Comparing Accelerometer and Self-Reported Treatment Effects in a Technology-Supported Physical Activity Intervention

To estimate and compare the change in moderate-to-vigorous physical activity (MVPA) between an accelerometer and technology-supported physical activity (PA) log across a 3-week PA intervention.

Participants (N = 204, 77% female, age = 33 ± 11 years, body mass index = 28.2 ± 7.1 kg/m²) were randomized to one of two activity-related intervention arms: (1) increase MVPA intervention or (2) decrease sedentary behavior active control. Participants wore an accelerometer while simultaneously completing a technology-based PA log every day for 5 weeks: a 2-week baseline assessment phase and a 3-week intervention phase. Bivariate linear mixed-effects models and correlations were used to characterize the relationship of MVPA between measurement methods throughout the intervention. Effect sizes were calculated to determine the intervention effect by measurement method.

At baseline, PA log MVPA was 28 minutes greater than accelerometer-based minutes of MVPA in the active control group. This difference was 35 minutes (95% CI [23.7, 46.1]) greater at follow-up than at baseline measurement in the MVPA intervention group. In the active control group, there was a significant 16-minute (95% CI [6.0, 26.5]) increase between the two measures from baseline to follow-up. The intervention effect size based on the PA log was 0.27 (95% CI [0.14, 0.39]) and 0.42 (95% CI [0.28, 0.56]) when using the accelerometer.

Our results indicate that PA log MVPA and accelerometer MVPA estimate significantly different minutes per day of MVPA. It is important researchers use caution when comparing MVPA intervention outcomes from different measurement methods.