Cochrane Public Health Project 1: Preferred and accepted risk of bias tools for assessing bias in non-randomised studies of interventions

"Comprehensive evaluations of public policies and programmes use a variety of methods to quantify impact. Non-randomised methods are at a higher risk of bias and confounding than randomised methods, but often are the only feasible methods to use. Decision makers and systematic reviewers therefore need a reliable way to assess the risk of bias and confounding in such studies. Although hundreds of different tools exist, there is no 'gold standard' and no guidance on how to select risk of bias tools. The lack of guidance causes confusion and inconsistency in Cochrane reviews, which are widely relied upon to inform decision making in policy and practice. The aims of this project are 1) Agree a list of criteria against which the acceptability of different NRSI critical appraisal tools can be judged;
(2) Produce an initial matrix of preferred and accepted tools mapped against their appropriateness for different types of NRSI;
(3) Outline additional work required to investigate further and/or provide more robust guidance."