| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
|
|
|
|||||||
|
|||||||||
Correspondence: Address correspondence to Nancy A. Miller, PhD, Department of Public Policy, University of Maryland, Baltimore County, Public Policy Building, Baltimore, MD 21250. E-mail: nanmille{at}umbc.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Home- and community-based care • Long-term care • State variation • State policy • Medicaid
States appear to have used the waiver program to extend community-based care to individuals with intellectual and developmental disabilities (IDDs) to a greater extent than to older and working-age individuals with primarily physical disabilities (O/WAIs). In 2001, of total Medicaid long-term-care expenditures on the IDD population (1915c waiver and expenditures of intermediate care facilities for the mentally retarded), waivers consumed 58% on average. In contrast, among Medicaid long-term-care expenditures on the O/WAI population (through 1915c waivers and skilled nursing facilities), waivers constituted only 11%. (This is our calculation, made on the basis of information from the Centers for Medicare & Medicaid Services [CMS], form 64.)
Previous empirical work has examined state-level factors related to rates of use and expenditures for Medicaid 1915c waiver services (Harrington, Carrillo, Wellin, Miller, & LeBlanc, 2000; Kitchener, Carrillo, & Harrington, 2004; Miller, Ramsland, Goldstein, & Harrington, 2001). Long-term-care supply is an important predictor of community-based-care use and expenditures. States with a greater availability of community-based-care providers (e.g., residential care facilities) and a more limited availability of nursing home facilities are found to have higher rates of use and expenditures for community-based care, all else being equal. Greater fiscal resources, as measured by state per capita income, are also a robust predictor of increased use and expenditures for community-based long-term care.
The empirical work has focused on use and expenditures across all segments of the population of people with disabilities. However, predictors of state investment might vary by target group (Braddock, 1992). Our purpose in this research is to compare state-level factors that are associated with states' provision of community-based care to O/WAIs relative to IDDs through the Medicaid 1915c waiver program. Prior to introducing the analytic approach, we begin with a brief overview of the 1915c waiver program.
Medicaid 1915c Program Structure
The Medicaid home- and community-based-care services waiver program was established under section 2176 of Public Law 97-35 (Omnibus Budget Reconciliation Act of 1981), which added section 1915c to the Social Security Act for the Medicaid program. This section authorizes the CMS to waive certain Medicaid statutory limitations to allow states to provide community-based services that individuals with disabilities may need to avoid institutional care. States may provide a range of nonmedical services, such as case management and personal care, as well as medical services, such as nursing care. States use this flexibility to tailor specific services to subgroups of the long-term-care population (Miller, Ramsland, & Harrington, 1999). States can define the target group being served, services provided under a specific waiver, and the geographic area in which services are provided. The ability to waive statewide requirements and comparability of services, combined with the enrollment cap set for each waiver, afford states the opportunity to manage growth in the number of participants served, as well as expenditures (Kitchener, Ng, & Harrington, 2004). In limiting the 1915c waiver to one or more eligibility groups (e.g., working-age individuals with physical impairments), states also may modify certain functional and financial eligibility requirements.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Variables
Independent Variables
We examine features of the state environment associated with long-term-care use and expenditures, such as demand (sociodemographic characteristics) and supply. We add to these utilization factors a measure of state resources, as well as state policy and programmatic features that may influence community-based-care use and expenditures.
Certain sociodemographic characteristics are expected to increase long-term-care use and expenditures. The age of a population is a strong predictor of formal home-care services (e.g., Houde, 1998; Kitchener, Carrillo, et al., 2004; Miller, Harrington, & Goldstein, 2002). Similarly, rates of disability among working-age individuals within a state can be expected to relate to both use and expenditures for long-term care. Other characteristics, such as race or ethnicity, may influence the probability of use of different forms of long-term care (Cagney & Agree, 1999; Miller et al., 2001; Murtaugh, Kemper, & Spillman, 1990; Wallace, Levy-Storms, Kington, & Andersen, 1998). In addition, there is some evidence that urban residents are more likely to use formal home health than rural residents (Kenney, 1993; Nyman, Sen, Chan, & Commins, 1991), although more recent research has found mixed effects (Kitchener, Carrillo, et al.; Miller, Harrington, & Goldstein, 2002).
