Accuracy of “high-quality” rating using the AMSTAR 2 tool for systematic reviews of prospective, controlled, clinical therapy trials. An empirical investigation (Protocol)

Abstract Aim: To investigate the likelihood that systematic reviews with “high-quality” AMSTAR 2 rating are affected by high risk of systematic error (bias). Methods: A systematic literature search of current meta-reviews (systematic reviews of systematic reviews) will be conducted. From the identified meta-reviews, systematic reviews with “high” and “low” or “critically low” AMSTAR 2 quality rating will be extracted. All meta-analysis of the systematic reviews will be statistically tested for selection bias and the reviewed trials re-appraised for bias risk using the CQS-2B tool. From the test and re-appraisal results true positive, true negative, false positive or false negative results will be established and the negative and positive likelihood ratios with 95% Confidence interval computed. From the likelihood ratios the likelihood that systematic reviews with “high-quality” AMSTAR 2 rating are affected by high risk of systematic error will be inferred. Ethics and dissemination: Ethical approval is not required for methodological studies. The results will be disseminated as a prior preprint version and subsequent peer-reviewed publication.