Systematic review of reporting benefits and harms of surgical interventions in randomized clinical trials

Abstract

Background Standardized reporting methods facilitate comparisons between studies. Reporting of data on benefits and harms of treatments in surgical RCTs should support clinical decision‐making. Correct and complete reporting of the outcomes of clinical trials is mandatory to appreciate available evidence and to inform patients properly before asking informed consent. Methods RCTs published between January 2005 and January 2017 in 15 leading journals comparing a surgical treatment with any other treatment were reviewed systematically. The CONSORT checklist, including the extension for harms, was used to appraise the publications. Beneficial and harmful treatment outcomes, their definitions and their precision measures were extracted. Results Of 1200 RCTs screened, 88 trials were included. For the differences in effect size of beneficial outcomes, 68 per cent of the trials reported a P value only but not a 95 per cent confidence interval. For harmful effects, this was 67 per cent. Only five of the 88 trials (6 per cent) reported a number needed to treat, and no study a number needed to harm. Only 61 per cent of the trials reported on both the beneficial and harmful outcomes of the intervention studied in the same paper. Conclusion Despite CONSORT guidelines, current reporting of benefits and harms in surgical trials does not facilitate clear communication of treatment outcomes with patients. Researchers, reviewers and journal editors should ensure proper reporting of treatment benefits and harms in trials.