The review had clear inclusion criteria. The authors searched a range of relevant databases, but no attempt to locate unpublished studies was reported and only English language studies were eligible for the review. This meant that the review could be at risk of publication and/or language bias. Studies were selected for the review by two independent reviewers, minimising risk of errors or bias, but the methods used for data extraction were not reported.
The authors did not assess the validity of the included studies, so the risk of bias in these trials and the synthesis derived from them is unclear. It was not reported whether the included RCTs used cluster randomisation (common in trials of educational interventions) and, if so, whether they were analysed correctly. Meta-analysis was performed for the main outcome. The heterogeneity of the included interventions and the presence of significant statistical heterogeneity suggest that meta-analysis was probably not appropriate. The results of the meta-analysis may not be meaningful or generalisable beyond the individual included trials.
Although the evidence presented tends to support the authors' conclusions, uncertainties about the strength of the evidence and methodological weaknesses in the review mean that the review should not be relied upon. The authors' recommendations for further research seem appropriate.