Predictors of medication nonadherence among patients with diabetes in Medicare Part D programs: A retrospective cohort study

Abstract Objectives: This study examined the prevalence of nonadherence with oral hypoglycemic agents, antihypertensive drugs (angiotensin-converting enzyme inhibitors [ACEIs] and angiotensin II receptor blockers [ARBs]), and statin medications among Medicare Part D enrollees with diabetes and analyzed the potential demographic and clinical factors that predict medication nonadherence. Methods: This was a retrospective cohort study of Medicare Part D enrollees with diabetes from 6 states (Alabama, California, Florida, Mississippi, New York, and Ohio). Adherence was calculated as the proportion of days covered (PDC; number of days with medication on hand/number of days in the specified time interval). A PDC was derived for each of the 3 categories of medications for patients who had at least 1 claim for the same class of medication. A comorbidity measure was created for each beneficiary using the Deyoadapted Charlson Comorbidity Index (CCI). Associations between nonadherence and patient characteristics including age, sex, race/ethnicity, and Deyo-adapted CCI were examined. Logistic regression models were constructed to identify predictors of nonadherence. Results: The study included 1,888,682 patients with diabetes. The mean (SD) age was 71.6 (11.6) years, and 59.5% (1,123,220/1,888,682) were female. A total of 66.4% (1,254,538/1,888,682) were white, 16.3% (308,158/1,888,682) were black, and 7.8% (147,498/1,888,682) were Hispanic. Estimated rates of non-adherence for oral hypoglycemic agents, ACEIs/ARBs, and statins were 35.1% (386,666/1,101,533), 41.8% (449,561/1,075,285), and 46.4% (447,106/962,877), respectively. In unadjusted analyses, patients aged