Long-term effectiveness of mHealth physical activity interventions: a systematic review and meta-analysis of randomized-controlled trials (Preprint)

BACKGROUND mHealth interventions can increase physical activity (PA), but their long-term impact is not well understood. The increasing number of primary studies reporting long-term follow-up measurements supports a meta-analysis of this evidence. OBJECTIVE This systematic review and meta-analysis aimed to understand the immediate and long-term impact of mHealth interventions on PA. The secondary objective was to explore potential effect moderators (population type, intervention design, control group type). METHODS We performed this systematic review according to the Cochrane and PRISMA guidelines. We searched PubMed, the Cochrane Library, SCOPUS and PsychINFO in July 2020. Eligible studies included randomized-controlled trials of mHealth interventions targeting PA as a primary outcome in adults. Eligible outcome measures were walking, moderate to vigorous PA (MVPA), total PA (TPA), and energy expenditure (EE). Where reported, we extracted data for three time points (ie, end of intervention, follow-up ≤6 months, follow-up >6 months). To understand effect moderators, we performed subgroup analyses by population, intervention design, and control group type. Results were summarized using random-effects meta-analysis. Risk of bias was assessed using the Cochrane Collaboration's tool. RESULTS Of the 2 828 identified citations, 117 studies were included. These studies reported on 21 118 participants with a mean age of 52.03 years (SD 14.14), of whom 59% were female. mHealth interventions significantly increased PA across all four outcome measures at the end of intervention (walking standardized mean difference 0.46, 0.36 to 0.55; P<.001; MVPA 0.28, 0.21 to 0.35; P<.001; TPA 0.34, 0.20 to 0.47; P<.001; EE 0.44, 0.13 to 0.75; P=.01). Only 33 studies reported short-term and eight studies reported long-term follow-up measurements in addition to end of intervention results. Effects were sustained short-term for walking (0.26, 0.09 to 0.42; P=.002), MVPA (0.20, 0.05 to 0.35; P=.008), and TPA (0.53, 0.13 to 0.93; P=.009). Long-term, effects were also sustained for walking (0.25, 0.10 to 0.39; P=.001) and MVPA (0.19, 0.11 to 0.27; P<.001). We found study population to be an effect moderator, with higher effect scores in sick and at-risk populations. PA was increased both in scalable and non-scalable mHealth intervention designs and regardless of the control group type. The risk of bias was rated high in 94 of 117 studies. In addition, heterogeneity was significant and substantial across outcome measures, resulting in low to very low quality of evidence. CONCLUSIONS mHealth interventions can foster small to moderate increases in PA. Effects are maintained long-term, but effect size decreases over time. The results encourage the use of mHealth interventions in at-risk and sick populations, and support the use of scalable mHealth intervention designs to affordably reach large populations. However, given low evidence quality and high heterogeneity, further methodologically rigorous studies are warranted to evaluate the long-term effects of mHealth interventions. CLINICALTRIAL This review is registered with PROSPERO (CRD42019124716).