Search tips
Search criteria 


Logo of bmjLink to Publisher's site
BMJ. 2002 November 23; 325(7374): 1184–1185.
PMCID: PMC1124675

Residents' hours of work

We need to assess the impact of the new US reforms
Ingrid Philibert, director of field activities
Accreditation Council for Graduate Medical Education, Suite 2000, 515 North State Street, Chicago, IL 60610-4322, USA (gro.emgca@TREBILIHPI)
Paul Barach, assistant professor

To many, “resident physician” conjures up an image of long hours of work, fuelled by caffeine and adrenaline. This overlooks the reality that residency is an educational experience that completes a physician's preparation for independent practice. About 100 000 resident doctors in the United States—as providers of care and as learners—will be affected by reforms regarding their hours of duty, which were recently announced by the US Accreditation Council for Graduate Medical Education.1 Under the new rules, set to take effect in July 2003, residents will work no more than 80 hours per week, have shifts that are no longer than 24 hours, and have 10 hours of rest between shifts.

The literature on sleep deprivation supports these reforms. Many articles show that sleep deprivation in laboratory and field studies has shown a negative effect on the performance of residents.24 Reduced performance due to sleep deprivation may be associated with increased errors and contribute to adverse events when fatigued members of staff participate in the care of patients.5

Accreditation is a voluntary approach of professional self regulation. The consequences of failed self regulation are often regulatory interventions, which are costly. The United States spends about $200bn (£128bn; €203bn) annually towards regulations related to health, safety, and the environment. Whether these achieve the desired outcome is quantified for only a third.6 The consequence of not assessing outcomes is that we do not know whether regulations have the intended effect. New York State, through its Department of Health, began to limit duty hours for resident doctors in 1989. Fifteen years later, the impact of this reduction on the safety of patients, education, and the professional lives of residents is still the subject of opinion and guesswork. Articles on New York's experience show the conflicting nature of the reports. Some noted that the limits had no effect on the care of patient and improved welfare of residents.7 Others found that they reduced the quality of care in teaching institutions.8 Reports that doctors trained under the limits were less familiar with their obligations to patients were countered by findings that residents did not want to leave patients until the process of care was completed.9

Absence of comprehensive objective data from the initiatives in New York State has reduced the value of this potential learning laboratory in guiding efforts to champion duty hours of residents. The academic community must not ignore the opportunity to benefit from the natural experiment that will result from the implementation of the new standards for hours of duty.

Beyond showing whether the standards will achieve their intended public goals, the results will help in developing new models for providing care with fewer hours for residents. Many institutions will use night float systems—a separate resident will be assigned to cover all or a part of the on-call hours of rotation and introduce other changes in how work is scheduled—or coverage by non-resident providers to comply with the standards, and some of these interventions, such as increased transfers of care and a growing workload for residents on call, in themselves have the potential to increase errors. Collecting data on costs, outcomes for patients, residents' education, and satisfaction for both groups will not be easy, and exploring the link between errors in health care and residents' hours or the interventions put in place to limit them is a needed, but highly complex, undertaking. Veasey et al noted that systematic research to address the causes of errors in health care is just beginning.10 Data are often incomplete and ill suited to making inferences about sleep deprivation in residents or that of interventions, such as more frequent transfers of the care of patients among physicians, as causes or contributing factors.

Metrics need to be developed for measuring the effect of the standards on safety of patients, learning, and wellbeing of residents. Residents' hours are a “political” matter as well as a scientific concern, and data collection must be objective and not influenced by a priori views of whether the changes are desirable or far reaching enough. The validity of a British survey was challenged when questions about the number of hours of duty were juxtaposed with questions about regret in choosing medicine as a career.11

Change is often sensationalised, as exemplified by an article on the regulation of working hours in France's restaurant industry, entitled “The death of French food.”12 In addition to negative views about the effort to limit hours, expectations may arise that it will have an immediate, profound, and measurable effect on the education of residents or safety of patients. When this does not occur, a too rapid consensus of “It did not make a difference” follows. What this does not consider is that we may not have measured the effect, may have looked in the wrong place, or may have interpreted the data incorrectly. In New York, the implementation of limits for duty hours did not result in a reported increase in residents' hours of sleep, but residents reported more time for reading and personal pursuits, and greater satisfaction. This has been interpreted as implying that the standards were superfluous. A better explanation is that the human need for sleep is pervasive, and individuals with long working hours will satisfy this need first (personal communication, Charles Czeisler, chief, Division of Sleep Medicine, Brigham and Women's Hospital, September 2002). The response to reductions in working hours is to resume some of the important activities that were foregone in favour of sleep. Studies are needed that combine institutional knowledge about how to create a better learning environment with broad based assessments that transcend those of individual institutions. This combination of information can contribute to improving the environment for training and patient care in teaching institutions.


Competing interests: PI is employed by the Accreditation Council for Graduate Medical Education.


1. ACGME Board of Directors. (accessed 5 November 2002) Chicago, Illinois.
2. Weinger M, Ancoli-Israel S. Sleep deprivation and clinical performance. JAMA. 2002;287:955–957. [PubMed]
3. Leung L, Becker C. Sleep-deprivation and house staff performance—update 1984-1991. J Occup Environ Med. 1992;34:1153–1160. [PubMed]
4. Samkoff J, Jacques C. A review of studies concerning effects of sleep-deprivation and fatigue on residents' performance. Acad Med. 1991;66:687–693. [PubMed]
5. Gaba DM, Howard SK. Fatigue among clinicians and the safety of patients. N Engl J Med. 2002;347:1249–1255. [PubMed]
6. Hahn RW, Burnett JK, Chan YI, Mader EA, Moyle PR. Assessing regulatory impact analysis: the failure of agencies to comply. Harv J Law Pub Policy. 2000;23:859–886.
7. Kelly A, Marks F, Westhoff C, Rosen M. The effect of the New York State restrictions on resident work hours. Obstetr Gynecol. 1991;78(3Pt1):468–473. [PubMed]
8. Holzman IR, Barnett SH. The Bell Commission: ethical implications for the training of physicians. Mt Sinai J Med. 2000;67:136–139. [PubMed]
9. Yedidia MJ, Lipkin M, Jr, Schwartz MD, Hirschkorn C. Doctors as workers: work-hour regulations and interns' perceptions of responsibility, quality of care and training. J Gen Intern Med. 1993;8:429–435. [PubMed]
10. Veasey S, Rosen R, Barzansky R, Rosen I, Owens J. Sleep loss and fatigue in residency training: a reappraisal. JAMA. 2002;288:1116–1124. [PubMed]
11. Beecham L. Junior doctors plan further hours' survey. BMJ. 1994;309:882.
12. Reynolds J. The death of French food (part 1). New York Times Magazine July 23 2000;149:57-8.

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Group