We include five demand measures—the share of the population age 85 or older (for O/WAI analyses), the per capita number of working-age individuals with disabilities eligible for Supplemental Security Income (SSI), the share of the population that is African American, the share of the population residing in a metropolitan area, and the state poverty rate. (We use SSI as a proxy for state-level demand among working-age individuals. State-level disability rates for working-age individuals are only available for 1990 and 2000 in the U.S. Census data.)
The increased availability of community-based-care supply (e.g., residential care beds) has been shown to influence community-based-care use and expenditures (Harrington et al., 2000; Kitchener, Carrillo, et al., 2004). In contrast, the greater availability of institutional bed supply has been associated with lower rates of use and expenditures for community-based care (Harrington et al.; Kitchener, Carrillo, et al.; Miller, Harrington, & Goldstein, 2002), as well as the share of dollars supporting community-based care (Miller et al., 2001; Miller, Harrington, & Goldstein). An increase in female labor-force participation is expected to reduce the supply of informal caregivers, increasing the use and hence expenditures for formal long-term care, both noninstitutional and institutional (Houde, 1998; Kitchener, Carrillo, et al.; Miller, Harrington, & Goldstein).
Four factors associated with supply are included in our estimations for O/WAI participants and expenditures—per capita nursing home beds, per capita residential care facility beds, per capita certified home-health agencies, and the female labor-force-participation rate. Additionally, we developed a measure to capture the relative change in per capita nursing home bed supply, as many view reduction in institutional capacity important to states' efforts to expand community-based care (Coleman, Kassner, & Pack, 1996; Ladd & Associates, 1997; Wiener & Stevenson, 1998). To measure the relative change, we calculated the percentage of change in per capita nursing home bed supply for the 10-year period for each state. We used similar supply measures for the estimations for IDDs, using intermediate care facility beds per capita in place of nursing facility beds, and the per capita bed change, constructed as the previously described relative change measure for nursing home bed supply.
State economic resources are associated with increased use and expenditures, as higher per capita income in a state is associated with increased funding efforts (Braddock, 1992; Kitchener, Carrillo, et al., 2004; Miller, Harrington, & Goldstein, 2002). The state per capita income is used as a measure of state fiscal resources.
State programmatic and policy actions also have been examined. The regulation of institutional bed supply is a long-standing strategy to constrain expenditures (Coleman et al., 1996; Grabowski, Ohsfeldt, & Morrisey, 2003; Harrington, Swan, Nyman, & Carrillo, 1997). Reduced institutional expenditures may be redirected to supporting community-based-care services. Some researchers have found a positive relationship between use of a certificate of need or moratorium for nursing facilities and Medicaid community-based-care expenditures (Miller, Harrington, & Goldstein, 2002), as well as a larger share of dollars supporting community-based care (Miller et al., 2001). The effect has not been consistent across studies, however (Harrington et al., 2000; Kitchener, Carrillo, et al., 2004). Alternatively, constraints on community-based care supply have been negatively associated with the share of long-term-care dollars supporting community-based long-term care (Miller, Harrington, & Goldstein) and the share of long-term-care dollars supporting 1915c waivers (Miller, Harrington, Ramsland, & Goldstein, 2002). We include measures of both nursing home bed and home-health-agency regulation (i.e., use of certificate of need or moratorium) in our estimations for O/WAI participants and expenditures. We utilize measures of intermediate care facility and home-health regulation in our estimations of IDD use and expenditures.
Following a policy described as Medicare maximization, states may pursue the use of Medicare home-health services to substitute for Medicaid services for individuals eligible for both programs. Although primarily a cost-containment strategy, it also may represent a policy to expand home- and community-based service access in a state by pursuing additional public-funding sources such as Medicare (Miller, in press). Previous research has been mixed. Some studies have demonstrated a complementary relationship between Medicare- and Medicaid-funded community-based-care expenditures as well as the share of long-term-care dollars supporting community-based care (Miller, Harrington, & Goldstein, 2002; Miller et al., 2001). Earlier research reported some evidence for Medicare substitution (Cohen & Tumlinson, 1997). The time frame analyzed in these studies was prior to the Balanced Budget Act (BBA) of 1997. Changes were made to the Medicare-payment methodology for home-health agencies and have been associated with a decline in the number of home-health agencies as well as expenditures (Levit et al., 2004). It may be that, for those individuals eligible for both programs, Medicaid services may have substituted for Medicare services in the years immediately following enactment of the BBA. We include two measures in our analyses. First, we include the number of Medicare home-health users per 1,000 Medicare beneficiaries in a state. To examine the BBA effect, we include an interaction between the users per capita and a dummy variable for years that followed the BBA (i.e., 1998–2000).
