35 An ultra-rapid review approach: an innovative solution to make evidence more relevant for time-sensitive policy-making and providing guidance for health systems?. Issue Volume 24:Issue Supplement 1(2019) (July 2019)

• Record Type: Journal Article
• Authors: Affengruber, Lisa; Wagner, Gernot; Gartlehner, Gerald
• Is Part Of: BMJ evidence-based medicine. Volume 24:Issue Supplement 1(2019)
• Page Start: A23
• Page End: A23
• Publication Date: 2019-07
• Subjects: Evidence-based medicine -- Periodicals; 616.005
• DOI: 10.1136/bmjebm-2019-EBMLive.43
• Languages: English
• ISSNs: 2515-446X

Abstract

Abstract : Objectives: To make evidence relevant to the time-sensitive needs of decision-makers in policy and health care systems, rapid reviews have become a practical alternative to systematic reviews because of their fast provision of results. By streamlining methodological steps, the results of rapid reviews are available within a few weeks or months. Ultra-rapid reviews employ multiple methodological shortcuts, are conducted in a very short timeframe (a few days), and therefore provide evidence at the time needed. We tested the accuracy of an ultra-rapid review approach that substantially reduced the effort for literature searches and study selection. Our approach combined an abbreviated literature search with single-reviewer literature screening. The aim of this methods project was to evaluate the sensitivity of an ultra-rapid review approach based on three case studies and to ascertain if it leads to different overall conclusions about beneficial and harmful treatment effects compared with a systematic review approach. Method: The ultra-rapid review approach consisted of strongly abbreviated literature searches and single screening of abstracts and full texts. As reference standards, we selected three ongoing Cochrane reviews that had not been published at the time of the methods study. Two Cochrane reviews addressed oncological topics (OT), one a public health topic (PHT). A search specialist conducted abbreviated literature searches consisting of a simple-structured Boolean search in combination with the similar articles function in PubMed. Authors of the three Cochrane reviews provided highly relevant, published studies as starter sets for the similar articles function in PubMed. Three reviewers screened records independently and in parallel. If they missed studies compared with the reference standard, we recalculated effect estimates and created new Summary of Findings tables. We surveyed the Cochrane review authors to solicit whether the detected bodies of evidence from the ultra-rapid reviews would change the conclusions of the Cochrane review. Results: The proportion of abstracts screened by ultra-rapid reviewers compared to the Cochrane reviewers ranged from 15% to 1%. The ultra-rapid review approach identified 11% to 22% of the relevant publications (OT: 11%-15%; PHT: 13%-22%) and 38%-100% of the relevant studies (OT: 40%-100%; PHT: 38%-48%). For case study one (OT), the ultra-rapid review approach led to the same conclusions as the Cochrane review. For case study two (OT), the authors would have drawn the same conclusion for all three ultra-rapid reviews, but with less certainty. For case study three (PHT), the authors would have drawn the same conclusion for one ultra-rapid review, but with less certainty. For the other two ultra-rapid reviews, the authors assessed that a conclusion could not be drawn. Conclusions: For oncological topics, the conclusions of the ultra-rapid review approach changed less often compared to the public health topic. Based on these three case studies, an ultra-rapid review approach seems to be viable for narrow questions that can be answered with RCTs. However, to assess whether an ultra-rapid review approach could be an expedited solution for evidence-based decisions for policy or health care professionals further research with a larger sample size is necessary to confirm our findings.