Interstate Variation and Progress Toward Balance in Use of and Expenditure for Long-Term Services and Supports in 2009. 3. Subgroup Differences

03/18/2014

As rates of HCBS use vary considerably across subgroups of LTSS users, policymakers may be particularly interested in the relationship between state factors and HCBS use for subpopulations of enrollees. Because elderly recipients generally make up a large portion of those in the LTSS system, they dominate the overall results. One question, however, is whether state factors are linked to LTSS system performance for those who are under age 65 and have physical disabilities, or for those with ID/DD.

As we found in 2006, the relationship between state factors and HCBS balance differs by subpopulation of enrollee (Table III.5). Most notably, some factors that were significantly associated with balance for the ID/DD population were not significant for other populations. For individuals with ID/DD, the number of people on waiver waiting lists, the relative size of the potentially eligible Medicaid aged population, and total taxable resources were significantly associated with HCBS use and expenditures. In comparison, consumer-direction rates, the availability of home health aides, and the availability of assisted living and residential care units were associated with higher levels of HCBS expenditures and use for the aged and those with physical disabilities, but were not significant for the ID/DD population. Such differences may point to the different needs of these subpopulations and reinforce the importance of targeting programs and policies to meet the needs of diverse populations of LTSS users.


TABLE III.3. Summary of State Constraints by State Rank in the Percentage of LTSS Expenditures for HCBS in 2009

Factor Mean for
  All States  
  Mean for Top 10  
(high HCBS)
Ranked States
Mean for
  Mid-Ranked  
Stats
Mean for
Bottom 10
(low HCBS)
  Ranked States  
  Relationship with  
Higher Levels
of HCBS
in 2006
  Expected  
(2009)
  Observed  
(2009)
Single-family housing price index, 2009 207 234 198 195 + - +*
Per-capita personal income, 2009 43,913 56,951 39,264 29,245 + + +
Average winter precipitation, 1971-2000 2.8 2.5 2.4 3.6 -* - -
Taxable resources per-capita, 2009 50,935 59,465 47,757 48,125 + + +*
LTSS expenditures as a share of total state spending, 2009 7.0% 7.0% 6.9% 7.2% NA - -
Percentage of potential eligibles age 75 or older, 2009 11% 10% 11% 13% none + -*
Home health aides per 1,000 elderly or persons with a disability, 2009 85 101 90 59 + + +
Personal and home care aides per 1,000 elderly or persons with a disability, 2009 63 112 55 31 +* + +*

SOURCE: Mathematica analysis of state constraints (see Table III.1) and 2009 MAX data for 37 states and the District of Columbia with representative LTSS data.

* Significant association at the 0.05 level. For continuous factors, we tested whether the correlation between balance and the constraint was significantly different from zero. For discrete factors, we used a t-test to identify significant differences between states with and without the constraint. We did not test for significant differences between top and bottom states.
NA = measure was not assessed for 2006 or used different data specifications.


TABLE III.4. Summary of State Policy and Supply-Side Variables by the Percentage of LTSS Expenditures for HCBS in 2009

Policy of Supply-Side Factor Mean for
  All States  
  Mean for Top 10  
(high HCBS)
Ranked States
Mean for
  Mid-Ranked  
Stats
Mean for
Bottom 10
(low HCBS)
  Ranked States  
  Relationship with  
Higher Levels
of HCBS
in 2006
  Expected  
(2009)
  Observed  
(2009)
Number of people consumer-directing services per 1,000 adults age 18+, 2010 14.2 36.3 7.2 4.7 NA + +*
Medicaid state plan personal care coverage, 2009 63% 70% 72% 40% +* + +
Any coverage for residential care, 2009 89% 100% 89% 80% +* + +
Waiver waiting list (2011) per HCBS users (2009) 374 71 611 218 - - -
State-administered optional SSI supplementation, 2009 76% 80% 83% 60% +* + +
Nursing home occupancy rates, 2009 84 87 81 86 NA - +
Percentage of total out-of-home placements in settings for 6 or fewer persons, 2011 76% 90% 74% 65% NA + +*
Assisted living and residential care units per 1,000 people age 65, 2010 29 32 31 23 NA + +*
ADRC/SEP functionality score, 2010 6.9 6.6 6.7 7.6 NA + -
Home health aide hourly rate, 2009 21 22 20 20 NA + +
Adult day service rate, 2009 68 86 60 66 NA + +*
Percentage of caregiving families receiving state ID/DD agency support, 2011 11% 16% 10% 8% NA + +