Dependent Variables
For both waivers serving O/WAIs and IDDs, we examine per capita beneficiaries using services as well as per capita expenditures for these waivers for each of the two target groups. We also examine the share of long-term-care dollars supporting 1915c waiver services. For waivers supporting O/WAIs, we combine skilled nursing facility and waiver expenditures for O/WAIs and then examine the waiver share of this total. For waivers serving IDDs, we combine intermediate care facility and waiver expenditures for IDDs and then examine the waiver share of this total.
Data Sources
We collected data on waiver participants and expenditures from CMS form 372, as described in Kitchener, Ng, and Harrington (2003). States report participant use, services, and expenditures on this report for each state waiver. We obtained nursing home expenditures, used to develop the measure of the share of long-term-care dollars supporting 1915c waivers, from CMS form 64 (www.HCBS.org), a report submitted by the states to the CMS for payment purposes that reports federal and total expenditures by date of payment for each federal fiscal year. Although used as the basis for payment, the reports may be amended by states at a later date, and some states amend their data. In addition, audits by the CMS may lead to some changes in the data but do not lead to adjustments in previously submitted reports. However, this is the only available data source for institutional spending. We also obtained expenditures for intermediate care services from this source. We obtained sociodemographic variables, female labor-force-participation rate, and state per capita income from the U.S. Census Bureau. We obtained information regarding bed supply and supply regulation from an annual survey of state officials collected by the University of California, San Francisco and Wichita State University (see Harrington, Anzaldo, Burdin, Kitchener, & Miller, 2005; Harrington, Chapman, Miller, Miller, & Newcomer, in press). We obtained home-health-use data from annual Statistical Supplements prepared by the CMS.
Estimation Techniques
We used multivariate regression analysis to explore the factors significantly associated with the dependent variables of beneficiary use, expenditures, and relative dollar investment in 1915c waivers. We used a fixed-effects panel model. This approach explicitly models state-level variables that may be fixed over time and that are not included as explanatory variables in the analyses. A fixed-effects model is also appropriate when the data exhaust the population, such as our use of all states in the anal-yses (Woolridge, 2003). The models took the following form:
|
|
Here, a stands for constant, E stands for error, i stands for state, t stands for year. To test for potential endogeneity of the certificate-of-need regulatory policies with state rates of use and expenditures, we followed a procedure suggested in Vita (2001) and employed by Grabowski and colleagues (2003) in their analyses of the effects on certificate-of-need repeal on Medicaid nursing home and long-term-care expenditures. Specifically, we created a dummy variable set equal to one in the year preceding the enactment of the regulatory policy. If the regulatory change is endogenous, then the coefficient of this lead variable should be positive. The lead variables we tested for nursing home and intermediate care bed supply regulation, as well as home-health-agency regulation, were not statistically significant in the models, suggesting that these regulatory variables are not endogenous.
The remaining independent variables are a set of sociodemographic, supply, and economic variables that we expect to influence long-term-care use and spending, as previously discussed. As they are not policy variables per se, we did not test for endogeneity as we did for the regulatory policies. We lagged these state-level variables by 1 year (e.g., we used 1991 data on state per capita income to predict 1992 use and expenditures). We used the independent variables to predict the dependent variables in the following year because we expected that their effects would require 1 year to have an impact.
