Evolution of reports of randomised clinical trials in plastic surgery

Abstract

Well-designed, well-conducted and well-reported randomised clinical trials (RCTs) can significantly impact medical care, by contributing to a strong evidence base from which clinical guidelines can be derived. in a previous study, we assessed the quality of reports of RCTs in plastic surgery published from 1966 to 2003. the aim of the present study was to verify what have changed over the last years. RCTs in plastic surgery published from 2004 to 2008 were identified through electronic searches, and classified according to their allocation concealment. Trials with allocation concealment appropriately described were evaluated as to their quality. Two independent reviewers performed the evaluations, using two tools: the Delphi List and the Jadad's quality scale. From 3840 identified studies, 96 were selected for classification according to allocation concealment; 28 (29%) of them appropriately described allocation concealment. From 1966 to 2003, 34 (17%) RCTs appropriately described allocation concealment (chi(2) = 22.98, p < 0.000). in the evaluation of the 28 RCTs by the Delphi List, the agreement coefficient between raters (kw) was 0.46 (z = 7.24, p < 0.000). Groups were similar at baseline in 96.4% of these trials, and this was the only item of the Delphi List, which significantly improved when compared with the period from 1966 to 2003 (chi(2) = 18.53, p < 0.000). When evaluated by Jadad's criteria, 14% of the RCTs were scored two points or less and thus considered of low quality (kw = 0.72, z = 8.57, p < 0.001). From 1966 to 2003, 59% of RCTs were scored two points or less (chi(2) = 17.07, p < 0.004). We concluded that the quality of reports of RCTs in plastic surgery (as measured by the Jadad's criteria and only one component of the nine components of the Delphi List) significantly increased over the last years. (C) 2010 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published by Elsevier B.V. All rights reserved.