Design, Conduct, and Analysis of Surgical Randomized Controlled Trials

Jiajie Yu,Wenwen Chen,Shidong Chen,Pengli Jia,Guanyue Su,Youping Li,Xin Sun
DOI: https://doi.org/10.1097/sla.0000000000002860
IF: 13.787
2019-01-01
Annals of Surgery
Abstract:Background: Randomized controlled trial (RCT) testing surgical intervention faced challenges due to complexities of surgery and made it more difficult for surgeons and methodologists than pharmaceutical providers to build a well-design, conduct RCT. Objective: We conducted a cross-sectional survey to address the methodological challenges of RCTs on surgical intervention and offer potential solutions. Methods: We searched PubMed in order to summarize 2-arm parallel randomized trials for surgical interventions published in 2013. The information regarding general characteristics, general methodological and special surgical characteristics related to surgical trials comparing alternative procedures was gathered. Results: Some 200 surgical trials were identified. The extent to which these trials in design, conduct and analysis differed substantially across items. The general information about sample size calculation (77.0%), lost to follow-up (71.5%), trial registration (55.5%), protocols of trials (56.0%), implementation of randomization (59.5%), concealment of randomization (56.0%); reporting of primary outcome as P value (67.0%). Surgery special information revealed that only 21.0% of trials considered surgeons’ preference, approximately 12% to 50% of them controlled the quality of surgical interventions and none evaluated the effect of the learning curve. Conclusion: There is much room for improvement concerning the reported designs, conduct, and analysis of surgical RCTs. Considering the difficulty of surgical RCTs, some other approaches, such as surgeons’ eligibility, performance of pilot studies, or implementation of pragmatic RCTs/expertise-based trials, should be feasibly implemented to overcome the presented challenges.
What problem does this paper attempt to address?