We logged the six dependent variables, because a test of the normality of the distribution for each indicated that the distribution was skewed. We standardized the number of beneficiaries served, expenditures, working-age SSI participants, and bed or agency supply by 1,000 state population to facilitate comparisons across states. We expressed all dollars in 2001 dollars, using the Consumer Price Index. We used Stata, Release 8, to conduct the analyses.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
Share of Long-Term-Care Dollars Supporting Waiver Services
A higher state poverty rate was associated with a smaller share of dollars supporting community-based care. As with the previous measures of use and expenditures, higher per capita supply of home-health agencies was positively associated with the share of dollars supporting waiver services. State per capita income was positively related to provision of greater dollars in community-based settings. The share of dollars supporting 1915c waiver services was negatively related to the number of Medicare home-health users.
1915c IDD Waivers
Rates of use and expenditures increased over the study period (Table 3). In 1992, 0.19 participants were served per capita, on average, increasing to 1.17 participants per capita in 2001. Expenditures increased from $11,459 in 1992 to $46,941 per capita in 2001 (in constant dollars). On average, states more than doubled the share of dollars targeted to 1915c waiver services, increasing from 24% to 58% over the study period. Forty-one states and the District of Columbia reduced per capita intermediate care bed supply over the study period. Regulation of intermediate care bed supply declined somewhat over the study period.
|
|
Share of Long-Term-Care Dollars Supporting IDD Waiver Services
Whereas an increase in the share of the state population residing in metropolitan areas was associated with relatively more dollars supporting community-based care, a higher poverty rate in a state was associated with relatively fewer dollars supporting such care. An increase in the per capita home-health-agency supply and an increase in the female labor-force participation were positively related to the share of dollars supporting waivers for IDDs. There was a negative association between residential care bed supply and the share of dollars supporting waiver services. Income was again positively related to the share supporting waiver services, as was intermediate care bed regulation.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We found long-term-care supply to be a predictor of home- and community-based service use and spending for both groups. Supply variables alone accounted for between 3.2% and 7.5% of total variance in rates of use and expenditures for O/WAI waivers. Supply variables accounted for between 13.6% and 34.4% of total variance in analyses of use and expenditures for IDD waivers (data not shown). We included two institutional supply measures—annual measures of per capita institutional bed supply and the 10-year change in per capita institutional bed supply. In models limited to supply variables, growth in institutional supply was negatively related to all IDD measures. Growth in nursing home bed supply was not related to any O/WAI measure in similar models limited to supply measures. These findings suggest that moderation of institutional bed supply may have played a greater role in the expansion of community-based long-term care for IDDs. Forty-one states and the District of Columbia reduced per capita intermediate care facility bed supply over the study period, and bed capacity in 2000, on average, was 71% of 1991 bed supply. In contrast, 27 states reported an increase in per capita skilled nursing facility bed capacity. Skilled nursing bed capacity in 2000, on average, was 99% of the 1991 bed supply.
The effects of institutional bed supply regulation were mixed. Intermediate care bed supply regulation was positively associated with both expenditure measures, as we expected. However, skilled nursing bed supply regulation had no effect on the use and expenditure measures. Other researchers have found states' use of nursing home bed certificate of need or moratoria to be associated with greater per capita community-based-care expenditures (i.e., 1915c waivers, home health, and personal care), a larger share of dollars supporting community-based care (Miller, Harrington, & Goldstein, 2002), and a larger share of long-term-care dollars supporting 1915c waivers (Miller et al., 2001). Some researchers have found no effect of supply regulation on rates of use and expenditures for either all community-based-care spending (Kitchener, Carrillo, et al., 2004) or waiver expenditures across all target groups (Harrington et al., 2000; Kitchener, Carrillo, et al.; Miller et al., 2001). Improving our understanding of the relationships among supply regulation, other mechanisms to moderate supply, per capita supply, and provision of community-based long-term care appears to be important.