SOURCE: Mathematica analysis of state policy or supply-side factors (see Table III.2) and 2009 MAX data for 37 states and the District of Columbia with representative LTSS data.

* Significant association at the 0.05 level. For continuous factors, we tested whether the correlation between the performance indicator and the factor was significantly different from zero. For discrete factors, we used a t-test to identify significant differences between states with and without the policy. We did not test for significant differences in rank or between top 10 and bottom 10 states.
NA = measure was not assessed for 2006 or used different source data or specifications.


TABLE III.5. Association of State Policies and Other Factors with Spending for HCBS in 2009, by Age and System Type

Constraint, Policy, or Supply-Side Factor   Expected     Overall     Aged  
(65+)
Enrollees with
Disabilities <65,
  Excluding ID/DD  
Enrollees
  with ID/DD  
Single-family housing price index, 2009 - 0.35 --- --- ---
Per-capita personal income, 2009 + --- 0.35 --- ---
Average winter precipitation, 1971-2000 - --- --- --- ---
Taxable resources per-capita, 2009 + 0.36 0.36 --- 0.35
Percentage of potential Medicaid eligibles age 75 or older, 2009 + -0.57 --- --- -0.56
Home health aides per 1,000 elderly or persons with a disability, 2009 + --- 0.42 0.38 ---
Personal and home care aides per 1,000 elderly or persons with a disability, 2009 + 0.50 --- 0.46 ---
Number of people consumer-directing services per 1,000 adults age 18+ with disabilities + 0.53 0.52 0.34 ---
Medicaid state plan personal care, 2009 + --- 0.34 --- ---
Any coverage for residential care, 2009 + --- --- --- ---
Persons on all waiver waiting lists (2011), per HCBS users (2009) - --- --- --- ---
Persons on aged and disabled waiver waiting list (2011), per aged and disabled HCBS users (2009) - --- --- --- NA
Persons on ID/DD waiver waiting list (2011), per ID/DD HCBS users (2009) - --- NA NA -0.35
State-administered optional SSI supplementation, 2009 + --- --- --- ---
Nursing home occupancy rates, 2009 - --- --- --- ---
Percentage of total out-of-home ID/DD placements in settings for 6 or fewer persons, 2011 + 0.67 NA NA 0.72
Assisted living and residential care units per 1,000 people age 65, 2010 + 0.40 0.34 0.35 ---
ADRC/SEP functionality score, 2010 + --- --- --- ---
Home health aide hourly rates, 2009 + --- --- --- ---
Adult day service rates, 2009 + 0.33 --- --- ---
Percentage of caregiving families receiving state ID/DD agency support, 2011 + --- NA NA ---

SOURCE: Mathematica analysis of state constraints, policy, and supply-side factors (see Table III.1 and Table III.2) and 2009 MAX data for 37 states and the District of Columbia with representative LTSS data. Analysis of enrollees with disabilities under 65 with ID/DD and those with other disabilities include 35 states (individuals with ID/DD could not be identified in the District of Columbia, Vermont, or Washington and these states were excluded from analyses of this population).
NOTE: Values in table represent the correlation coefficient between the factor and HCBS share of LTSS expenditures. All values shown are significant at the 0.05 level. For continuous factors, we tested whether the correlation between the performance indicator and the factor was significantly different from zero. For discrete factors, we used a t-test to identify significant differences between states with and without the factor. We did not test for significant differences in rank or between top 10 and bottom 10 states.

