A Preliminary Investigation of an Automated Online Program for Increasing Physical Activity among Adults: A Randomized Trial

ABSTRACT Introduction Given the numerous health benefits associated with regular physical activity (PA), effective, affordable, and scalable PA interventions are needed to address the high prevalence of inactivity among adults within the United States. The purpose of this study was to evaluate the feasibility, acceptability, and preliminary efficacy of an Internet-delivered program for increasing moderate-to-vigorous PA (MVPA) among inactive adults. Methods Participants ( n = 55; age: 53.5 ± 10.2 yr, body mass index: 32.0 ± 6.1 kg·m −2 , 56.4% female, 83.6% non-Hispanic White) were randomized to the Energize! exercise program, a 12-wk, fully automated program for increasing MVPA, or a newsletter control condition. Each week, Energize! participants completed behaviorally based video lessons, submitted an exercise plan, reported their exercise minutes, and completed a homework assignment. The PA goal progressed from 75 to 200 min·wk −1 and automated, personalized feedback was provided. Control participants received PA education newsletters twice per month (six in total). Assessments of psychosocial PA constructs and MVPA (self-report, accelerometer-derived total MVPA, and accelerometer-derived bouted MVPA (>10 min bouts)) occurred at baseline, 3 months (post-intervention), and 6 months (following a 3-month no-contact follow-up). Results Retention was 85.5% and 81.8% at 3 and 6 months, respectively. On average, Energize! participants submitted an exercise plan in 76.4% of weeks, watched 84.5% of video lessons, and completed 75.0% of homework assignments. Program satisfaction was higher in Energize! than the control group (all P < 0.001). A significant interaction effect favoring Energize! was observed from baseline to 3 months for bouted MVPA ( P = 0.01) and total MVPA ( P = 0.01). No significant differences between groups were observed for changes from baseline to 6 months. At both 3 and 6 months, medium to large effect sizes were observed for most psychosocial PA variables (Cohen’s d = 0.57–1.08). Conclusion This fully automated program was feasible, acceptable, and effective at increasing MVPA in the short-term, but effects were diminished once the intervention was removed.