JMIR Publications

ABSTRACT

Background:

Self-Monitoring of Dietary Behavior (SMDB) is typically a central component of behavioral weight loss programs, widely recognized for its effectiveness in promoting healthy behavior changes and improving health outcomes. Providing tailored feedback and emotional social support are potent behavior change strategies for enhancing adherence to SMDB. However, analyzing the adherence dynamics of overweight or obese individuals using digital SMDB and the effects of different strategies remains challenging.

Objective:

This study analyzed the intervention effects of a 28-day digital behavioral weight loss program targeting overweight and obese individuals. Utilizing the Adaptive Control of Thought-Rational (ACT-R) cognitive architecture, it modeled and examined the SMDB adherence dynamics under the impact of various intervention strategies, considering both goal pursuit and habit formation perspectives.

Methods:

The digital behavioral weight loss program employed tailored feedback and emotional social support strategies to enhance participants' adherence to SMDB. Participants were divided into three groups: the Self-Management (SM) group, the Tailored Feedback (TF) group, and the intensive support (IS) group. The Kruskal-Wallis test was used to analyze the inter-group differences in adherence. Furthermore, an ACT-R-based mathematical model was constructed to examine the adherence sequences and the effects of different strategies The model fit was evaluated using mean squared error, root mean squared error, and goodness of fit.

Results:

Ninety-seven participants completed the intervention (participation rate 0.84). After 28 days of intervention, there were significant reductions in both weight and body fat percentage among the overweight/obese participants. The average SMDB adherence rate was 0.55 (SD=0.38) for the SM group, 0.72 (SD=0.29) for the TF group, and 0.83 (SD=0.20) for the IS group. The adherence rate in the IS group (0.83, SD=0.20) was significantly higher than that in the SM group (0.55, SD=0.38) (P=.04). The modeling results based on ACT-R for participants' SMDB adherence sequence showed that the goodness of fit for the models in all three intervention groups exceeded 0.7.

Conclusions:

This study is a significant endeavor in the dynamic analysis of user behavior within digital behavioral interventions. Future research should prioritize the optimization of diversity of feedback and social support through advanced technological approaches, ensuring that interventions are aligned with participants' actual needs. Clinical Trial: ChiCTR2200055548