Improving GRADE Evidence Tables Part 4: A Three-Arm Non-Inferiority Randomized Trial Demonstrates Improved Understanding of Content in Summary of Findings Tables with a New Format

Objectives: To evaluate alternative formats of summary of findings (SoF) tables for single comparison with multiple outcomes.Study Design and Setting: We conducted a three-arm randomized controlled non-inferiority trial (RCT) in systematic review (SR) users, researchers, clinical practice guideline developers, policy makers, and knowledge transfer organizations to measure understanding, accessibility, satisfaction, and preference across the current GRADE SoF, an alternative GRADE SoF, or an adapted Evidence-based Practice Center (EPC) Program SoF table.Results: 179 participants were randomized, and 129 participants completed the RCT (n=47 current GRADE, n=41 alternative GRADE, n=41 adapted EPC). Understanding the certainty of evidence and treatment effect was comparable across groups. The adapted EPC SoF table was inferior for quantifying risk and risk difference compared to the alternatives (<35% correct vs >85% correct). Participants reported increased satisfaction when SoF tables presented NNT, anticipated absolute effect differences, and narrative syntheses for evidence that could not be meta-analyzed. Participants reported accessibility to information as significantly better in both GRADE SoF tables, when compared with the adapted EPC SoF table. Participants preferred the alternative GRADE SoF table format.Conclusions: The alternative GRADE SoF table is a promising format for SR users preferring a comprehensive presentation of SR results for single comparisons.