The authors did assess publication bias, and found evidence of it for some analyses, but no attempt was made to adjust the pooled effects. The authors investigated the sources of heterogeneity but only through subgroup analysis for intervention type.