The Association Between Depression and Medication Non-Adherence in Type 2 Diabetes: A Population-Based Cohort Study

Background: Depression has been correlated with suboptimal adherence to antidiabetic drug (AD) treatment in many studies. However, most of them were cross-sectional and the association and directionality of this relationship has not yet been established. Purpose: To measure the association between depression and non-adherence to AD treatment. Methods: We performed a population-based cohort study using the administrative data from the Quebec health insurance board (RAMQ) among adults newly using ADs. Depression was determined using a validated algorithm. To avoid immortal time bias, we carried out a depression diagnosis-time distribution matching by assigning a random fictive date of depression to individuals without depression. Non-adherence to AD treatment (i.e. <90% of days covered by at least 1 AD) in the 1-year period following depression diagnosis (real or fictive) was the main outcome. To estimate the effect of depression on AD treatment non-adherence, we conducted a multivariate logistic regression analysis, adjusting for baseline adherence and other potential confounders. Results: Between 2000 and 2006 we identified 3106 patients with a diagnosis of depression matched to 70,633 patients without depression. Among patients with depression, 52.0% were subsequently non-adherent with their AD treatment vs. 49% of patients without depression. Patients with depression were 24% more likely to be non-adherent to their AD treatment (adjusted odds ratio=1.24, 95% confidence interval: 1.13 to 1.37). Conclusions: Results suggest that depression is an independent risk factor of non-adherence to AD treatment. Patients with type 2 diabetes might benefit from regular detection for depression and adherence-enhancing interventions.