Table 3 The general FEAT principles for critical appraisal of studies in evidence reviews and their specific interpretation for assessing threats to internal validity (risk of bias)

FOCUSED

Critical appraisal should be directed at key quality constructs that are relevant to the evidence review. Each construct should be appraised separately

In comparative quantitative systematic reviews, internal validity should always be assessed; the assessment of internal validity should be separate, not conflated with other quality constructs

EXTENSIVE

All important elements of the target quality construct should be identified and evaluated

All relevant threats to internal validity (i.e. all important individual sources of bias and confounding) relevant to the studies being assessed should be included in the appraisal

APPLIED

The appraisal process should logically inform the data synthesis, with accurate, consistent descriptions of the extent to which each element of each construct has been fulfilled

The internal validity assessment should inform the data synthesis in an appropriate format (e.g. to support sensitivity or subgroup analyses)

TRANSPARENT

Judgements should be made against explicit, unambiguous criteria. The reason for each quality judgement made by the reviewers should be clearly justified and transparently reported

All internal validity judgements should be based on pre-specified and agreed criteria detailed in the review protocol; each judgement should be supported with a concise explanation and grounded in evidence of the practices used in the study that is being appraised