The long hours worked by resident physicians, and the mandated reduction of those hours in recent years, are widely known. A related issue—perhaps an even more important issue—has so far received less attention. This was brought home to me one day, many years after my residency was over, during the care of an infant with a rare and frequently fatal birth defect in the newborn intensive care unit of the teaching hospital where I am an attending neonatalogist.

My day began around quarter after seven in the morning. I started by taking a look at the x-rays from overnight, then examining some of my patients before rounds. At eight o\'clock I began rounding with my team (one of three in the newborn intensive care unit), which consisted of a neonatal nurse practitioner, two pediatric residents, a neonatology fellow, and several nurses. We discussed the recent events, current status, outstanding problems, and plan for each patient.

Once rounds were done, the remainder of the day was filled with more physical exams, speaking with parents and consultants, teaching, documentation, and supervision of medical management and procedures for sixteen intensive care patients. The team regrouped at two o\'clock in the afternoon, as is our daily practice, to review the x-rays from the previous twenty-four hours with a pediatric radiologist. I was on call that night, so at around half past five, I took sign-out from the attendings for the other teams. I made it home by half past six, but at ten o\'clock that evening I received a call that a child with a prenatal diagnosis of congenital diaphragmatic hernia would soon be delivered, so I headed back to the hospital.

Diaphragmatic hernia is a birth defect that essentially consists of a hole in the diaphragm. During gestation, some of the abdominal contents (stomach, intestines, sometimes liver) migrate through the defect into the chest. This is usually associated with some degree of underdevelopment of the lungs (pulmonary hypoplasia). Sometimes the pulmonary problems are so severe that these newborns cannot survive despite all of our efforts, and even among those who do survive, the first hours and days are often quite difficult.

The child was born at about one o\'clock in the morning. The team did a good job in the resuscitation and management of the patient, who, as feared, was very sick from the outset. The resident, nurse practitioner, nurses, and respiratory therapist all acted quickly and skillfully, ably directed by the neonatology fellow. My role was to oversee their effort and provide back-up and guidance as needed. If they could not get the endotracheal tube in quickly, it would fall to me. If they were unsure how to prioritize the management, or optimize the mechanical ventilation, I would tell them. The fellow in this case, nearly finished with her neonatology training, did a fine job, with only an occasional suggestion or redirection from me. As the attending, though, it clearly fell to me to decide if and when to stop, which we did before the sun came up. And it fell to me to explain to the parents that we were unable to save their daughter.

Once the dust settled, it was time to begin my exams. Then came rounds, and essentially a repeat of the day before, supervising the care of the same sixteen intensive care patients. At two in the afternoon, we gathered to review the films. As each x-ray came up on the screen, a resident or practitioner provided a brief presentation of the case, and the radiologist gave his reading. After going through several films, a remarkable image appeared on the screen, and the room was silent. Many in the room recognized the pattern of intestinal gas in the chest as being diagnostic for diaphragmatic hernia, but no one else knew the case. The fellow, residents, and nurse practitioners who had cared for this child through the night had long since gone home.

When I was a first-year resident in the early 1980s, I was asked by an attending what I thought of the residency program. I told him I thought the residents were great, the attendings were great, and the system that required residents to work up to thirty-four hours at a stretch, and sometimes well over one hundred hours in a week, would be illegal in our lifetime. He laughed, gave me some much-appreciated encouragement, and sent me back to work. It had always (it seemed) been that way, and it always would. But that\'s not how things turned out.

As most readers are likely aware, that system, while not illegal, is no longer allowed in American training programs. Residents and fellows\' hours, while still sometimes grueling, have been significantly reduced. There is ample evidence that the long shifts can lead to increased medical and surgical errors. I wholly endorse the shorter shifts, and in fact I think there may come a time when we further reduce the maximum length of resident shifts. The current twenty-four-hour limit (plus six more allowed for sign-out, paperwork, and education) still seems inconsistent with what is known about human physiology and performance. In any case, the present limit is an improvement over the old system.

Presumably, the reduction is intended to protect patients from the mistakes that exhausted physicians are likelier to make. Nursing errors are, of course, also a concern, but to my knowledge, it has (thankfully) never been standard practice in American hospitals to require, or even allow, nurses to work anything approaching a twenty-four-hour shift. Nurse practitioners and physician assistants, who in many units function similarly to residents, are rarely if ever required to work in excess of twenty-four hours at a time.

The irony is that one exception to these rules persists. It struck me that day in x-ray rounds. Every clinician involved in patient care the previous night had gone home, for the protection of the patients and perhaps for their own good, except the oldest member of the team and the one with final responsibility for patient management in the unit.

It has been a few years since that day, but the rules remain the same. That is to say, on this topic, there are no rules. Attending physicians in many hospitals may work all day, through the night, and still have significant clinical responsibilities the next day.

What is the rationale? Continuity of care—trying to avoid passing a patient from one physician to another? Those who were around before the resident hours were reduced will recognize this as one argument used to justify the long shifts. There may be some truth to the idea that patient care could benefit from having one resident follow a patient over many hours. Perhaps there is some benefit for resident education, too. In the end, however, these potential benefits were trumped by concerns about medical errors and declining physician performance. It is difficult, then, to see how the same argument could justify allowing extended hours for attending physicians.

Another possible justification is manpower shortage. This one may at times be legitimate, particularly in isolated or underserved locations or one- and two-person subspecialty groups. If there are only one or two obstetricians in a hospital or community, for example, to limit allowable shifts does not seem feasible. But in hospitals with ample staff long shifts seem hard to justify given what is well described regarding physician fatigue and medical errors. And even if we acknowledge that limiting attending hours would sometimes be impractical, it does not follow that such guidelines should therefore never be put in place. Is it in the patients\' best interests for an obstetrician in a large group practice to see patients during the day, work in the delivery room much of the night, and then perform surgery the next day? Is it necessary?

A generation ago, many residents who were told to work long hours suspected that they really did it, at least in part, to save money for the hospital. And save money they did. When their hours were reduced and some resident shifts were replaced by nurse practitioners, these welcome reinforcements worked fewer than half the hours for nearly twice the salary. Financial concerns may similarly play a role in the current reluctance to change the rules for attendings. Private practices, medical schools, and hospitals that employ physicians could experience an increase in salary costs, or lose income, by limiting attending shifts, and in the current era, this is no small matter. But if it was not decisive in the debate about resident hours, then it is even less persuasive when we consider attendings, with their greater age and responsibility. If allowing a twenty-eight-year-old resident to work around the clock and beyond is unsafe for patients or the physician, then allowing a fifty- or sixty-year-old attending to do it makes little sense.