Outcome Indicator Development
AHRQ developed the set of indicators used in this report through an extensive process that included an environmental scan, expert input, and validity and reliability testing. All but two of the indicators (Pressure Ulcer and Injurious Falls) build on existing AHRQ Prevention Quality Indicators (PQIs; go to ). The development process is described in detail elsewhere (Schultz, et al., 2012). The final set of indicators represents potentially avoidable inpatient admissions for 13 conditions, drawing on both the MAX and MedPAR data to identify admissions.
The 13 outcome indicators judged to be of sufficient validity and reliability for the HCBS population, based on testing to date, include:
- Short-Term Complications of Diabetes.
- Asthma or Chronic Obstructive Pulmonary Disease.
- Congestive Heart Failure.
- Composite: Potentially Preventable Infection:
- Bacterial Pneumonia.
- Urinary Tract Infection.
- Infection Due to Device or Implant.
- Dehydration.
- Composite: Ambulatory Care Sensitive Condition (ACSC): Chronic Conditions.
- Composite: ACSC: Acute Conditions.
- Composite: ACSC: Overall.
- Pressure Ulcer.
- Injurious Falls.
For each measure, the numerator represents the number of potentially avoidable inpatient hospital admissions and the denominator represents the number of HCBS participants divided by 100,000. Thus, the indicator as a whole reflects the rate of potentially avoidable inpatient admissions per 100,000 HCBS participants.
Because the HCBS population is dynamic, with individuals changing eligibility, enrollment, and service use throughout the year, we calculate the indicators for each quarter rather than for the entire year. In quarters when an individual does not meet our HCBS population definition, that individual is excluded from the denominator and any corresponding hospital admissions are excluded from the numerator. The quarterly rates are then annualized for presentation in this report. Excluding noneligible quarters means that the total number of people in the HCBS population at any point in 2005 is greater than the total number of people included in the outcome indicator denominators.
Details of each of the measure specifications, including numerator exclusions, are available in Schultz, et al. (2012). We note that one validated numerator exclusion, same-day admissions to a hospital for the same individual, was finalized after the preparation of the tables for this report. Thus, the admission rates in our report may be slightly larger than admission rates reflected in the measures reports, but qualitative conclusions are not affected by this minor difference.
All analyses that use the outcome indicators have several additional exclusions from the denominator, over and above the exclusions used in defining the HCBS population. We exclude individuals in a managed acute care plan or a managed long-term care plan. Thus, we exclude the entire State of Arizona, which has managed acute care for all Medicaid beneficiaries through an 1115 waiver. We also exclude individuals who exhibit institutional care use but no HCBS use during a quarter as demonstrated by MAX claims; individuals under age 18; and data from Wisconsin, Washington, and Maine due to missing data on HCBS use.
These indicators are intended for use by policymakers as tools for monitoring the welfare of recipients and should be applicable across diverse HCBS populations and across States. In their current state of development, they are not appropriate for use in establishing accountability of individual providers or State programs. The selected outcomes are clearly affected by the availability and quality of primary, acute, and long-term care and the underlying health and behaviors of the individuals, but none of the measures used in this report are risk adjusted.
Outcome Indicator Rates
Overall outcome indicator rates in the HCBS population vary considerably across measures, as shown in Table A. Short-Term Complications of Diabetes and Injurious Falls exhibit the lowest rates, with fewer than 300 potentially avoidable hospital admissions per 100,000 HCBS participants. The ACSC Overall Composite exhibits the highest rate, at almost 18,000 admissions per 100,000 HCBS participants. Rates for most of the other indicators fall between 2,000 and 8,000 admissions per 100,000 HCBS participants.
Table A also displays outcome indicator rates for the full Medicaid population. Although the pattern from measure to measure is similar in the two populations, the rates in the HCBS population are consistently and dramatically higher than in the full Medicaid population, more than twice as high for almost all measures. Thus, while the HCBS population constitutes just 4 percent of the overall Medicaid population, it is a particularly sick part of the Medicaid population and one where efforts to reduce potentially avoidable hospitalizations may be most needed.
In the sections that follow, we examine these outcome indicator rates with respect to specific participant, policy, and area characteristics.
