Inclusive Care for the Elderly (PACE) and other managed care plans because information on their use of services (HCBS or institutional care) is missing or unreliable in MAX in many states. We also reviewed MAX 2009 data documentation to identify data quality concerns related to LTSS use in each state. Based on these assessments, we excluded from the analysis 13 states with MAX fee-for-service (FFS) data that are potentially unrepresentative or unreliable due to high levels of managed care penetration among the aged and disabled population or due to data quality concerns. The excluded states are Arizona, Hawaii, Maine, Massachusetts, Michigan, Minnesota, Montana, New Mexico, Oregon, Pennsylvania, Rhode Island, Tennessee, and Wisconsin. Finally, we could not differentiate enrollees with physical disabilities from those with ID/DD in the District of Columbia, Vermont, and Washington, and excluded these states from the relevant subgroup analyses. A more detailed discussion of the MAX data, analyzed measures, and methods used is in Appendix B. Appendix C lists state-specific MAX data anomalies for 2009.

The analysis of state constraints and policies related to LTSS provision relied on a wide range of publicly available data sources. When available, we used data from 2009 to capture policies in place and state characteristics at the time that services were being used.


8 An individual can receive both HCBS and institutional care during the year.

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

10 The population of enrollees eligible for full Medicaid benefits excludes the following enrollees with restricted Medicaid benefits: Medicare-Medicaid enrollees who are eligible only for Medicare cost-sharing, aliens eligible only for emergency services, individuals eligible only for family-planning services, and individuals eligible only for premium assistance support toward the purchase of private health insurance.

G. Road Map to This Report

In the following chapters, we characterize Medicaid LTSS 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 findings for the overall LTSS 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 Services and Supports System Performance

Policymakers are interested in learning about states' progress in incorporating HCBS into their LTSS systems, and the extent to which they can provide LTSS in the community for important subgroups of Medicaid enrollees who are aged or have disabilities. In this chapter, we summarize the variation in LTSS balance across states in 2009, overall, and for key subpopulations of enrollees, including aged individuals and, for those under 65 with disabilities, by whether or not they used LTSS designed for people with ID/DD. To assess recent state-level progress on balancing LTSS systems toward HCBS use, we compare our results from 2009 to previously published results that used 2006 data.

A. Measures Characterizing LTSS System Performance

As Wenzlow et al. (2011) argue, no single measure fully captures LTSS system performance in terms of the breadth of the population covered, and the breadth and intensity of services provided. For this reason, we used a combination of measures to capture variation in system performance across states, including the:

  • Percentage of Medicaid LTSS expenditures allocated to HCBS.

  • Percentage of LTSS users receiving HCBS.

  • Ratio of per-recipient spending on HCBS to spending on institutional care.

Although the first two measures are commonly used indicators of the degree to which states have balanced their LTSS systems toward HCBS use and spending relative to institutional care, each is limited, to some degree, and therefore should be interpreted with some caution. The share of LTSS expenditures allocated to HCBS is expected to increase as community care becomes a more frequently used component of LTSS. However, variation in expenditures can arise both from variation in the number served and from variation in payment rates. Thus, a state that increased its payments to institutional providers, perhaps as a result of imposing new minimum staffing standards, would show a decline in the percentage of LTSS expenditures allocated to HCBS, even though one might argue that overall balance had not changed or that this policy change is beneficial for people who need LTSS.

The percentage of LTSS users who receive HCBS is a similarly imperfect measure of differences across states. HCBS users are far more heterogeneous in their care needs than are users of institutional care. Some may be as impaired as those in nursing homes or ICFs/IID. Others may simply be people who received a few home health visits at some point during a year. States that allow many users to access HCBS, but impose limits on its quantity, might show a high value for the percentage of LTSS users who receive HCBS. In those states, much of HCBS is provided in amounts too low to substitute for institutional care. As a result of the problems just noted, it would appear reasonable to look to states with high values for both measures when seeking to identify those that had most successfully transitioned to a high level of community LTSS. The third measure captures the extent to which state spending on HCBS per user is similar to state spending on institutional care.

