Evaluation of complex community-based childhood obesity prevention interventions

Background: Multi-setting, multi-component community-based interventions have shown promise in preventing childhood obesity; however, evaluation of these complex interventions remains a challenge. Objective: The objective of the study is to systematically review published methodological approaches to outcome evaluation for multi-setting community-based childhood obesity prevention interventions and synthesize a set of pragmatic recommendations. Methods: MEDLINE, CINAHL and PsycINFO were searched from inception to 6 July 2017. Papers were included if the intervention targeted children <= 18 years, engaged at least two community sectors and described their outcome evaluation methodology. A single reviewer conducted title and abstract scans, full article review and data abstraction. Directed content analysis was performed by three reviewers to identify prevailing themes. Results: Thirty-three studies were included, and of these, 26 employed a quasi-experimental design; the remaining were randomized control trials. Body mass index was the most commonly measured outcome, followed by health behaviour change and psychosocial outcomes. Six themes emerged, highlighting advantages and disadvantages of active vs. passive consent, quasi-experimental vs. randomized control trials, longitudinal vs. repeat cross-sectional designs and the roles of process evaluation and methodological flexibility in evaluating complex interventions. Conclusions: Selection of study designs and outcome measures compatible with community infrastructure, accompanied by process evaluation, may facilitate successful outcome evaluation.