Assessing the contribution of self-monitoring through a commercial weight loss app: mediation and predictive modeling
Consistent dietary and physical activity self-monitoring is an essential component of successful weight loss in traditional in-person behavioral weight loss programs and technology-based programs. Commercial technology-based dietary and physical activity monitoring programs have grown in quantity and popularity with applications like Lose It!TM and MyFitnessPalTM. Our objective was to investigate the effects of electronic self-monitoring behavior (using the commercial Lose It!TM application) and weight-loss interventions (with differing amounts of counselor feedback and support) on 4 and 12-month weight loss. In this secondary analysis of the Fit Blue study, adapted from the Look AHEAD Intensive Lifestyle Intervention for a military lifestyle, we compared the results from two interventions of a randomized controlled trial. Counselor-initiated treatment participants received early, consistent support from interventionists, and self-paced treatment participants received assistance on request. Participants (n= 191) who were active-duty military personnel were encouraged to self-monitor diet and exercise with the Lose It!TM application or website. We examined logging trends throughout the study for associations between intervention assignment and app usage. We conducted a mediation analysis of the intervention assignment on weight loss through multiple mediators: app usage, number of self-weighing, and 4-month weight change. We used linear regression to predict weight loss at 4 and 12 months, and we measured the model's accuracy using cross-validation. We found that app usage and daily reported caloric intake substantially impact weight loss prediction at 4 months. Our analysis did not find evidence of an association between participant self-monitoring exercise information and weight loss. Since the mediation analysis showed that 4-month weight loss completely mediated 12-month weight loss, intervention targets should promote early and frequent dietary intake self-monitoring and self-weighing to target early weight loss goals leading to long-term success.
The directory contains the supporting code necessary to run the analysis contained in the publication:
Farage, G., Gale1, C., Kocak, M., Klesges, R.C., Talcott, G.W., Richey, P., Hare, M., Johnson, K., Sen, S., Krukowski, R.A. (2021) Assessing the contribution of self-monitoring through a commercial weight loss app: mediation and predictive modeling. (submitted)
The directory src contains the code files that can reproduce the figures in the manuscript. The file description.Rmd generates the descriptive demographic table and frequency logging figure in the sub-directory' prediction'. The prediction.Rmd file produces the principal component analysis results, the accuracy 8-weeks models tables, and the linear regression coefficients tables. The mediation sub-directory contains the Jupyter notebooks that generate the results for the mediation analysis. They include the missing data's sensitivity results using two different imputations methods: the last observation carried forward analysis, and the baseline observation carried forward analysis for the 4-month and 12-month weight loss.