Abstract P047: Depressive Symptom Trajectory Associated With Step Counts From Smartwatch: The Electronic Framingham Heart Study

Introduction: Depression is a common and serious medical illness world-wide. Studies have demonstrated a protective effect of physical activity on depression, but few studies have examined the prospective association between depressive symptoms and objectively measured physical activity. Objective: To investigate if antecedent depressive symptoms predict subsequent daily step counts among participants enrolled in the electronic Framingham Heart Study (eFHS). We hypothesize that the trajectory group with a larger proportion of participants having depressive symptoms and antidepressant use is associated with a lower level of physical activity. Methods: We used group-based multi-trajectory modeling to construct depressive symptom trajectory patterns using both depressive symptoms and antidepressant use collected from eFHS participants who attended three FHS health exams over fourteen years. The presence of depressive symptoms was defined at each exam with a Center for Epidemiological Studies-Depression (CES-D) score > 16. At the final exam, participants were provided with a study smartwatch to measure daily step counts. We excluded: a) person-day records with less than 5 hours of wear-time, or less than 1000 steps recorded in a day; b) participants who failed to return data for ≥1 day per week in the first month. We performed linear mixed effects models to examine the association between depressive symptom trajectories and average daily step counts during the first 30 days of smartwatch use and one-year of follow-up adjusting for age, sex, wear-hour, BMI, and smoking status. We used p<0.05 for significance. Results: We identified two depressive symptom trajectory groups for the 724 eFHS participants (mean age 53 years, 60% women). The “low-stable” group (n=578; mean follow-up 286±110 days) consisted of <5% of participants with depressive symptoms and <3% of those with antidepressant medication use across the three exams. The “high-increasing” group ( n = 146; 269±114 days) consisted of >21% of participants with depressive symptoms and >46% of those with antidepressant medication use across the three exams. Compared to the low-stable trajectory group, the participants in the high-increasing depressive symptoms group walked 475 (95% CI: 24-927) fewer daily steps during the first 30 days of smartwatch use ( p = 0.04) and walked 550 (95% CI: 109-991) fewer daily steps during one-year of follow-up ( p = 0.01). Conclusion: Participants in the group with a larger proportion of antecedent depression and anti-depressant use walked significantly fewer steps compared to the group with a smaller proportion of antecedent depression and anti-depressant use. Our findings suggest that interventions via mHealth technologies that target mood may play an important role in promoting physical activity.