Active Video Game Interventions Targeting Physical Activity Behaviors: Systematic Review and Meta-analysis

BACKGROUND: Research on digital games designed to increase physical activity (PA), also known as exergames or active video games (AVGs), has proliferated over the past 2 decades. As a result, reviews of literature in this field can become outdated, revealing the need for updated high-quality reviews that identify overarching insights. Furthermore, given the significant heterogeneity in AVG research, study inclusion criteria may significantly influence conclusions. To the best of our knowledge, no prior systematic review or meta-analysis has specifically focused on studies of longitudinal AVG interventions targeting increases in PA behaviors. OBJECTIVE: The aim of this study was to obtain insights into when and why longitudinal AVG interventions are more or less successful for sustained increases in PA, especially for public health. METHODS: Six databases (PubMed, PsycINFO, SPORTDiscus, MEDLINE, Web of Science, and Google Scholar) were reviewed until December 31, 2020. This protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO: CRD42020204191). For inclusion, randomized controlled trials had to prominently (>50% of intervention) feature AVG technology, involve repeated AVG exposure, and target changes in PA behavior. Experimental designs had to include ≥2 within- or between-participant conditions with ≥10 participants per condition. RESULTS: A total of 25 studies published in English between 1996 and 2020 were identified, with 19 studies providing sufficient data for inclusion in the meta-analysis. Our findings indicated that AVG interventions had a moderately positive effect, thereby increasing overall PA (Hedges g=0.525, 95% CI 0.322-0.728). Our analysis showed substantial heterogeneity (I(2)=87.7%; Q=154.1). The main findings were consistent across all subgroup analyses. The comparison between PA assessment type groups showed a moderate effect for objective measures (Hedges g=0.586, 95% CI 0.321-0.852) and a small effect for subjective measures (Hedges g=0.301, 95% CI 0.049-0.554) but no significant difference between the groups (P=.13). The platform subgroup analysis indicated a moderate effect for stepping devices (Hedges g=0.303, 95% CI 0.110-0.496), combination of handheld and body-sensing devices (Hedges g=0.512, 95% CI 0.288-0.736), and other devices (Hedges g=0.694, 95% CI 0.350-1.039). The type of control group showed a wide range of effects sizes, ranging from a small effect size (Hedges g=0.370, 95% CI 0.212-0.527) for the passive control group (nothing) to a moderate effect size for the conventional PA intervention group (Hedges g=0.693, 95% CI 0.107-1.279) and ultimately to a large effect size for sedentary game as control groups (Hedges g=0.932, 95% CI 0.043-1.821). There was no significant difference among the groups (P=.29). CONCLUSIONS: AVGs represent a promising tool for PA promotion among the general population and clinical subpopulations. However, significant variabilities in AVG quality, study design, and impact were also detected. Suggestions for improving AVG interventions and related research will be discussed. TRIAL REGISTRATION: PROSPERO CRD42020204191; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=204191