We are intrigued by Mr Venkatesh's and Dr Noskin's hypothesis that further reductions in resident work hours might improve patient safety, thus producing cost savings that could mitigate the magnitude of expense that we have projected for this change.1 We hope that this will be the case.
We are concerned that it will not be so. The 2003 ACGME guidelines, in terms of patient safety, were based on supposition and common sense because there were few data linking resident fatigue to patient danger. Subsequent publications have identified fatigue as a risk to the residents themselves2- 3 and as a cause of inattentiveness.4 There is an association of long shifts with increased rates of medical error.5 It seems reasonable that extreme fatigue, as once might have been the case, combined with a lack of stimulation and a lack of supervision, could have led to mistakes. But does extreme fatigue happen now, with the enforcement of the 80-hour week and consequent limitation of extended shifts? More to the point, is a reduction from 80 hours to 70 or 60 hours going to have any impact on fatigue-related errors and therefore error-related costs? Any such beneficial result, in our opinion, would be far under the 19% to 31% error reduction required to make this “break even” by the Nuckols estimate of costs.6