This may be a proxy indicator -- although an imperfect one -- for generosity of HCBS coverage. (It is imperfect as a measure of coverage generosity because states that provide predominantly agency-delivered aide services typically pay a higher hourly rate than states that rely primarily on consumer-directed independent providers; it thus could mix price and quantity effects.) In any case, consumer advocates have long argued that "money should follow the person" (not the provider), by which they mean that state Medicaid programs should be willing to spend as much on HCBS as they would be willing to spend on institutional care for someone with comparable disabilities. However, because HCBS users tend to be less severely disabled than nursing home residents, and nursing home care also encompasses room and board in addition to the cost of providing functional assistance, a ratio of per-user cost of HCBS to per-user cost of ILTC approaching 1:1 might raise the question of whether the average level of spending on HCBS per user could make it difficult for the state to afford to serve all Medicaid beneficiaries who qualify for HCBS coverage. An exceptionally high ratio of per-user spending on HCBS compared to per-user spending on ILTC might be associated with more restrictive level-of-care need criteria for coverage of both HCBS and institutional care and/or with HCBS waiver enrollment caps that have required establishing waiting lists.

B. Interstate Differences

Across the 38 study states, about 45 million enrollees were eligible for full Medicaid services in 2009, with about 10.5 million eligible on the basis of age or disability. About 7 percent of all full-benefit enrollees and almost 30 percent of enrollees who were aged or disabled used any FFS LTSS -- with higher rates of HCBS than of ILTC use (Table II.1). (See Appendix Table D.1 for state-level detail.) Medicaid-financed LTSS included in these estimates include HCBS (including 1915(c) waiver services and state plan services for personal care, residential care, home health, adult day care, and private duty nursing), as well as institutional services (including services provided in nursing homes and ICFs/IID). The list of states differs slightly from those appearing in Wenzlow et al. (2011), but LTSS utilization rates are consistent with findings in that report, which showed that just over 7 percent of all full-benefit enrollees used LTSS, and almost 5 percent used HCBS.

TABLE II.1. Number of Enrollees Who Were Aged or Eligible on the Basis of Disability Using Medicaid FFS LTSS Compared to the Total Number of Full-Benefit Enrollees in 2009

MeasureAll Full-Benefit
  Medicaid Enrollees  
Full-Benefit Aged
  or Medicaid Enrollees  
with Disabilities
Aged or Disabled
  with Any FFS LTSS  
Aged or Disabled
  with Any FFS HCBS  
Aged or Disabled
  with Any FFS ILTC  
Number, in thousands45,08110,5153,1302,0851,205
Percentage of all full-benefit Medicaid enrollees100.023.36.94.62.7
Percentage of full-benefit aged or Medicaid enrollees with disabilities---100.029.819.811.5

SOURCE: Mathematica Policy Research analysis of 2009 MAX data for 37 states and the District of Columbia with representative FFS LTSS data (excludes data from Arizona, Hawaii, Maine, Massachusetts, Michigan, Minnesota, Montana, New Mexico, Oregon, Pennsylvania, Rhode Island, Tennessee, and Wisconsin).
NOTES: Enrollees in managed LTSS and those eligible for only restricted Medicaid benefits are excluded. HCBS include 1915(c) waiver services and state plan services for personal care, residential care, home health, adult day care, and private duty nursing. ILTC includes services provided in nursing homes, ICFs/IID, mental hospitals for the aged, and inpatient psychiatric facilities for people under age 21

TABLE II.2. Expenditure and Utilization-Based Measures of LTSS System Performance Among Enrollees Who Were Aged or Had Disabilities and Were Eligible for Full Medicaid Benefits in 2009, Ranked by HCBS Share

