Early change in daily step counts after antidepressant initiation among adults with depression

Objective Early real-world activity changes after starting antidepressants are not well described. This study aimed to quantify short-term changes in daily ambulatory activity around antidepressant initiation using wearables. Methods We analyzed the All of Us Research Program CDR v8 (data cutoff October 1, 2023). Adults with incident major depressive disorder and a first antidepressant (ATC N06A) record within 90 days were included. Prior-year antidepressant use and history of schizophrenia or bipolar disorder were excluded. Fitbit step counts were summarized daily from 60 days before to 60 days after the index date. Valid days required ≥600 min of non-missing minute-level heart rate data. We fit a within-person mixed-effects piecewise interrupted time-series model on log steps with breakpoints at day 0 and day 14, adjusting for day of week. Results Among 168 participants (mean age 44.2 years; 78% female), steps declined in pre-index period (−0.16%/day; p = 0.027). During days 0–14 after initiation, the slope increased versus pre-index ( p = 0.012), with predicted mean steps rising from 6634 (day 0) to 7341 (day 14; ∼0.74%/day). After day 14, changes were minimal (∼0.07%/day). An immediate decrease at initiation was observed (−9.7%; p = 0.031). After additionally adjusting for days 0–2, the increasing trend over days 0–14 remained. Conclusions Daily Fitbit step counts captured distinct activity patterns in the first two weeks after antidepressant initiation, suggesting wearables can support high-frequency monitoring during treatment start.