To explore the potential of applying smartphone-based monitoring for depressive symptom tracking in patients with recent ischemic stroke (IS) and transient ischemic attack (TIA) syndromes.

Passive smartphone sensor data enables clinicians to longitudinally observe real-world mobility and physical activity and may facilitate routine screening for post-stroke depression (PSD), a risk factor for poor functional outcomes.

We enrolled 54 patients with recent TIA or IS symptoms who enabled the Beiwe smartphone application to collect passive data from accelerometer, GPS, and touch screen sensors. Participants completed weekly Patient Health Questionnaire-8 (PHQ-8) surveys remotely via the app and were evaluated remotely every 30-days with the staff-administered Montgomery–Åsberg Depression Rating Scale (MADRS). Repeated measures correlation evaluated associations between longitudinal sensor data and PHQ-8 scores. Using sensor data and PHQ-8 scores, linear mixed effects models predicted depression severity.

The analytic sample (n = 41) consisted of mostly males (56.1%) and iOS users (63.2%) with a mean age of 58 years. PHQ-8 survey completion time (seconds) was positively correlated with depressed mood and self-blame.  Average time spent in one location was positively correlated with total PHQ-8 score, disinterest, depressed mood, sleep problems, lethargy, and self-blame. Negative correlations were observed between places visited (n) and lethargy as well as average commute time and depressed mood. For MADRS score prediction, the average root-mean-squared error (RMSE) was 2.12 using the antecedent PHQ-8 score alone, 2.10 using only passive accelerometer and GPS data, and 1.92 when using passive data and the PHQ-8 score with survey timing.