Regulations intended to reduce resident physicians' work hours have been accompanied by controversy since their introduction in 2003. Although large-scale studies have demonstrated that the regulations were not associated with adverse clinical outcomes, their effect on patient safety remains unclear. This prospective cohort study combined data on actual hours worked and burnout (voluntarily reported by pediatric residents at three hospitals), with data on medication errors gathered through active surveillance. The most notable finding was that, despite the regulations, residents' total hours worked and sleep habits did not change. Extended-duration shifts (of more than 24 hours) remained common, and the majority of residents met the criteria for burnout, although the incidence did decrease. Medication errors and self-reported medical errors did not improve. The authors interpret these findings as demonstrating a need for further reduction in extended-duration shifts, which have been associated with an increased rate of errors.