Adherence to Medications to Treat Diabetes, Hypertension, and Hypercholesterolemia among Medicare Part D Beneficiaries

Background : Poor adherence to medication is associated with significant increased morbidity, mortality and health care utilization. Older adults are more vulnerable to medication nonadherence because this population utilizes more medications and has lower health literacy rates than younger adults. The purpose of this dissertation is to examine adherence to pharmacologic therapies used to treat diabetes, hypertension, and hyperlipidemia among Medicare Part D beneficiaries. Specific objectives are 1) to evaluate the effectiveness of a coordinated patient- directed medication adherence intervention on adherence rates and a Medicare-Advantage Prescription Drug (MAPD) plan's CMS Part D star ratings, 2) to assess the impact of a quasi-experimental multi-channel adherence intervention on beneficiaries' medication adherence and health plan quality performance measures among two MAPD plans, and 3) to critically review recently published adherence interventions with specific focus on measurement methods and theoretical fidelity. Methods : To address the first two objectives, two quasi-experimental interventions (a daily prescriber-directed 90-day program and a patient- directed counseling program), were assessed by using pharmacy claims and member eligibility data. Pre-post changes in adherence were adjusted for demographics, comorbid conditions and secular adherence trends. To address the remaining objective, articles of adherence interventions in the past five years identified by a systematic review were examined and scored based on quality of measurement method and use of a behavioral theoretical framework. Results : For the patient-directed intervention, pre-post adherence rates increased an average 20.7 percentage points (P <.001). For the prescriber-directed program, adherence for the intervention group increased 2.0 (P <.001) and 1.8 percentage points (P <.001) for antihypertensives and antihyperlipidemics, respectively, relative to control. In both interventions, naïve to treatment and younger age were risk factors for nonadherence. Subjective indirect measurement methods (e.g., self-reports and electronic monitoring) are mostly used in randomized clinical trials whereas objective indirect methods (e.g., pharmacy claims) are employed in observational studies. Few interventions identified a behavioral model. Conclusions : Large scale interventions may offer an effective approach for health plans to address common adherence barriers and improve both adherence and quality performance ratings. Future clinical trials addressing medication adherence should incorporate theoretical frameworks to address the complexity of medication use behavio