--- = no significant relationship was found.
NA = factor was not relevant for subgroup.


Correlations summarize the overall nature and strength of relationships between state factors and LTSS balance, but these associations provide limited insight into the underlying composition of these relationships. Figure III.1, Figure III.2 and Figure III.3 illustrate the differences in significant relationships between balance of HCBS expenditures and three state factors. These figures show the relative importance outliers have on these associations and how different patterns can lie beneath the overall relationships seen in the data. Understanding the nuances of how policies are related to LTSS system balance may be useful for policymakers who are interested in using these approaches to shift LTSS systems toward HCBS.


FIGURE III.1. Percentage of Out-of-Home Placements for Individuals with ID/DD in Settings for Six or Fewer People in 2011 by Percentage of Medicaid LTSS Expenditures for HCBS, 2009

FIGURE III.1. Percentage of Out-of-Home Placements for Individuals with ID/DD in Settings for Six or Fewer People in 2011 by Percentage of Medicaid LTSS Expenditures for HCBS, 2009

SOURCE: Mathematica analysis of MAX 2009 data for 35 states with representative LTSS data; University of Colorado 2013.

ALT TEXT for FIGURE III.1, Scatter Plot: Shows the state abbreviations labeling each data point. It shows the relationship between the percentage of total out-of-home placements in settings for 6 or fewer people (Y-axis) and HCBS share of spending for individuals with ID/DD (X-axis). In 2011, out-of-home placements in small ID/DD facilities ranged considerably, from 37% of placements in Mississippi to 98% of placements in Alaska and Washington. Similarly, rates of HCBS expenditures as a share of LTSS costs ranged from 15% in Mississippi to 99% in New Hampshire. As the figure shows, states with the highest rates of placements in small out-of-home settings also tended to have the highest rates of HCBS spending, with somewhat more variation in the relationship between small-facility placement and HCBS spending for middle-ranked states. Although there is some variation in the link between HCBS spending and small-facility placements, overall it appears that, throughout the range, the 2 are positively associated.


Figure III.1 shows one of the strongest relationships we identified, between the percentage of total out-of-home placements in settings for six or fewer people and HCBS share of spending for individuals with ID/DD. In 2011, out-of-home placements in small ID/DD facilities ranged considerably, from 37 percent of placements in Mississippi to 98 percent of placements in Alaska and Washington. Similarly, rates of HCBS expenditures as a share of LTSS costs ranged from 15 percent in Mississippi to 99 percent in New Hampshire. As the figure shows, states with the highest rates of placements in small out-of-home settings also tended to have the highest rates of HCBS spending, with somewhat more variation in the relationship between small-facility placement and HCBS spending for middle-ranked states. Although there is some variation in the link between HCBS spending and small-facility placements, overall it appears that, throughout the range, the two are positively associated.


FIGURE III.2. Percentage of Medicaid LTSS Expenditures for HCBS in 2009 by the Number of People Consumer-Directing Services in 2010

FIGURE III.2. Percentage of Medicaid LTSS Expenditures for HCBS in 2009 by the Number of People Consumer-Directing Services in 2010

SOURCE: Mathematica analysis of MAX 2009 data for 37 states and the District of Columbia with representative LTSS data; Reinhard et al. 2011.