Home-health-agency supply was a positive predictor of rates of use, expenditures, and the share of dollars supporting community-based care for O/WAIs, and residential care bed supply was also related to per capita use and expenditures. Home-health-agency supply was a positive, but less robust, predictor for IDD waiver use and expenditures. However, residential care bed capacity was negatively related to both expenditure measures. Previous studies have suggested the importance of community-based supply (Harrington et al., 2000; Kitchener, Carrillo, et al., 2004). We cannot assume through our analyses a causal relationship between expanded home- and community-based service supply and use. It is possible that increased demand leads to increased supply. However, state case studies and surveys identify capacity building as a state policy to increase service access and use (Coleman et al., 1996: Fox-Grage, Folkemer, Straw, & Hansen, 2002; Ladd & Associates, 1997; Murtaugh et al., 1999). The CMS Systems Change grants include funding for infrastructure development, similarly suggesting the importance of state capacity building. In addition, over 55,000 O/WAIs were on waiver waiting lists in 2002, as were over 80,000 IDDs (Kitchener, Ng, et al., 2004), suggesting that demand may not be leading supply in most states. Efforts to enhance the availability of home- and community-based care supply, particularly through home-health agencies, appear to be important.
A greater female labor-force-participation rate was associated with all measures of use and spending for waivers serving IDDs but was not related to measures of use and expenditures for O/WAIs. This suggests that the female labor-force decision may more strongly influence informal caregiving practices for IDDs. Although many caregivers of older individuals are older themselves and no longer in the labor force, caregivers of IDDs may more often be working aged. In addition, demand variables were a more consistent predictor of IDD waiver use and spending.
State resources, as measured by average per capita income, were a robust positive predictor of utilization and expenditures for waivers serving both groups. This finding is consistent with previous studies of Medicaid community-based long-term-care spending (Kitchener, Carrillo, et al., 2004; Miller et al., 2001; Miller, Harrington, & Goldstein, 2002) as well as broader studies of Medicaid expenditures (Braddock, 1992; Buchanan, Cappellini, & Ohsfeldt, 1991; Schneider & Jacoby, 1995). Some studies have found that states with the fewest financial resources also may experience greater demand for long-term care (Miller, Harrington, & Goldstein). For example, the 10 states associated with the lowest per capita expenditures on home- and community-based services in 1997 had a larger share of individuals in their population who were aged 85 and older, a higher rate of disability among the working aged, and a greater share of African Americans (for whom rates of disability are higher), yet fewer resources on average than the 10 states reporting the highest per capita service expenditures. Although the federal matching assistance percentage provides relatively greater dollars to states with a lower average per capita income, it does not adjust for the state-level variation in long-term-care demand, which can be substantial. Policy efforts to address state fiscal resource issues are key to expanding access. Grants to states provide one mechanism. A change to the federal matching assistance percentage that incorporates state differences in long-term-care demand also could address state resource concerns.
In analyses focused on O/WAIs, Medicare and Medicaid appear to act as substitutes for individuals eligible for both. The relationship between Medicare and Medicaid for IDDs was positive for per capita rates of use but was not associated with the expenditure measures. Continuing to examine the relationship between Medicare and Medicaid home- and community-based services and their impact on access for those eligible for both programs is important.
States are increasing their provision of community-based long-term care through Medicaid 1915c waivers. States have used the 1915c waiver program to more effectively transform systems of care for IDDs, relative to O/WAIs. Our findings suggest that increased community-based-care capacity is important to this effort. Our findings were less definitive on the role of institutional supply moderation in supporting increased community-based long-term care. Institutional bed supply moderation may play a more important role in fostering community-based care for IDDs, and this possibility warrants additional study. Our research suggests that effort on the part of states to enhance community-based-care capacity, coupled with targeted federal fiscal support, may facilitate further development of community-based systems of long-term care across states.
|
Footnotes |
|---|
1 Department of Public Policy, University of Maryland, Baltimore County.
2 Department of Social and Behavioral Sciences, University of California, San Francisco.
3 Arnold School of Public Health, University of South Carolina, Columbia.
4 Intercampus Doctoral Program in Gerontology, University of Maryland, Baltimore.
Decision Editor: Linda S. Noelker, PhD
Received for publication November 5, 2004. Accepted for publication July 21, 2005.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. A. Miller, K. T. Elder, M. Kitchener, Yu Kang, and C. Harrington Medicaid 1915(c) Waiver Use and Expenditures for Persons Living With HIV/AIDS Med Care Res Rev, June 1, 2008; 65(3): 338 - 355. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| All GSA journals | Journals of Gerontology Series A: Biological Sciences and Medical Sciences | Journals of Gerontology Series B: Psychological Sciences and Social Sciences |
