An Investigation of Interstate Variation in Medicaid Long-Term Care Use and Expenditures Across 40 States in 2006. B. Factors and Policies Associated with Long-Term Care System Performance

07/01/2013

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

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

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

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

TABLE III.3. Summary of State Constraints by State Rank in the Percentage of LTC Expenditures for HCBS in 2006
Factor Mean for
  All States  
State Rank in the Percentage of
LTC Expenditures for HCBS in 2006
  Expected     Observed  
Mean for Top
  10 (High HCBS)  
Ranked States
Mean for
  Mid-Ranked States  
  Mean for Bottom  
10 (Low HCBS)
Ranked States
Single-family housing price index, 2006 367 405 372 319 - +
Per-capita personal income, 2006 35,369 36,929 35,424 33,701 + +
Average winter precipitation, 1971-2000 2.8 2.0 2.8 3.7 - -*
Taxable resources per-capita, 2006 50,626 54,088 50,477 47,462 + +
Percentage of potential eligibles age 75 or older, 2006 26% 26% 28% 24% + none
Home health aides per 1,000 elderly or persons with a disability, 2009 84 103 83 68 + +
Personal and home care aides per 1,000 elderly or persons with a disability, 2009 67 129 50 40 + +*
SOURCE: Mathematica analysis of state constraints (see Table III.1) and 2006 MAX data for 39 states and the District of Columbia with representative LTC data.

* Significant association at the 0.05 level, one-tailed test. 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 10 and bottom 10 states.

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

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

TABLE III.4. Summary of State Policy and Supply-Side Variables by the Percentage of LTC Expenditures for HCBS in 2006
Policy or Supply-Side Factor Mean for All States State Rank in the Percentage of LTC Expenditures for HCBS in 2006   Expected     Observed  
Mean for Top
  10 (High HCBS)  
Ranked States
Mean for
  Mid-Ranked States  
  Mean for Bottom  
10 (Low HCBS)
Ranked States
Any consumer direction, 2006 78% 100% 80% 50% + +*
Stricter functional limits for HCBS waivers than nursing facilities, 2006 15% 20% 10% 20% - none
Personal care services in state plan, 2006 58% 80% 60% 30% + +*
Any coverage limits for home health care, 2006 35% 20% 35% 50% - -
Any coverage for residential care, 2003 85% 100% 85% 70% + +*
Waiver waiting list per 1,000 HCBS enrollees 142 102 103 293 - -
Higher SSI supplement for community living, 2006 77% 90% 85% 50% + +*
Maximum days bed hold, 2000 8.8 8.4 8.8 9.2 - -
Nursing home beds per 1,000 elderly, 2003 52 43 53 59 - -*
Percentage of ICFs/IID with 16 or more beds, 2006 35% 28% 42% 28% - -
Medicaid payment per day for nursing facility care, 2007 161 181 159 146 - +*
Medicare reimbursement per home health aide visit, 2006 140 145 141 136 + +*
Average private pay daily rate for adult day care, 2008 56 63 55 52 + +
SOURCE: Mathematica analysis of state policy or supply-side factors (see Table III.2) and 2006 MAX data for 39 states and the District of Columbia with representative LTC data.

* Significant association at the 0.05 level, one-tailed test. 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.

3. Subgroup Differences

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

TABLE III.5. Summary of Select State Measures by the Percentage of LTC Expenditures for HCBS in 2006, Overall and 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, 2006 -   +*    
Per-capita personal income, 2006 +       +*
Average winter precipitation, 1971-2000 - -*      
Taxable resources per-capita, 2006 +       +*
Home health aides per 1,000 elderly or persons with a disability, 2009 +   +* +*  
Personal and home care aides per 1,000 elderly or persons with a disability, 2009 + +* +* +* +*
Any consumer direction, 2006 + +* +* +*  
Personal care services in state plan, 2006 + +* +*    
Any coverage for residential care, 2003 + +* +*    
Higher SSI supplement for community living, 2006 + +* +*   +*
Nursing home beds per 1,000 elderly, 2003 - -* -*   -*
Percent of ICFs/IID with 16 or more beds, 2006 -   -*    
Medicaid payment per day for nursing facility care, 2007 - +* +*   +*
Medicare reimbursement per home health visit, 2006 + +* +*    
Average private pay daily rate for adult day care, 2008 +   +*    
SOURCE: Mathematica analysis of state constraints, policy, and supply-side factors (see Table III.1 and Table III.2) and 2006 MAX data for 39 states and the District of Columbia with representative LTC data.

* Significant association at the 0.05 level, one-tailed test. 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.

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

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

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

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

View full report

Preview
Download

"40State.pdf" (pdf, 787.2Kb)

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®