Automated Physical Activity Support for Adults and Youth From Low-Income Communities: Single-Arm Pilot Study

Background Mobile health (mHealth) interventions are growing in popularity, but less research has focused on low-income families, particularly interventions integrating wearable devices with automated personalized messages. Objective We tested a preliminary wearable-integrated mHealth intervention with initial personalization elements among adults and youth from low-income urban communities, focusing on feasibility, acceptability, and preliminary evidence of physical activity behavior. Methods Participants were 83 adults and 31 youth recruited through community health events held in low-income urban communities. Using a single-arm pre-post design, participants were enrolled into a 7-week beta-version mHealth intervention that integrated a Garmin activity monitor with automated text messages. Messages were sent 4 days/week and focused on increasing step counts using theory-based behavior change techniques related to goal setting, self-monitoring, reinforcement, contextual factors, and self-efficacy. Most messages were personalized by including calculations based on the step-count and step-goal data, using branching logic, and using 2-way question-and-response messages. Feasibility measures included enrollment, retention, fidelity of message delivery, and adherence to wearing the Garmin device. Acceptability measures included survey items and engagement with responding to 2-way messages. Changes in daily steps were explored using mixed-effects linear regression. Results Enrollment and eligibility rates were 64% (84/132, adults) and 63% (31/49, youth), retention for physical activity measures was 84% (70/83) and 77% (24/31), and 99% (3910/3955) of the intended messages were delivered. Adults and youth adhered to wearing the Garmin on 82% (45/56) and 79% (44/56) of the study days, respectively. Overall acceptability ratings were 83% to 100%, with 97% (75/77) of adults and 100% (27/27) of youth indicating they would recommend the program. Adults and youth replied to a mean of 2.6 (SD 2.2) and 3.2 (SD 2.7) of the 7 text messages that asked for a reply, with higher engagement among adults who participated with their child. Pre-post changes in daily steps were β=240 (95% CI –387 to 866) for adults and β=413 (95% CI –877 to 1703) among youth, with larger changes observed among those in the highest tertile of engagement (adults: β=584, 95% CI –784 to 1952; n=19; youth: β=941, 95% CI –827 to 2709; n=11) and those who were meeting less than two-thirds of the physical activity guideline at baseline (adults: β=609, 95% CI –30 to 1247; n=47; youth: β=1406, 95% CI –94 to 2907; n=22). Conclusions Personalized mHealth physical activity interventions integrating wearable step trackers with automated text messaging appear to be feasible and acceptable among adults and youth from low-income communities. Step-count findings show promise for the intervention’s ability to support individuals who are further from meeting physical activity guidelines and warrant more research among parent–child dyads. Overall, findings support additional research to optimize and evaluate similar interventions within this population group using fully powered randomized controlled trials. Trial Registration ClinicalTrials.gov NCT05110508; https://clinicaltrials.gov/ct2/show/NCT05110508