Objectives: To develop activity limitation clusters for case-mix adjustment of health care ratings and as a population profiler, and to develop a cognitively accessible report of statistically reliable quality and access measures comparing the health care experiences of adults with and without disabilities, within and across health delivery organizations.
Design: Observational study.
Setting: Three California Medicaid health care organizations.
Participants: Adults (N = 1086) of working age enrolled for at least 1 year in Medicaid because of disability.
Interventions: Not applicable.
Main outcome measures: Principal components analysis created 4 clusters of activity limitations that we used to characterize case mix. We identified and calculated 28 quality measures using responses from a proposed enabled version of the Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey. We calculated scores for overall care as the weighted mean of the case-mix adjusted ratings.
Results: Disability caused a greater bias on health plan ratings and specialist ratings than did demographic factors. Proxy respondents rated care the same as self-respondents. Telephone and mail administration were equivalent for service reports, but telephone respondents tended to offer more positive global ratings. Plan-level reliability estimates for new composites on shared decision making and advice on healthy living are .79 and .87, respectively. Plan-level reliability estimates for a new composite measure on family planning did not discriminate between health plans because respondents rated all health plans poorly. Approximately 125 respondents per site are necessary to detect group differences.
Conclusions: Self-reported activity limitations incorporating standard questions from the American Community Survey can be used to create a disability case-mix index and to construct profiles of a population's activity limitations. The enabled comparative report, which we call the Assessment of Health Plans and Providers by People with Activity Limitations, is more cognitively accessible than typical CAHPS report templates for state Medicaid plans. The CAHPS Medicaid reporting tools may provide misleading ratings of health plan and physician quality by people with disabilities because the mean ratings do not account for systematic biases associated with disability. More testing on larger populations would help to quantify the strength of various reporting biases.