Effects Of Mhealth Apps On Physical Activity And Weight Loss Outcomes

Abstract

PURPOSE: Little systematic review has been conducted on mobile device health applications (a.k.a., mHealth apps) frequently used in health promotion. This meta-analysis synthesized evidence regarding the effectiveness of mHealth apps on physical activity (PA) and weight loss (WL). METHODS: A total of 150 published articles regarding mHealth apps and PA/WL were found. Twelve studies met the following inclusion criteria: 1) empirical articles published between 2000 and 2016 in English; 2) examined the effectiveness of an mHealth app on PA or WL outcomes; and 3) each study contained ≥ 1 comparison. Data extraction for comparisons was completed for the following outcomes: 1) PA; 2) WL; and 3) body mass index (BMI). Moderator analyses were performed for studies reporting intervention fidelity frequency of: 1) greater than every two weeks; 2) less than every two weeks; or 3) not reported. Calculation of effect size (ES; Hedge’s g) was completed with Comprehensive Meta-Analysis software for each entry. Analyses were run separately between mHealth apps and control (i.e., standard care or no treatment) or comparison (i.e., another experimental treatment) conditions. RESULTS: mHealth apps were most commonly used for WL (n= 10). Compared to control and comparison, mHealth apps had no effect on WL (ES= .02 and ES= .08, respectively; all p > .05). Seven and six studies investigated mHealth apps effect on BMI and PA, respectively (one study examined both outcomes). Regarding BMI, mHealth apps had no effect versus control (ES= -.08, p > .05) or comparison (ES= .09, p > .05). The same result was seen when using mHealth apps to promote PA outcomes. Specifically, when compared to control, mHealth apps demonstrated no effect in promoting increased step counts (ES= -.10, p > .05) and moderate-to-vigorous PA (ES= -.12, p > .05). Intervention fidelity was only reported by 4 of the included studies, with moderator analyses revealing no effect of intervention fidelity frequency on the study outcomes (all p > .05). CONCLUSIONS: Findings indicate mHealth apps to be at least as effective as standard/no treatment or another experimental treatment condition. To improve effectiveness of mHealth apps in promoting PA and WL, future studies need to improve intervention fidelity measures and use established behavioral theory to implement the study.