The Effectiveness of Interactivity in Improving Mediating Variables, Behaviors and Outcomes of Web-Based Health Interventions: A Meta-Analytic Review

ABSTRACT

Despite the increasing amount of research investigating health interventions that applies to interactive computer technology, the effect sizes in Cohen’s d obtained across these studies range from −0.32 to 1.74. The lack of systematic review of interactive health interventions leaves their overall effectiveness unknown. To address this, a meta-analysis of 67 studies examining the effects of web-based interactive health interventions was conducted. Results indicated that web-based interactive health interventions were effective in general, but the effects were moderated by health topic, theoretical framework, and design of treatment and control groups. The unique advantage of interactivity was small but significant when comparing to health interventions with comparable information in non-interactive version. Theoretical and practical implications of findings were discussed.

Notes

1. The inter-coder reliability of coded variables was .85 for the operationalization of media, message, and source interactive features, and 1.0 for the ESs and all other moderators (i.e., human and medium interactivity, information comparability of the control group, study design, and frequency of intervention messages).

2. According to Borenstein et al. (2009), fixed-effect model with categorical moderator assumes that all studies in one subgroup share a common effect size, while mixed-effected model allows true variation of effects within the subgroups of studies.

3. The categorization of health topics was based on Healthy People 2020 (2018) at http://www.healthypeople.gov/2020/topicsobjectives2020/default.

4. Nutrition, physical activity, and obesity was combined as one category based on the Prevention Status Reports (PSPs) by Centers for Disease Control and Prevention (CDC, 2015).