I would like to thank the authors for drawing attention to an important yet very controversial issue. There is no doubt that the introduction of the EWTD into practice in UK has raised several valid concerns. In their discussion, the authors state that they were surprised that they could not demonstrate a difference in the caseload difference before and after the introduction of the EWTD and as a person with some interest in statistics, I would like to ask the authors a question regarding their power analysis. First is regarding the clinically relevant difference in caseload, which statistically represents the magnitude of the departure f from the null hypothesis. I think the authors had set themselves a high target of detecting 20% reduction in caseload which, after doing simple mathematics,  means a reduction of about 88 cases per trainee per year. As a doctor in training, I would probably regard half of this number significant to me! The second issue is that such a large difference would be very difficult to prove statistically. I wonder if the authors think that another analysis aiming for a smaller (10% or 15%) reduction in caseload work would result in a different result.  My other comment is that the study, apparently, was aimed to achieve 80% power to detect a realistic and worthwhile difference (the authors accepted a type II error of 20%) which is the usual requirement. However, such obviously hard work undertaken by the authors might deserve keeping this figure (the probability of a Type II error) as low as feasible.