Methods: We investigated potential moderators of these effects, by taking into consideration the length of interventions, other programme components, the type of SMTs (commercial and non-commercial), the type of target behaviour, and the quality of the RCTs. We excluded correlation studies and research carried with children populations or those having medical conditions.
Findings: Our analyses on 34 studies showed that overall, the quality of RCTs was satisfactory, with some underpowered studies. The overall effect size on behavioral outcomes in lifestyle interventions was small, d= 0.195, CI95%= [0.110, 0.280] though significant (Z=4.49, p< 0.001). There were no differences in effectiveness between types of interventions using SMTs, such as for physical activity, weight loss, quitting smoking, or alcohol intake, Q(2)= 1.823, p= .402. Commercial SMTs, such as Facebook had medium to large effects in interventions, d= 0.410, CI95%= [0.198, 0.622], while non-commercial ones such as blogs, internet discussion forums or message boards had small to moderate effect sizes, d= 0.199, CI95%= [0.018, 0.381].
Discussion: Introducing SMTs in behavior change interventions is a promising approach. However, well conducted RCTs that incorporate SMTs in theory-driven, well designed and adequately powered interventions should be prioritized before integrating SMTs in actual health promotion interventions.