Longitudinal Oral Anticoagulant Adherence Trajectories in Patients With Atrial Fibrillation

Abstract

Background: Conventional adherence summary measures do not capture the dynamic nature of adherence.

Objectives: This study aims to characterize distinct long-term oral anticoagulant adherence trajectories and the factors associated with them in patients with atrial fibrillation.

Methods: Adults with incident atrial fibrillation were identified using linked population-based administrative health data in British Columbia, Canada (1996-2019). Group-based trajectory modeling was used to model patients' 90-day proportions of days covered over time to identify distinct 5-year adherence trajectories. Multinomial regression analysis was used to assess the effect of various demographic and clinical factors on exhibiting each adherence trajectory.

Results: The study cohort included 19,749 patients with AF (mean age: 70.6 ± 10.6 years), 56% male, mean CHA₂DS₂-VASc stroke risk score 2.8 ± 1.4. Group-based trajectory modeling identified 4 distinct oral anticoagulants adherence trajectories: "consistent adherence" (n = 14,631, 74% of the cohort), "rapid decline and discontinuation" (n = 2,327, 12%), "rapid decline and partial recovery" (n = 1,973, 10%), and "slow decline and discontinuation" (n = 819, 4%). Very few patient variables were found to be associated with specific adherence trajectories.

Conclusions: There is heterogeneity among nonadherent patients in the rate and timing of decline in their medication taking. Clinical and demographic characteristics were found to be inadequate to predict patients' adherence trajectories. Insights from this study could be used to inform the design and timing of adherence interventions, and qualitative studies may be needed to better understand the psychosocial determinants and reasons for the behaviors reflected in the identified trajectories.

Keywords: adherence; anticoagulants; atrial fibrillation; group-based trajectory modeling; proportion of days covered; trajectory.