Critical appraisal of AMSTAR: challenges, limitations, and potential solutions from the perspective of an assessor

Systematic reviews are pivotal components in the development of evidence-based clinical guidelines. To evaluate the methodological quality of these systematic reviews, several tools have been proposed. Among them, the assessment of multiple systematic reviews (AMSTAR) checklist is probably used most frequently. This tool comprises 11 items related to the steps taken when conducting a systematic review, and it is claimed to have good face and content validity. The objective of this debate paper was twofold: (a) to critically evaluate the ability of all AMSTAR checklist items to adequately determine the methodological quality of a systematic review; and (b) to describe difficulties regarding interpretation of the checklist, and provide potential solutions for these difficulties.

Some items of the AMSTAR checklist seem to assess the quality of reporting of a systematic review more than its methodological quality. For example, item 7 may not "capture" the true methodological quality of primary studies included in the systematic review. Item 10 does not likely result in the collection of in-depth information on the presence of publication bias in the systematic review. Furthermore, some items may be difficult to interpret, hindering accurate assessment. For example, item 5 does not explicitly indicate whether a list of documents excluded in each phase of selection (i.e., after evaluation of titles and abstracts, and after full-text assessment) should be reported. The present debate paper evaluated and discussed some methodological limitations of the AMSTAR checklist, as well as challenges involved in evaluation of the checklist's items. Several suggestions are also made to optimize the use of this checklist. The information in this paper may stimulate further discussion among systematic reviewers, methodologists and clinicians.