Abstract
BACKGROUND: Individuals with depression who do not respond to initial antidepressant may switch to a different antidepressant, add a second antidepressant, or add an atypical antipsychotic. Previous studies comparing these strategies' efficacy and safety reported conflicting results, and the impact of these strategies on subsequent health care utilization is unknown. OBJECTIVE: To compare health care utilization between individuals with depression who switched antidepressants, added a second antidepressant (ie, combination), or added an atypical antipsychotic (ie, augmentation) following their initial antidepressant. METHODS: This retrospective cohort study used a 25% random sample of the IQVIA PharMetrics Plus for Academics health plan claims database. The study cohort included individuals aged 10-64 years who newly initiated an antidepressant at any point from January 2016 to December 2020. New use was defined as no evidence of an antidepressant in the 180 days preceding the antidepressant dispensing. Individuals had to have a depression diagnosis and a treatment change in the 180 days following the initial antidepressant. The index date was the date of the first observed antidepressant change, which was defined as a switch, combination, or augmentation. Health care utilization, measured as the number of outpatient visits, any all-cause hospitalization, and any emergency department (ED) visit, was assessed in the 180 days after the index date. Negative binomial regression models evaluated the rate ratio of the number of outpatient visits. Logistic regression models estimated the odds ratio of a hospitalization, and modified Poisson regression estimated the relative risk of an ED visit. Models were adjusted for demographics, clinical characteristics, and previous health care utilization. RESULTS: Among 3,847 individuals with depression who had the first treatment change following the initial antidepressant, we identified 2,418 (62.9%) who switched, 1,268 (33.0%) who combined, and 161 (4.2%) who augmented their antidepressant. The augmentation group had a significantly higher rate of outpatient visits than the combination group (adjusted rate ratio = 1.14, 95% CI = 1.04-1.25). There was no statistically significant difference in hospitalizations or ED visits between the switch and augmentation vs combination groups. CONCLUSIONS: Augmentation comprised 4% of our antidepressant cohort but had higher outpatient health care utilization than those who combined treatment.