Examining the quality of evidence to support the effectiveness of interventions: An analysis of systematic reviews

Abstract

Objective: This analysis examines the quality of evidence (QOE) for 1472 outcomes linked to interventions where the QOE was rated in 42 systematic reviews of randomised clinical trials and/or observational studies across different topics. Setting: Not applicable. Participants: 76 systematic reviews. Primary and secondary outcome measures: Strength of evidence ratings by initial reviewers. Results: Among 76 systematic reviews, QOE ratings were available for only 42, netting 1472 comparisons. Of these, 57% included observational studies; 4% were rated as high and 12% as moderate; the rest were low or insufficient. The ratings varied by topic: 74% of the surgical study pairs were rated as low or insufficient, compared with 82% of pharmaceuticals and 86% of device studies, 88% of organisational, 91% of lifestyle studies, and 94% of psychosocial interventions. Conclusions: We are some distance from being able to claim evidence-based practice. The press for individual-level data will make this challenge even harder.