Background: Digital interventions integrating gamification features hold promise to promote daily steps. However, results regarding the effectiveness of this type of intervention are heterogeneous and not yet confirmed in real-life contexts.

Objective: This study aims to examine the effectiveness of a gamified intervention and its potential moderators in a large sample using real-world data. Specifically, we tested (1) whether a gamified intervention enhanced daily steps during the intervention and follow-up periods compared to baseline, (2) whether this enhancement was higher in participants in the intervention than in nonparticipants, and (3) what participant characteristics or intervention parameters moderated the effect of the program. Methods: Data from 4819 individuals who registered for a mobile health Kiplin program between 2019 and 2022 were retrospectively analyzed. In this intervention, participants could take part in one or several games in which their daily step count was tracked, allowing individuals to play with their overall activity. Nonparticipants were people who registered for the program but did not take part in the intervention and were considered as a control group. Daily step counts were measured via accelerometers embedded in either commercial wearables or smartphones of the participants. Exposure to the intervention, the intervention content, and participants' characteristics were included in multilevel models to test the study objectives.

Participants in the intervention group demonstrated a significantly greater increase in mean daily steps from baseline than nonparticipants (P<.001). However, intervention effectiveness depended on participants' initial physical activity. The daily steps of participants with <7500 baseline daily steps significantly improved from baseline both during the Kiplin intervention (+3291 daily steps) and the follow-up period (+945 daily steps), whereas participants with a higher baseline had no improvement or significant decreases in daily steps after the intervention. Age (P<.001) and exposure (P<.001) positively moderated the intervention effect.

In real-world settings and among a large sample, the Kiplin intervention was significantly effective in increasing the daily steps of participants from baseline during intervention and follow-up periods compared to nonparticipants. Interestingly, responses to the intervention differed based on participants' initial steps, with the existence of a plateau effect. Drawing on the insights of self-determination theory, we can assume that the effect of gamification could depend of the initial motivation and activity of participants.