Tracked Physical Activity Levels Before and After a Change in Incentive Strategy Among UK Adults Using a Rewards App: Retrospective Quasi-Experimental Study.

BACKGROUND

Financial incentives delivered via apps appear to be effective in encouraging physical activity. However, the literature on different incentive strategies is limited, and the question remains whether financial incentives offer a cost-effective intervention that could be funded at the population level.

OBJECTIVE

This study aimed to explore patterns of tracked physical activity by users of an incentive-based app before and after a change in incentive strategy. A business decision to alter the incentives in a commercially available app offered a natural experiment to explore GPS-tracked data in a retrospective, quasi-experimental study. The purpose of this exploratory analysis was to inform the design of future controlled trials of incentives delivered via an app to optimize their usability and cost-effectiveness.

METHODS

Weekly minutes of tracked physical activity were explored among a sample of 1666 participants. A Friedman test was used to determine differences in physical activity before and after the change in incentive strategies. Post hoc Wilcoxon tests were used to assess minutes of physical activity in the 2 weeks before and after the change. A secondary analysis explored longitudinal patterns of physical activity by plotting the mean and median minutes of physical activity from 17 weeks before and 13 weeks after the change in incentive strategy. CIs were calculated using bias-corrected bootstraps. Demographics were also explored in this way.

RESULTS

There were significant differences in the weekly minutes of activity before and after the change in incentive strategy (Friedman χ=42, P<.001). However, a longitudinal view of the data showed a more complex and marked variation in activity over time that undermined the conclusions of the before/after analysis.

CONCLUSIONS

Short-term before-and-after observational studies of app-tracked physical activity may result in misleading conclusions about the effectiveness of incentive strategies. Longitudinal views of the data show that important fluctuations are occurring over time. Future studies of app-tracked physical activity should explore such variations by using longitudinal analyses and accounting for possible moderating variables to better understand what an effective incentive might be, for whom, and at what cost.