Journal de gastroentérologie clinique et d'hépatologie Libre accès

Abstrait

An Analysis of Fragility and Risk of Bias in Randomized Clinical Trials in Bowel Preparation Guidelines

Chris Chapman, Benjamin Howard, Cole Wayant and Matt Vassar

Objective: In this study, we use the Fragility Index and Cochrane’s Risk of Bias Tool 2.0 to analyze the randomized controlled trials underpinning the American Gastroenterological Association’s clinical practice guideline on bowel preparation before colonoscopy.
Design: All citations within the guideline were screened for specific criteria. We extracted bowel preparation outcome data from the included studies and used an online calculator to determine the FI and Fragility Quotient (FQ) (fragility index relative to study sample size). Risk of bias assessments was made using the Cochrane Risk of Bias Tool 2.0.
Results: The median FI for the 30 included trials was 7.5 events (IQR 3-11.75). The median FQ was 3.5 per 100 patients. The Risk of Bias Assessments resulted in the following classifications: 12: Low Risk, 2: Some Concerns, 16: High Risk.
Conclusion: RCTs in ACG Bowel Preparation guidelines were found to contain moderate fragility and relatively high risk of bias. Reporting fragility in RCTs will help appraisers of guidelines by indicating the robustness of the results. In this way, guideline writers will be in a better position to make recommendations. Likewise, preemptive evaluation of risk of bias will help identify key weaknesses underlying RCTs and add to their credibility in formulating recommendations.