ALT TEXT for FIGURE III.2, Scatter Plot: Shows the state abbreviations labeling each data point. It shows the relationship between Number of People Consumer-Directing Services per 1,000 people Age 18+ with Disabilities, 2010 and the Percentage of Medicaid LTC Expenditures for HCBS, 2009. The HCBS share of spending appears to be driven by a small number of states that have very large numbers of consumers directing services. Most states reported rates from 0 to 20 people consumer-directing services per 1,000 adults age 18 and older with disabilities. The 5 states with the highest overall shares of HCBS spending in 2009 (Alaska, California, Colorado, Vermont, and Washington) reported much higher rates of consumer-direction, from a rate of 44.7 (per thousand people with disabilities) in Colorado to 142.7 in California. When these states are removed from the analysis, the correlation coefficient drops from 0.53 to 0.20. Thus, there appears to be a particularly strong relationship between very high rates of consumer direction and very high rates of balance toward HCBS spending, with less of a relationship between moderate levels of consumer direction and balance.


In comparison, Figure III.2 shows how the positive correlation between consumer-direction and HCBS share of spending appears to be driven by a small number of states that have very large numbers of consumers directing services. Most states reported rates from 0 to 20 people consumer-directing services per 1,000 adults age 18 and older with disabilities. The five states with the highest overall shares of HCBS spending in 2009 (Alaska, California, Colorado, Vermont, and Washington) reported much higher rates of consumer-direction, from a rate of 44.7 (per thousand people with disabilities) in Colorado to 142.7 in California. When these states are removed from the analysis, the correlation coefficient drops from 0.53 to 0.20. Thus, there appears to be a particularly strong relationship between very high rates of consumer-direction and very high rates of balance toward HCBS spending, with less of a relationship between moderate levels of consumer-direction and balance.

Finally, Figure III.3 examines the relationship between waiver waiting lists and share of LTSS expenditures for HCBS among individuals with ID/DD. We found a negative relationship between the number of people on waiting lists for ID/DD waivers and the share of HCBS expenditures. The figure shows how this negative relationship is affected by 11 states that reported no one on waiting lists for these waivers in 2009 combined with very large waiting lists in a few states with relatively low HCBS spending for this population (including Illinois, Indiana, Ohio, and Texas). When the states with no waiting list are removed from the analysis, the correlation changes from -0.35 to -0.45, suggesting that, among the states with waiting lists for ID/DD waivers, longer lists may be even more closely related to lower rates of HCBS spending than the all-state correlation suggests.19 In this case, the relationship between waiting lists and the composition of LTSS expenditures was biased downward by the presence of states that did not use waiting lists at all.


FIGURE III.3. Percentage of Medicaid LTSS Expenditures for HCBS in 2009 by the Number of People on Waiting Lists for ID/DD Waivers Per ID/DD Waiver Users in 2011

FIGURE III.3. Percentage of Medicaid LTSS Expenditures for HCBS in 2009 by the Number of People on Waiting Lists for ID/DD Waivers Per ID/DD Waiver Users in 2011

SOURCE: Mathematica analysis of MAX 2009 data for 35 states with representative LTSS data. KFF 2012.

ALT TEXT for FIGURE III.3, Scatter Plot: Shows the state abbreviations labeling each data point. It examines the relationship between waiver waiting lists and share of LTSS expenditures for HCBS among individuals with ID/DD. We found a negative relationship between the number of people on waiting lists for ID/DD waivers per ID/DD Waiver User and the share of HCBS expenditures among Individuals with ID/DD. The figure shows how this negative relationship is affected by 11 states that reported no one on waiting lists for these waivers in 2009 combined with very large waiting lists in a few states with relatively low HCBS spending for this population (including Illinois, Indiana, Ohio, and Texas). When the states with no waiting list are removed from the analysis, the correlation changes from -0.35 to -0.45, suggesting that, among the states with waiting lists for ID/DD waivers, longer lists may be even more closely related to lower rates of HCBS spending than the all-state correlation suggests. In this case, the relationship between waiting lists and the composition of LTSS expenditures was biased downward by the presence of states that did not use waiting lists at all.


19 When the states with the largest waiting lists (Illinois, Indiana, Ohio, and Texas) are removed from the analysis, the correlation changes only slightly.

View full report

Preview
Download

"ProgBal.pdf" (pdf, 1.6Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®