State RankStates Ranked by Percentage of
LTSS $ for HCBS
States Ranked by Percentage of
LTSS Users Receiving HCBS
States Ranked by the Ratio of Per-User $
on HCBS Relative to Per-User $ on ILTC
$#  Ratio  StateTotal LTSS $% of Medicaid
  LTSS $ Allocated  
to HCBS
StateTotal LTSS Users% of LTSS
  Users Receiving HCBS  
State  Per-User $  
on HCBS
  HCBS $ Per-User/  
ILTC $ Per-User
   All 38 States    90,014,728,763  45.3All 38 States    3,130,010  66.6All 38 States  19,5470.478
132Washington1,930,549,58774.8Alaska7,76489.6New Hampshire31,6641.018
2128Alaska344,563,44773.8California664,24984.9Washington21,2840.857
376Vermont298,904,02361.2Washington82,97181.7Indiana28,9280.829
4235California12,064,245,82960.8Idaho19,70380.0Utah28,5050.771
5816Colorado1,300,230,15957.6Iowa55,29974.5Wyoming26,7360.738
6  15  5Wyoming206,001,29556.4North Carolina149,37173.8Vermont23,2430.726
7128Kansas1,006,566,45856.2Vermont10,70173.6Nebraska21,9320.695
8261New Hampshire463,202,43853.6Colorado46,66573.3Kansas20,0870.683
91118District of Columbia590,462,73350.5Virginia59,02872.6Louisiana21,5690.664
10  1612New York20,116,573,29250.0Missouri96,06870.0South Dakota20,2850.663
11913Virginia1,775,061,68348.7District of Columbia11,12168.5Delaware37,7580.653
12634North Carolina3,182,180,36645.8Kansas41,45767.9New York45,1500.630
132214Maryland1,996,568,84945.7Nevada13,40367.9Virginia20,1810.602
141329Nevada339,047,04245.6Alabama63,91867.7Maryland28,4230.599
151026Missouri1,836,511,96445.0Wyoming6,56166.3Georgia17,9870.594
161820Oklahoma1,151,255,03343.7New York345,39864.5Colorado21,8870.578
17314Utah369,135,85542.7South Carolina45,07263.1Ohio21,4920.567
18436Idaho461,085,97042.5Oklahoma53,60462.5District of Columbia39,1420.536
19533Iowa1,355,006,85842.2Texas246,81461.1Connecticut27,9610.525
20277Nebraska612,132,76642.1West Virginia28,75061.1Oklahoma15,0300.513
211921Texas5,505,748,86442.1New Jersey102,38960.2Texas15,3560.503
221727South Carolina1,075,756,27641.4Maryland53,82259.6North Dakota23,2230.496
232023West Virginia902,596,72841.2Ohio168,01159.3West Virginia21,1420.489
242317Ohio5,415,944,20639.6Illinois155,28757.7Kentucky16,9140.478
253010South Dakota286,371,34538.2Connecticut57,37256.3Illinois13,9740.474
26329Louisiana2,035,834,50238.0New Hampshire13,94156.2Missouri12,2890.461
272519Connecticut2,498,041,84736.2Nebraska21,54954.5South Carolina15,6580.442
28383Indiana2,160,937,79636.0Florida140,19851.9Alaska36,5470.438
293311Delaware340,623,57935.9Arkansas41,00051.2Nevada16,9900.435
302425Illinois3,540,342,81835.4South Dakota10,68050.5Florida17,1640.411
312132New Jersey3,835,410,40233.6Utah10,97550.4Arkansas12,8670.388
323615Georgia1,683,701,04533.0Louisiana72,17449.7New Jersey20,8970.362
332830Florida4,155,577,85830.0Delaware6,71048.2Iowa13,8820.354
341437Alabama1,332,176,14629.7Mississippi40,50646.6North Carolina13,2040.346
353522North Dakota338,845,93928.1North Dakota8,84646.4California12,9930.343
362931Arkansas989,454,10327.3Georgia66,75546.2Idaho12,4250.261
373724Kentucky1,342,891,91627.2Kentucky47,68145.3Alabama9,1540.244
383438Mississippi1,175,187,74614.7Indiana64,19741.9Mississippi9,1720.208