Trainee fatigue: Are new limits on work hours enough?

It should come as no surprise that working in a busy hospital environment for 60 to 80 hours each week engenders a significant amount of sleep deprivation, with all of the attendant physiologic consequences. Sleep deprivation of the magnitude experienced by medical trainees has been shown to impair performance, worsen mood and compromise patient and provider safety.¹ In North America, new guidelines and requirements should eliminate the egregious 100- to 120-hour workweek that used to be all too common for residents. Despite this, clinicians, academicians and training-program coordinators are still grasping for the right balance of safety, education and service in residents' workloads. Established methods of residency training may have to be altered substantially to accommodate the reduction of work hours.

Resident well-being, patient safety and the anticipation of government regulations have all been cited as motivation for limiting residents' work hours. The agreement set forth in July 2000 by the Professional Association of Internes and Residents of Ontario and the Ontario Council of Teaching Hospitals limits the number of call periods required of residents.² Similarly, in July 2003 the Accreditation Council for Graduate Medical Education in the United States instituted common program requirements that limit work hours and call periods for all residency programs.³ These imposed limits still allow for very long work hours that are outside the limits imposed by other hazardous industries.⁴ In fact, it is now commonplace in the transportation industry for investigations to evaluate the work and sleep schedules of those involved in a crash.⁵ If we were to impose such investigative techniques to error analysis in health care, fatigue would certainly be cited as a contributing factor in many adverse events.

Many studies on the effects of fatigue and long work hours on medical trainees have been published over the past 25 years. Most have attempted to focus on the performance of subjects in both a fatigued and rested state, testing the hypothesis that performance will be worse in a condition of greater fatigue. However, many of these studies have failed to define fatigue rigorously and have not included the cumulative effects of weeks and months of sleep deprivation. This can mean that some studies actually compare groups that have similar levels of cumulative fatigue.⁶ The physical stress of sleep deprivation and fatigue has been studied in other populations, but has been largely absent from the literature evaluating the effects of fatigue on clinicians.⁷

In this issue (page 965), Parshuram and colleagues describe a small prospective study of critical care fellows who were working under the Ontario guidelines that limit call periods, on average, to every fourth night.⁸ Beyond the largely expected results (long shifts, little sleep, many interruptions), some new and important physiologic information has been revealed that opens the door for larger studies. In a subset of fellows studied with continuous Holter monitoring, heart rate variability showed dominance of the sympathetic nervous system. This physiologic alteration might have been due to the inherent stress of being on call, or to the increased physical activity of the subjects, who walked an average of 6.3 km during their shift. The authors also noted that some fellows neglected their own basic needs, as their urine specific gravity and ketone levels indicated that they were dehydrated.

Their study suggests that a potential strategy to mitigate some of the effects of fatigue would be to add another physician to the call team, thus distributing the workload among more people. The cost of such a strategy, which would likely be quite substantial, is not addressed.

Although the number of subjects was small, Parshuram and colleagues' work provides further evidence that, even when it is compliant with “new and improved” requirements, clinical work alters normal physiology. The authors should be congratulated: field studies of this type are difficult to perform on busy clinicians but are an important adjunct to laboratory-based investigations. Although this study extends existing research, larger studies are still required to better elucidate the effects of our work environment on practitioner physiology and performance and on the safety of both practitioners and patients. In this way we will be better able to monitor the changes that will be promulgated within health care circles, whether adopted voluntarily or imposed externally through regulation or legislation, and to determine whether they are sufficient to improve performance and enhance safety.

β See related article page 965