No Doctor Should Work 30 Straight Hours Without Sleep

The American medical system requires dangerous feats of sleep deprivation. It doesn’t have to.

Fabrizio Bensch / Reuters

When Larry Schlachter was a 31-year-old neurosurgeon, he was driving to the hospital early one morning and “just blacked out.” He crashed his car and crushed his chest; broken ribs punctured his thorax, which filled with air and blood. “I almost died.”

Instead he was left with 14 fractured bones and a lingering loss of balance. He attributes the blackout to working 120-hour weeks that left him often on the brink of awareness. He put it to me clinically: “I was a victim of physician fatigue and exhaustion.”

Getting five or six hours of sleep—substantial by many physicians’ self-standards—can leave drivers impaired to a degree that’s similar to drunkenness. That’s according to findings of a study released this month from the AAA Foundation for Traffic Safety: Drivers who sleep only five or six hours in a 24-hour period are twice as likely to crash as those who got seven or more.

The finding led AAA’s director of Traffic Safety Advocacy and Research Jake Nelson to recommend on NPR:“If you have not slept seven or more hours in a given 24-hour period, you really shouldn't be behind the wheel of a car.”

So, should you be performing neurosurgery?

When the young Schlachter did come back to work, his damaged vestibular system proved less than optimal. “I lost my balance and just fell on top of one or two patients in the operating room,” he recalls.

Even if a surgeon doesn’t physically collapse on top of a person, drowsy doctors are more likely to experience lapses in memory and judgment that can prove critical. In other words, the brains of doctors are subject to the limits of physiology in much the same way as other human brains.

In this month’s issue of The Atlantic, I wrote about my experience with sleep deprivation during medical training, and since publication, I keep hearing iterations of the same response—a version of what this caller asked on a Wisconsin Public Radio show on which I was a guest yesterday: “I remember 30 years ago in a human physiology class, it seemed like there was a good understanding then of sleep cycles and how harmful it can be to mess them up. I wonder why the medical profession—the one that should understand this the best—seems to be the one that kind of abuses this the most?”

It’s an especially timely question, because right now things stand to get only more extreme for medical residents. The organization that makes the rules for medical trainees—the Accreditation Council for Graduate Medical Education (ACGME)—is proposing increasing the current number of consecutive hours that young doctors can work, from 16 hours to 28 hours.

When I was a medical intern (the first year after graduating medical school) in 2009, the limit for people in my position was 30 continuous, sleepless, busy hours. The Institute of Medicine had issued a report the year prior saying that was unsafe. At the request of Congress, the physician body had audited the ACGME rules and said that the limit for shifts should be 16 hours. (Or 30 hours with a “5-hour protected sleep period” in the middle. Which sounds meager, but there were times I would have sold my soul for even 20 minutes of sleep.)

In 2010, the ACGME changed its rules accordingly—for first-year residents, at least. Hearing that, I thought I’d be the last class to have his first-ever hospital shift be 30 hours in the ICU. That mix of panic, inadequacy, and exhaustion that I wish on no person—for the new class, that would simply be a mix of panic and inadequacy. The exhaustion from 16-hour days would be more chronic than acute.

But now the ACGME is proposing raising that limit back up to 28 hours. The group is currently accepting public comments on proposed revision, until December 19. After that, the task force will use the comments to inform final recommendations.

I asked the ACGME why this is happening. The group’s spokesperson said no one was available to talk to me for a few days, but they were happy to answer my questions in written form, which they did (and that’s why the quotes here sound stilted).

The group said the 28-hour maximum is “based on new evidence, research and expert input.” At a national meeting in March of 2016, the ACGME  heard perspectives from specialty societies, certifying boards, patient-safety organizations, resident unions, and medical student organizations. Among the new evidence since 2011, the most influential study was a large survey of surgical outcomes. Published earlier this year The New England Journal of Medicine, it found that for surgical residents, “Less-restrictive duty-hour policies were not associated with an increased rate of death or serious complications.”

The study began in the fall of 2014, when Northwestern researchers compared programs that allowed residents to work longer shifts. They also didn’t have to be given eight hours off between shifts, or 14 hours off after a 24-hour shift. In these hospitals, rates of death and surgical complication were comparable. So the authors concluded that “flexible duty-hour policies for surgical residents were non-inferior to current ACGME duty-hour policies with respect to patient outcomes.”

Of course, non-inferior does not mean superior. The study did not compare the actual hours worked by residents, only the guiding policies; and it didn’t assess the effects of exhausted residents on nurses and other clinical colleagues, who may have served as safeguards against error. The trial also didn’t test the 16-hour versus 28-hour maximum. Another trial is doing that currently—called iCOMPARE, it is a large collaboration between the University of Pennsylvania, Johns Hopkins University, and Harvard Medical School. But those results are not yet known.

