(PS2-86) Smartphone GPS Tracking as Objective Measure of Social Participation in Bipolar Disorders: A Pilot Study

Abstract Text: Despite the call for more objective measures of social participation, i.e., engagement in social roles/activities and community, most research continues to rely on retrospective self-report measures (Dijkers, 2010). Digital technology and mHealth methods are a growing area of promising objective measures of clinical outcomes, but these studies rarely focus on social participation measurements in bipolar disorders (BD). This is a significant gap in the literature, considering that people with BD experience relatively more “out of role days” and that people with BD view social functioning as one of the most important components of recovery (Alonso et al., 2011; Mezes et al., 2022). This study is the first to examine smartphone GPS tracking as an objective measure of social participation and explore its associations with self-report measures of social participation in individuals with BD.
Study participants were 44 patients with a BD diagnosis receiving care within a VA medical center (mean age = 48; 68% male). Clinical interviews using SCID-5 confirmed primary bipolar disorder diagnosis (25 = BD I, 15 = BD II, 4 = Other Specified BD) and established comorbid psychopathology. Participants provided passive GPS-tracking using a smartphone app during a 3-month follow-up. Participants also completed well-established self-report measures of social participation, PROMIS Satisfaction with Social Roles and Activities and PROMIS Ability to Participate in Social Roles and Activities (version 2.0, 8a), and Location Context Interviews (LCI) in 7 biweekly assessments. LCIs provided information on the type of location visited to categorize locations into bins: home, occupational (e.g., work), daily chores (e.g., grocery stores), and social engagement (e.g., friend’s house, museum).
Hierarchical linear models examined associations between GPS-tracking indices (time spent in each location bin) and PROMIS measures. For example, average time per day spent at home (in hours) had a positive association with PROMIS Ability T-scores (b = 0.80, R2c = 0.65), but an inverse association with PROMIS Satisfaction T-scores (b = -1.18, R2c = 0.65). Lower marginal R2 values (0.006--0.014) indicated that fixed effects explained little of models’ variance. To better visualize GPS-tracking data and its relationships with self-report indices, figures will illustrate smartphone GPS-tracking data across the follow-up separately for participants who exhibited consistently high social participation, consistently low social participation, or changing/variable social participation based on PROMIS measures. Figures will also illustrate significant between-participant differences observed across the follow-up.
GPS-tracking appears to be a promising objective measure of social participation given its passive nature requiring low patient effort and no insight into current functioning both in BD and non-BD clinical populations. Still, these preliminary findings show the need for more understanding how much GPS-tracking indices capture individual/trait-like differences versus clinical state-dependent within-person changes in social participation.