Using Medication Adherence as a Measure of Healthcare Quality

Policymakers are increasingly using measures of the quality of healthcare to try and provider health plans and providers with incentives to improve healthcare delivery.  New payment models are emerging that directly incorporate these quality data into coverage and reimbursement decisions.  Many such measures have been developed to monitor adherence to prescription medicines treating chronic diseases.  Medication adherence is a natural candidate for quality measurement, as individuals with higher adherence consistently demonstrate better clinical outcomes and lower use and spending on other medical services.  However, there is little evidence on whether explicitly measuring and tracking medication adherence and outcomes at the plan-level will actually lead to improvements in patient outcomes. In particular, it is not known whether providers whose patients have better average medication adherence will also exhibit better health outcomes or lower health spending.

We used data medical and pharmacy claims from two sources to examine the relationship between provider-level measures of medication adherence on outcome measures for patients with diabetes and coronary heart failure (CHF).  We compared mean adherence to diabetes and CHF medications to mean health outcomes within health plans.  Provider performance was measured based on rates of disease complications, hospitalizations, and use of emergent care.  Outcomes were adjusted for average differences inpatient age, gender and co-morbidities using multivariate regression.  Medication adherence was calculated as the fraction of patients with at least 80% proportion of days covered for medications in classes used to manage diabetes or CHF.  To assess the relationship between plan-level adherence and quality, we used claims data from a variety of health plans from 46 large employers.  To assess the relationship between the adherence of individual physicians’ patients and their average outcomes, we used data from Medicare Part A and B claims linked to Part D pharmacy claims. 

We found that average adherence varied significantly across providers and plans, even after conditioning on individual patient characteristics.  We also found that there was a strong relationship between medication adherence measures and health outcomes at both the plan and provider level.  For example, plans with low adherence to diabetes medications had adjusted rates of uncontrolled diabetes admissions of 13.2 per 1,000 patients, compared to 11.2 in moderate adherence plans and 8.3 in high adherence plans (p<0.001).  The adjusted rate of CHF-related hospitalization was 15.3% in low adherence plans, compared to 12.4% in moderate adherence plans and 12.2% in low adherence plans (p<0.001). These patterns were consistent across different types of complications for both diabetes and CHF, at both the provider and plan level.  These results indicate that medication adherence measures are potentially useful tools for improving the performance of health plans.  More work is needed to understand the mechanisms through which health plans or providers could actually use to improve their patients’ medication adherence, and how these respond to financial incentives.