Nonetheless, the ACGME has decided to propose repealing the 16-hour cap for first-year residents. As the group explained it to me, the cap “may not have had an incremental benefit in patient safety, and there might be significant negative impacts to the quality of physician education and professional development.” Letting first-year doctors work 28 hours “puts first-year residents on the same schedule with other residents, and is a commitment to team-based care and seamless continuity of care that promotes professionalism, empathy, and commitment among new physicians.”

In other words, that’s the culture. Patients and colleagues feel bad, and you will, too. That may be less absurd than it sounds. Even Schlachter agrees this cultural component is important. Part of medical education is teaching dedication. “I should be at the front of the line saying that residents shouldn’t be pushed to the point where they can't take care of themselves,” Schlachter told me, “or when their safety is endangered.”

He injured his hand several years ago and had to give up neurosurgery, so he went to law school and now works as a medical malpractice attorney. In that capacity, he is profoundly critical of hospital culture. But ultimately he’s on the fence about the work-hour restrictions: “When I came up, we worked 120 hours a week as residents. We were committed—it was like Marine boot camp. But that kind of training follows through into your care of patients. Things have evolved now to the point where doctors are shift workers. They don't care as much. They don't feel that responsibility as much.”

Is so many hours in an inpatient setting really what takes to teach dedication, responsibility, and commitment? Especially when most American illness is chronic, and the most cost-effective and underutilized solutions are preventive? Are other professions—where people sleep at night, every night—failing to instill dedication?

And what about the health and safety of young physicians? The evidence that sleep deprivation is a serious health hazard is mounting daily. For just one example, a study in Science that haunts me is one suggesting a function of sleep is to flush metabolic byproducts and toxins from the brain—including the beta-amyloid plaques that accumulate in Alzheimer’s disease. Sleep-deprived people are at higher risk of diabetes, obesity, depression, and cardiovascular disease.

I put the question of resident health concerns to the ACGME directly. They answered less so: “The ACGME is committed to addressing physician well-being and recognizes that many factors contribute to well-being, beyond hours worked.” (Though what we’re talking about is hours worked.) The group went on to detail ways that residents will have support if they are feeling exhausted or burned out, like provisions for transitioning the care of patients to other doctors when a resident is fatigued or ill; and the requirement that hospitals “must provide adequate sleep facilities and safe transportation options for residents who may be fatigued.”

It’s delicate language—suggesting that a person may occasionally be fatigued after running around a hospital for 28 hours. And the problem for me was almost never that there wasn’t a bed, but that if I had chosen to use it, patients would’ve gone neglected. If I said I was too tired, one of my already beleaguered colleagues would bear that burden.

Sweeping changes to this complex system are clearly impractical; inpatient hospital work is a tapestry of personnel dynamics, patients in need 24-7, multidisciplinary teams to be coordinated and bottom lines to be met. Doctors today see ever more patients in ever shorter visits and spend ever more time on paperwork.

In that light, the less discussed factor in work-hour debates is that residents are a cheap source of labor for hospitals, as compared to senior doctors. Over the years, representatives from the ACGME and Association of American Medical Colleges have been emphatic that hospitals do not profit from the labor of residents. This has been the long-accepted idea, though it has not born out in independent analyses or basic economic arguments.

Even though residents are licensed M.D.’s often working 80-hour weeks—often on the least desirable tasks at the least desirable hours—resident physicians make $50,000 to $65,000. On a per-hour basis, that breaks down to less than most ancillary staff at the hospital. Immediately upon completing the residency program, though, the same doctors command a salary of four, six, or eight times as much.

Resident labor is made cheaper because salaries are in most cases paid by the federal government, drawn from Medicare and Medicaid. The cost to taxpayers is around $5 billion, though profit from residents’ work done goes to the hospital. The money given to hospitals actually exceed the residents’ salaries by as much as $100,000 per resident. The rest goes to the hospital, officially  to cover administrative costs of running a residency program—a staff-administrator to oversee the program, often a daily lunchtime lecture, malpractice insurance, and some time allotted for senior physicians to see patients alongside the residents.

In either case, the workflow in many hospitals would crumble if residents instantly started working 10-hour days and rarely overnight. A 16-hour maximum, though, represented incremental movement toward a change in culture. The medical profession is rife with stubborn adherence to tradition, but it is an especially dramatic failure of imagination to think that a well-slept physician workforce is simply precluded by the nature of the work. It is clearly true that shift changes are a source of miscommunication and error, but it’s deeply unimaginative to consider that the solution is to keep people working beyond the point that neurobiology tells us our systems can function well, even adequately.