Research waste in surgical randomized controlled trials.

Abstract

Research waste is a major challenge for evidence-based medicine. It implicates misused resources and increased risks for research participants. The aim of this study was to quantify constituent components of waste in surgical RCTs and explore targets for improvement. ClinicalTrials.gov was searched for RCTs registered between January 2011 and December 2012 using the keyword 'surgery'. The primary outcome was research waste, defined as non-publication, inadequate reporting or presence of an avoidable design limitation. Serial systematic searches of PubMed and Scopus databases were performed to determine publication status. Adequacy of reporting was assessed using the CONSORT checklist. Avoidable design limitations were evaluated according to the presence of bias and/or the absence of a cited systematic review of the literature. Of 5617 registered RCTs, 304 met all eligibility criteria. Overall, 259 of 304 (85·2 per cent) demonstrated at least one feature of waste. Of these, 221 (72·7 per cent) were published in a peer-reviewed journal and 219 were accessible for full-text review. Only 73 of 131 (55·7 per cent) RCTs with a pharmacological intervention and 24 of 88 (27 per cent) with a non-pharmacological intervention were reported adequately, and 159 of 219 (72·6 per cent) demonstrated an avoidable design limitation. Multicentre (odds ratio 0·31, 95 per cent c.i. 0·11 to 0·88) and externally funded (OR 0·35, 0·15 to 0·82) RCTs were less associated with research waste. This study identified a considerable burden of research waste in surgical RCTs. Future initiatives should target improvements in single-centre, poorly supported RCTs.