Duty Hours: Where Do We Go From Here?

In this issue of Mayo Clinic Proceedings, 2 articles explore the effects that limiting resident physician duty hours have had on the learning environment and patient care.^1,2 These articles will prompt discussion among resident physicians and their faculty and evoke memories for those of us who completed residency training in an earlier era. In this regard, I am no exception.

As a “rotating” intern at Harborview Medical Center in Seattle, WA, 40 years ago, I endured 4 months on the surgery service of “2 nights on, 2 nights off.” Starting rounds at 6 am on Monday and coming home at 6 pm on Wednesday meant a 60-hour shift. Admissions and emergency surgeries at all hours left little time for food, showers, or naps. Thursday was just a 12-hour day with no admissions, but the cycle resumed early Friday morning. My 4 months on the internal medicine service were more humane: I was on call just every other night. Once, after I had finished a marathon shift, my wife took a picture of me asleep in a recliner with my infant daughter in my arms.

Although these schedules were arduous, I never felt alone or abandoned. A seasoned resident was always on service with me, and my learning curve was steep. I have no regrets and still revere the experience. This crucible, shared by many in my generation, taught me that a good and effective learning environment has many aspects—and that hours worked is only one. (But that is enough personal nostalgia; there is no way we will return to such schedules.)

Journalist Sidney Zion brought the issue of resident duty hours to public attention after the death of his 18-year-old daughter, Libby, at a New York hospital in 1984. Although the root cause of her death was really lack of supervision, not excessive duty hours, New York instituted duty hour regulations for residents in 1989, limiting weekly service to 80 hours and shift lengths to 24 hours. At that time, I was an associate dean and medical director of a large community teaching hospital in Massachusetts. Our faculty physicians began to emulate New York duty hour limits for our residents, anticipating that similar laws could be enacted in the Commonwealth.

Despite a tenuous link to the quality of patient care, resident hours have become the all-too convenient explanation for a variety of systemic ills and inefficiencies in health care. Outside agencies have pressured the Accreditation Council for Graduate Medical Education (ACGME) to limit duty hours as a way to improve patient care, even though the number of hours that residents work is only one aspect of their clinical learning environment. Furthermore, frequent changes to the duty hour requirements have made the evaluation of the actual effect of past changes almost impossible.

See also pages 185 and 192

For example, in a misguided attempt to increase patient safety and reduce medical errors, Public Citizen compared residents' schedules to those of workers in other industries with regulated work hours, such as truck drivers and airline pilots, and in 2002 asked the Occupational Safety and Health Administration (OSHA) to regulate and reduce work hours for resident physicians.³ However, a physician is not a truck driver, and a very sick patient is not an 18-wheeler that can be pulled into a rest stop on the expressway. Although OSHA wisely declined Public Citizen's request, regulation was in the air, and resident duty hours legislation was introduced in 2003 by US Senator Jon Corzine and Representative John Conyers. Moving quickly to avert a Congressional mandate, the ACGME developed and implemented new standards, effective July 2003, that limited residents' duty to 80 hours per week and in-house call to every 3 nights, as well as requiring that 1 in 7 days be free of patient care and educational obligations (all averaged over 4 weeks). Additionally, ACGME standards called for a 10-hour rest period between duty periods and a 24-hour limit on continuous duty, with up to 6 additional hours for continuity of care and education. Admissions could not be accepted after 24 hours of continuous duty. To allow for flexibility, some specialties (eg, neurologic surgery) were permitted to apply to the ACGME for an 8-hour increase in weekly duty hours.⁴

The new standards were not embraced enthusiastically by the entire graduate medical education (GME) community. Many residents were reluctant to leave tasks unfinished or to shift care for sick and unstable patients to their colleagues. Program directors and educators were legitimately concerned that shorter shifts would mean more hand-offs and transfers of care, which are associated with their own risks for adverse events, and that educational goals would be compromised. These opinions persist, as documented in the current article by Antiel et al¹ in this issue of the Proceedings. Furthermore, ACGME enforcement of the standards at some institutions was perceived to be lax, and the issue of confidentiality for, and whistleblower protection of, residents who reported program violations became a concern. In addition, true anonymity is hard if not impossible to ensure for residents in smaller programs. Some residents reported intimidation and pressure by attending physicians and senior residents to underreport actual duty hours. These pressures, combined with residents' fears of negative consequences for programs, program directors, and their own careers (in the event of program probation or withdrawal), created powerful disincentives to honest and accurate reporting.

Amid these controversies, in September 2007, at the request of the US Congress and the Agency for Healthcare Research and Quality, the Institute of Medicine (IOM) appointed a committee to synthesize current evidence on medical resident schedules and health care safety and to develop strategies for implementing optimal work schedules to improve safety in health care.

