Group-based trajectory modeling to identify adherence patterns for direct oral anticoagulants in Medicare beneficiaries with atrial fibrillation: a real-world study on medication adherence.

Abstract

Suboptimal adherence to direct oral anticoagulants (DOACs) among atrial fibrillation (AF) patients remains currently a major concern due to the increased risk of cardiac and thromboembolic events. To identify longitudinal distinct trajectories of DOAC adherence and sociodemographic and clinical factors associated witheach trajectory. Patients with AF who were prescribed withDOAC from July 2016-December 2017 were identified among patients enrolled in the Medicare Advantage Plan. Patients were followed up for a year after the index date to calculate the monthly proportion of days covered (PDC). The monthly PDC was incorporated into the logistic group-based trajectory model to evaluate distinct patterns of adherence. A multinomial regression model was carried out to assess various predictors associated with each trajectory. Sub-group analysis was conducted among incident DOAC users. Total of 1969 patients with AF, four distinct trajectories of adherence were selected: adherent 36.8%, gaps in adherence 9.3%, gradual decline in adherence 29.7%, and rapid decline in adherence 24.2%. Significant predictors associated with suboptimal adherence trajectories were age (75 years or older), gender (male vs female), low-income subsidy health plan, prevalent users, and presence of comorbidities. Among 933 incident users, three adherence trajectories were identified: adherent trajectory (31.8%), rapid decline in adherence (32.5%), and gradual decline in adherence (35.6%). The significant predictors among incident users were gender (male vs female), low-income subsidy health plan, HAS-BLED score ≥ 2, and presence of coronary artery disease. Adherence to DOACs was suboptimal among the total population and incident users.