Aaron Bowers1, Chase Meyer2, Daniel Tritz2, Courtney Cook2, Kaleb Fuller2, Caleb Smith2, Brian Diener3, Matt Vassar2. 1. College of Osteopathic Medicine, Oklahoma State University Center for Health Sciences, Tulsa, Oklahoma. Electronic address: aaron.bowers@okstate.edu. 2. College of Osteopathic Medicine, Oklahoma State University Center for Health Sciences, Tulsa, Oklahoma. 3. Department of Surgery, Oklahoma State University Medical Center, Tulsa, Oklahoma.
Abstract
BACKGROUND: Recent studies have highlighted the risk of bias and the fragility of results in randomized controlled trials (RCTs). The aim of our study was to evaluate the clinical practice guidelines created by the Society for Gastrointestinal and Endoscopic Surgeons (SAGES) for fragility, statistical power, and risk of bias. MATERIALS AND METHODS: We screened the SAGES clinical practice guideline references for qualifying RCTs. RCTs were assessed for risk of bias using the Cochrane Collaboration Risk of Bias tool 2.0. We used the fragility index and fragility quotient to evaluate the robustness of trial results and conducted a power analysis using G*Power to determine if trials were adequately powered. RESULTS: Twenty-two (40.7%) of the 54 trials that we assessed were rated as having a high risk of bias, 17 (31.5%) were rated as having a low risk of bias, and 15 (27.8%) were rated as having some concerns. The median fragility index was 2.5 (interquartile range 1-7). The median fragility quotient was 0.021 (interquartile range 0.003-0.045). Mean sample size was 108, and the mean loss to follow-up was eight patients. Eight of 33 trials (24.2%) were found to be underpowered according to the sample size used in the primary outcome. CONCLUSIONS: Guidelines created by SAGES are supported by RCTs that are frequently fragile or underpowered or have a high risk of bias. Future RCTs should utilize the Consolidated Standards of Reporting Trials statement, implement strategies to minimize loss to follow-up, and use properly powered sample sizes.
BACKGROUND: Recent studies have highlighted the risk of bias and the fragility of results in randomized controlled trials (RCTs). The aim of our study was to evaluate the clinical practice guidelines created by the Society for Gastrointestinal and Endoscopic Surgeons (SAGES) for fragility, statistical power, and risk of bias. MATERIALS AND METHODS: We screened the SAGES clinical practice guideline references for qualifying RCTs. RCTs were assessed for risk of bias using the Cochrane Collaboration Risk of Bias tool 2.0. We used the fragility index and fragility quotient to evaluate the robustness of trial results and conducted a power analysis using G*Power to determine if trials were adequately powered. RESULTS: Twenty-two (40.7%) of the 54 trials that we assessed were rated as having a high risk of bias, 17 (31.5%) were rated as having a low risk of bias, and 15 (27.8%) were rated as having some concerns. The median fragility index was 2.5 (interquartile range 1-7). The median fragility quotient was 0.021 (interquartile range 0.003-0.045). Mean sample size was 108, and the mean loss to follow-up was eight patients. Eight of 33 trials (24.2%) were found to be underpowered according to the sample size used in the primary outcome. CONCLUSIONS: Guidelines created by SAGES are supported by RCTs that are frequently fragile or underpowered or have a high risk of bias. Future RCTs should utilize the Consolidated Standards of Reporting Trials statement, implement strategies to minimize loss to follow-up, and use properly powered sample sizes.
Authors: Muhammad Shahzeb Khan; Rohan Kumar Ochani; Asim Shaikh; Muhammad Shariq Usman; Naser Yamani; Safi U Khan; M Hassan Murad; John Mandrola; Rami Doukky; Richard A Krasuski Journal: Circ Cardiovasc Qual Outcomes Date: 2019-12-11
Authors: Muhammad Shariq Usman; Muhammad Shahzeb Khan; Gregg C Fonarow; Stephen J Greene; Tim Friede; Muthiah Vaduganathan; Gerasimos Filippatos; Andrew J Stewart Coats; Stefan D Anker; Javed Butler Journal: ESC Heart Fail Date: 2022-01-13
Authors: Matthew Vassar; Sam Jellison; Hannah Wendelbo; Cole Wayant; Harrison Gray; Michael Bibens Journal: BMJ Open Date: 2019-09-06 Impact factor: 2.692