Using digital measures to predict sleep impairment in adolescents with mood disorders treated with antidepressants.

BACKGROUND: Evidence on the use of consumer-based wearables to inform care for youth with mood disorders on antidepressants is limited. We assessed whether longitudinal data from a mobile app and a wearable could be used to characterize sleep among these youth. METHODS: Prospective longitudinal study of a convenience sample of 32 adolescents (12-17 years) with (n = 22) and without (control, n = 10) mood disorders. Sleep measures (total sleep time, later sleep onset) were captured over 28 days using an accelerometry-based wearable digital health technology. The PROMIS® Pediatric Sleep-Related Impairment questionnaire was used at baseline and every 3 days throughout follow-up via the mobile app. We compared differences in sleep measures between groups and their association with sleep impairment over time using linear mixed models that clustered longitudinal measures at the participant level. RESULTS: Mean (SD) age was 16 (1.5) years, with 62.5% female. Mean total sleep time was similar between mood and control groups. Mood youth had a lower percentage of nights within the recommended sleep duration range (7-11 h) (28%) than controls (47.5%). A higher percentage of nights with later sleep onset (2+ hours past 11 pm) was observed in the mood group (14.4%) than in controls (7.6%). Poor baseline sleep impairment scores and female sex, predicted worse sleep impairment scores at the end of follow-up. LIMITATIONS: Lower engagement with the app and wearable tracker among youth with mood disorders. CONCLUSIONS: Digital technologies create an opportunity to enable longitudinal biopsychosocial data collection, with the potential to inform management of mood disorders in youth.