We thank Chen and colleagues for their letter relating to our published article . While we agree that risk of bias and assessment of publication bias are important elements in systematic reviews, available tools to assess these were not designed to address meta-analyses of subgroup data and are not validated in this setting. For example, a randomised controlled trial may be of low risk of bias, but if a subgroup is not stratified, then many of the benefits of randomisation would not apply to the results of the subgroup data.
Response to: Assessing the risk of bias and publication bias should be integral parts of the systematic review
Leggi l'articolo originale