Are We Paying for What Matters: The Relationship of HEDIS Measures and Health Outcomes

Background. Recent policy proposals have called for paying health care providers on the basis of patient health outcomes. While there has been wide spread adoption of clinical process measurement, we know little about the role of clinical care on subsequent health outcomes.

Objective. To examine the association of adherence to HEDIS outcome measures in Medicare Advantage plans on patient health outcomes with respect to five Health Plan Employer and Data Information Set (HEDIS) outcome measures.

Study Design. We examine the relationship between HEDIS outcome measures and health outcomes using multivariable regression models. The study sample included 182,762 respondents to the Medicare Health Outcome Survey baseline survey in 2011 and follow up survey in 2013.  We examine these relationships in the overall sample as well is key subgroups: race, education, and chronic disease.

Main Outcome Measures: Reported number of unhealthy days, reported level of bodily pain, and presence of depressive symptoms.

Results. Overall adherence to HEDIS outcome measures ranged from50% to 70%. We found inconsistency in changes in health over time: the average number of reported unhealthy days decreased from 7.61 at baseline to 7.45 at follow up and the presence of depressive symptoms decreased from 7.4% to 7%, while the reported proportion of individuals in extreme bodily pain increased from 4.4% to 5%. In the overall sample, we did not find a significant relationship between individual adherence to HEDIS measures and health outcomes. Among specific disease groups, we found that the odds of extreme pain were significantly lower in older adults with arthritis, COPD, or heart disease who received advice to maintain or increase physical activity. Similarly, reported efforts to address fall risk were associated with fewer unhealthy days in older adults with COPD.

Conclusion. We find some promising evidence that adherence with recommended clinical care process is associated with better health outcomes, but that the association depends on health profiles. In a two-year time period, there limited evidence that clinical processes can change health status. Additional research is needed to examine the association between clinical care processes and health outcomes over a longer time frame.