Abstract
OBJECTIVE: the Medicare Health Outcomes Survey (HOS), a new quality measure in the Health Plan Employer Data and Information Set, is designed to assess physical and mental functional health outcomes of Medicare beneficiaries enrolled in Medicare+Choice organizations. We discuss the rationale for the HOS measure together with methodologic challenges in its use and interpretation, using descriptive data from the baseline Medicare HOS to illustrate some of these challenges. DATA SOURCES/STUDY DESIGN: The 1999 Cohort 2 Medicare HOS baseline data were used for a cross-sectional descriptive analysis. A random sample of 1,000 beneficiaries from each health plan with a Medicare+Choice contract was surveyed (N = 156,842; 282 organizations included in these analyses) . PRINCIPAL FINDINGS: The HOS measure is designed to assess a previously unmeasured dimension of quality. Plan-level variation was seen across all baseline measures of sociodemographic characteristics and illness burden. At the individual level socioeconomic position as measured by educational attainment was strongly associated with functional status. The least educated beneficiaries had the highest burden of illness on all measures examined, and there was a consistent and significant gradient in health and functional status across all levels of education. In analyses stratified by race and ethnicity, socioeconomic gradients in f un ct ion persist ed. CONCLUSIONS Despite limitations, by focusing at t en t ion on the need to improve functional health out comes among elderly Medicare beneficiaries enrolled in Medicare+Choice, the HOS can serve as an important new tool to support efforts to improve health care quality. The HOS provides valuable information at the federal, state, and health plan levels that can be used to identify, prioritize, and evaluate quality improvement interventions and monitor progress for the program overall as well as for vulnerable subgroups. To interpret the HOS as a quality measure individual-and plan-level differences in functional status and illness burden, as well as methodologic issues in health status measurement, need to be recognized and addressed.
Full Text
The Full Text of this article is available as a PDF (96.7 KB).