Guideline development and Cochrane reviews: criteria and principles for selecting systematic reviews to answer guideline clinical questions

Article type
Authors
McInnes E, Gibbons E, Westby M, Gray W
Abstract
Background: This poster will build on previous work which has outlined the uses, benefits and disadvantages of using Cochrane reviews in guideline development. Guideline developers need to be aware of the shortcomings associated with using existing systematic reviews. There is frequently a tension between the answerable review question provided by existing systematic reviews and the question that would meaningfully address the multifaceted clinical questions set by a clinical guideline. Time-pushed guideline developers therefore need to guard against an overly expedient approach when selecting and using available systematic reviews and avoid being data-driven by available reviews. Existing reviews are a valuable resource for guideline developers often saving much work, but clear decisions about their use, application and revision are needed.

Objectives: To present a) criteria for selecting appropriate reviews to help answer guideline clinical questions; b) a decision table to assist guideline developers about whether to update, revise or re-do Cochrane reviews and c) guidance on how to interpret and report existing reviews, particularly when only using selected information from the review.

Methods: Several scenarios will be presented to illustrate the use of the suggested criteria and decision principles. For example: review fits with clinical guideline question; review only partly fits in with clinical guideline question; review reports secondary outcomes of interest; review results can be extrapolated to answer clinical questions. These scenarios will be matched with a suggested approach for using, reporting and interpreting the information from the systematic reviews. Examples will be derived from recent guidelines which have used existing Cochrane reviews and will include instances of misuse of reviews where inadequate thought has been given to their appropriateness and relevance because of lack of a systematic selection process.

Conclusions: The criteria and principles presented will assist guideline developers to make valid decisions about the use of existing reviews and also suggest appropriate reporting formats. It is expected that adhering to these criteria and principles will avoid a disproportionate amount of time being spent on reproducing or updating reviews that may mislead or compromise the information needed to answer the guideline clinical questions.