Ten studies (4,984 participants) were included in the review. There were seven randomised controlled trials and five pre-and post-test pilot studies. The attrition rate ranges from zero to 52%. Four studies reported adequate sequence generation, allocation concealment and adequate protection against contamination. Incomplete data was addressed in five studies. Blinding was described in seven studies, as was freedom from selective reporting and incentives. Similarity in baseline characteristics were reported in seven studies, selection bias was accounted for in four studies. No single study scored positive on all quality criteria.
Effect size calculation was possible in seven studies, in which six studies presented eight effect sizes indicating the intervention was effective. The mean effect size for online intervention groups compared with offline control groups was 0.19 (±0.21), and compared with online control groups was 0.39 (±0.37).
Single component interventions (both online and the online comparison groups) had mean effect sizes of 0.15 (±0.20) and the complex multi-component offline interventions had effect sizes of 0.19 (±0.21) compared with online interventions. The complex multi-component interventions had a mean effect size of 0.51 (±0.33) compared with online interventions.
Two studies provided information on dose-response relationships. Meeting a login goal over 10 weeks significantly increased weight loss compared with those who used logins for less than 10 weeks. In another study, the intervention group with more exposure to a web site increased physical activity significantly more than the control group.
No large differences between the different focus areas were found.