Transparency in surgical randomised controlled trials: cross-sectional, observational study