Quality of Reporting in Surgical Randomized Clinical Trials.

J. Yu,X. Sun,Y. Li,X. Li

DOI: https://doi.org/10.1002/bjs.10331

2016-01-01

Abstract:Background: RCTs testing surgical interventions can change clinical practice. The adequate reporting of surgical trials is an important issue.Methods: A cross-sectional survey was undertaken by searching PubMed for two-arm parallel randomized trials assessing surgical interventions published in 2003 or 2013. Quality of reporting was evaluated against the CONSORT checklist containing 29 items (standard CONSORT plus CONSORT Extension for Trials Assessing Non-Pharmacological Treatments). Univariable and multivariable linear regression was undertaken to explore factors associated with quality of reporting measured with CONSORT scores.Results: Some 120 trials were identified and included. The mean(s. d.) CONSORT score was 12.7(4.0). Trials published in 2013 achieved a higher CONSORT score than those published in 2003 (mean 14.5(3.8) versus 10.8(3.4) respectively; P < 0.001). The extent to which these trials met the requirement for CONSORT reporting differed substantially across items: four of 29 items were reported adequately across trials, and seven were reported adequately in less than 20 per cent of trials. Less than 40 per cent of the trials described the additional items related to surgical interventions and care providers (such as nursing care and anaesthetic management). In multivariable regression analysis, trials published in 2013 (coefficient 3.05, 95 per cent c. i. 1.89 to 4.20) and multicentre studies (coefficient 208, 0.79 to 3.37) were associated with statistically higher quality of reporting.Conclusion: The quality of reporting in surgical trials has improved in the past decade. Overall quality, however, remains suboptimal, particularly in relation to details regarding surgical interventions and management.