The IOM committee gathered testimony from the accreditation and certification community, organized medicine, medical students, residents, patient safety advocates, and researchers on sleep and patient outcomes, as well as from program directors in primary and surgical specialties. Its report, released in December 2008, did not recommend a further reduction in residents' work hours from the ACGME's 80-hour limit (as some had speculated) but called for reducing the maximum number of hours without time for sleep to 16, allowing overnight call only with a required 5-hour nap period, increasing the number of days off, and restricting moonlighting during residents' off-hours. Other recommendations called for greater supervision of residents, limits on patient caseloads, increased interdisciplinary teamwork, and overlap in schedules during shift changes to reduce errors during hand-offs. The report noted that the major barriers to implementation of these changes were cost ($2.3 billion) and an insufficient health care workforce to substitute for residents. Nonetheless, the report called for action on all recommendations by December 2010.

In response, the ACGME initiated a systematic review of duty hours and the GME learning environment, with a goal to create more appropriate and flexible standards that recognized the challenges of training in each specialty.⁴ Since then, ACGME activities include the following:

• A duty hours symposium, held in March 2009, to obtain input from multiple perspectives and stakeholders

• A duty hours congress, held in June 2009, with more than 120 professional associations, program director organizations, and other groups submitting formal position papers and 44 providing oral testimony

• Three comprehensive reviews of the literature on duty hours and related topics (Darcy Reed, MD, coauthor of the Antiel article, was on one review team)

• Consultations with leading ethicists on the issues of professionalism surrounding duty hours

In June 2010, the ACGME released its proposed duty hour standards for comment; these were approved with minimal modifications by the ACGME Board at its September 2010 meeting for implementation in July 2011. The new standards retain the 80-hour limit per week, averaged over 4 weeks, but reduce shift lengths for first-year residents to no more than 16 hours and set stricter requirements for duty hour exceptions. In addition, they specify in greater detail the levels of supervision necessary for first-year residents; set higher requirements for teamwork, clinical responsibilities, communication, professionalism, personal responsibility, and transitions of care; establish graduated requirements for minimum time off between scheduled duty periods; expand requirements regarding patient care hand-offs; and call for alertness management and fatigue mitigation strategies to promote continuity of patient care and resident safety.

However, in September 2010, Public Citizen (in an apparent attempt to upstage the ACGME) again requested that OSHA assume regulatory powers over resident hours. In its letter, joined by the Committee of Interns and Residents and the American Medical Student Association, Public Citizen argued that resident health and well-being were endangered by existing ACGME rules and that OSHA regulations were needed. The American Medical Association, with input and support from both its Medical Student Section and Resident and Fellow Section, wrote OSHA to oppose this petition and to express strong support of the ACGME and its system of oversight.⁵ Other member organizations of the ACGME have also written to OSHA, supporting the ACGME's role and arguing that it is impossible to separate duty hours from the overall educational and clinical environment.

The fundamental issue for GME is the successful formation of competent physicians and surgeons who are well prepared to practice in their specialties without supervision. They must be ethical, altruistic, and capable of exercising personal responsibility (including the ability to recognize their personal limits). Resident duty hours are only one metric (albeit the one most easily measured) in the quality of GME and its delivery of patient care. Other equally important issues in the GME learning environment include quality of supervision, intensity of workload, rigor of hand-offs, sleep and rest (including circadian rhythms), work-life balance, peer interaction, teamwork, moonlighting, and at-home call.

Changes should be made on the basis of the best evidence. Although some (unsurprising) studies show that sleepy physicians do not perform quite as well as rested ones,⁶ data are scant on all the other factors that enter into both human cognitive and motor performance. Plus, we know that biologic variability among humans is great; we all know individuals who perform at a high level, day after day, on little sleep—the late Michael DeBakey comes to mind.⁷ Furthermore, human sleep science is still in its infancy. For example, we are just beginning to learn about the multiple biologic clocks that govern human circadian rhythms.⁸

Looking forward, what additional data are needed to improve the learning environment of residents while striving to improve patient safety and quality? I believe we need information from many sources to guide any subsequent changes in the regulations. This should include qualitative and descriptive studies about what works and what does not work, like the report in this issue of the Proceedings by McCoy et al.²

Our inquiries should address questions such as: What effect do resident duty hour limits have on the workload and learning of residents and on medical students? Do duty hour limits change the workload and teaching of attending physicians? For both resident education and patient care safety and quality, what is the relative importance of duty hour limits compared with other factors, such as appropriate supervision, hand-offs, patient continuity, attendance at rounds, teaching conferences, sleep, and work-life balance? Are residency training periods still adequate, especially in procedural specialties? If not, should some specialties consider extending the length of their training programs? Can we find ways to actually accelerate learning with use of modalities like simulation? Are current GME graduates prepared for their transition into practice? What relationships exist between duty hour limits and sleep? Do residents actually sleep more, and, if so, do they learn and perform better? Finally, and perhaps most importantly, have duty hour limits improved or degraded the professionalism of residents and their mentors?

Common sense tells us that schedules like the one I trained under are neither sustainable nor reasonable. Changing the rules in the absence of evidence is no strategy either. Maybe it is time to step back, take a deep breath, and wait until better evidence becomes available before we make wholesale changes to a GME system that, despite any shortcomings, is still among the best in the world at training physicians.