Sleep-Deprivation: Can Residents “Learn” to Function with Less Sleep?

Imagine that you been in labor for 18 hours, and the resident who will be delivering your baby tells you that he hasn’t had any sleep for 25 hours. Would you ask for a new doctor?  On the one hand, this resident has been checking you since the beginning of labor. You like him. But, he confides, that this is his second 28 hour shift in three days. He adds that he hasn’t had 24 hours off for two weeks.
An Institute of Medicine (IOM) investigation released in December 2008 concluded that physician fatigue accounts for a growing number of medical errors: “the scientific evidence base establishes that human performance begins to deteriorate after 16 hours of wakefulness” wrote the authors of the report. They called for elimination of resident shifts exceeding 16 hours without sleep. The Accreditation Council for Graduate Medical Education (ACGME), the organization charged with setting and enforcing standards for graduate medical education, has finally responded. New rules regulating how long residents and interns can be asked to work while caring for patients went into effect last month.

In the past, the ACGME let residents work 30 hours at a stretch. But as of July 1, interns (a.k.a first year residents) will be limited to 16-hour tours of duty, following IOM’s recommendation. But, to the dismay of patient safety advocates, under the new rules second and third-year residents still will be permitted to work 28 hours at a time. Moreover, “time off” in between long shifts has been cut from ten hours to eight. This gives residents eight hours to go home, sleep, shower, grab something to eat, and get back the hospital for another 28 hour shift. At best, they may catch seven hours of sleep.

Many doctors believe that long shifts prepare students for the real world of medicine. In a recent post titled “Why work hour restrictions won’t improve patient safety,” Kevin Pho (a.k.a. Kevin M.D.), an internal medicine physician who has become one of the best health care bloggers out there seems to agree: “I’ve often said that there are no work hour restrictions in the real world, so residents used to shift work may find themselves in for a bit of a surprise when they graduate.”

I would argue that lack of work hour restrictions in the real world is part of the problem. Patient safety advocates note that when fifty-something surgeons perform elective surgery after being up all night, patients may be at risk. A meta-analysis cited by the IOM, and conducted by (ACGME)’s own Senior Vice President of Field Activities, Dr. Ingrid Philibert, concludes that after 24–30 hours of sleep deprivation, individual cognitive performance of the average person drops from the 50th to the 15th percentile of their performance when rested, and that clinical performance of physicians drops from the 50th to the 7th percentile of their performance when rested. Typically, patients have no idea that their surgeon has had so little sleep.

Over the past four decades, scientific evidence linking fatigue with accidents and errors in industries ranging from aeronautics to medicine, nuclear power and transportation has spawned regulations to help ensure public safety “with the notable exception of medicine” observe the authors of a white paper on resident hours published in June in the peer-reviewed journal, Nature and Science of Sleep.

The IOM report, and ACGME’s response has re-opened a long-standing debate: How are doctors different from airline pilots? Can they, in fact, learn to function with less sleep? Many physicians argue that learning to remain alert for 28 hours is one of the demands of a profession that expects physicians to be responsible and accountable for their patients. Training residents to work long shifts teaches them endurance, they argue, as well as how to recognize when fatigue begins to degrade their cognitive and physical capabilities. As one physician told Medscape Medical News, "If I've been on call, and done a case, and I have a ruptured aneurysm to do in the morning, I lie down in my office for 15 minutes and come back fresh.” 

On the other side of the argument ACGME’s critics point out that other developed nations, such as members of the European Union, have instituted shifts that max out at 13 hours, with 11 hours off afterward. If they can do it, why can’t we? We have a stack of research which reveals that if residents work more than 16 hours, the odds that they will hurt a patient—or themselves—spiral. Nationwide, interns show a 168% increased risk of a motor vehicle crash driving home from shifts lasting more than 24 hours. At the very least, hospitals should provide transportation. Though this raises a question—if the resident is too impaired to drive a car, why would the hospital let him treat a patient 30 minutes before getting into that car? 

Dr. Thomas Nasca, CEO of ACGME defends his group’s decision, explaining that maintaining long shifts is crucial for training residents to practice medicine. While limiting shifts for the youngest residents was important to ease them into the profession, he told NPR in an interview at the beginning of the month, older residents must be taught to "recognize and manage fatigue" they will encounter regularly in their actual clinical practice, where hours are not regulated. 

Dr. Chris Landrigan a pediatric hospitalist at Children’s Hospital, Boston who has been studying health care quality and sleep deprivation for a dozen years, rejects the argument. “There is a substantial body of medical literature regarding senior residents which shows that if their shifts are limited to 16 hours patient safety and quality of care improves,” says Landrigan, who is an associate professor at Harvard’s medical school. 

For example, in one national study more than 2,700 residents in their first postgraduate year were asked to, report on a monthly basis: 1) how much they were working;  2) how much sleep they were getting;  and 3) what medical errors they might have made. When researchers compared the months in which a residents didn’t work any extended shifts to months when the same resident worked five or more overnight shifts, the analysts found that when residents worked more extended shifts they were about 7 times as  likely to make a fatigue-related mistake that injured a patient. And 1 in 20 reported a fatigue-related mistake that led to a patient’s death.

“I was very disappointed to see ACGME discount the research,” Landrigan told me in a phone interview not long ago. “They say, ‘We don’t have strong data on 2nd and 3rd year residents.’ But we do have eight or nine years of research. ACGME cherry-picks the studies it wants to talk about.”

Perhaps most importantly, Landrigan argues, the notion that 1st year residents should be limited to 16 hours, but 2nd year residents can stay on their feet for 28 hours without impairment “doesn’t make sense physiologically. It seems to me irrational to say that on July 1 of your second year, suddenly, you are able to tolerate sleep deprivation.”

        “It’s not like exercise.  . . It’s like drinking”

There is no science showing that if you go without sleep over a period of time, you can learn to tolerate it,” he explains. “It’s not like exercise–with practice you get better. In fact, if you are sleep-deprived over a period of time, say you are getting only four or six hours per night, you become vulnerable to bouts of sleeplessness. When you have 24 hours off, you find that you just can’t sleep.” 
As for the belief that residents who work long shifts learn to “recognize” fatigue, Landrigan points out that the science reveals that the longer you are awake, the less aware you are that you are impaired. “It’s like drinking,” he says.

Moreover, instead of slowing down to preserve accuracy as reaction time increases, sleep deprivation often leads people to make hasty decisions based on inadequate information, leading to increased rates of errors.

Those who champion the old rules suggest that if residents work limited “shifts” patients are put at risk during the “hand-off” when one resident goes off-duty and another resident takes responsibility for the patient. The first resident may fail to communicate essential information. But hospitals and medical schools have been focusing on teaching the art of the hand-off. As Kevin Pho acknowledges, “Residency programs have worked hard to close that loophole. At Children’s Hospital in Boston, for instance, errors fell by 40% after the patient handoff process was standardized.” Unlike functioning without sleep, the hand-off is something that can be learned.

Still, ACGME’s Nasca argues that longer shifts are important for patients, whose care often does not fit into 16-hour episodes. For example, a difficult labor and delivery might last much longer. "You may have complications," he says. "Who is better to take care of you during the final phases of the labor than the physician who's been caring for you all the way through?"

This brings me back to the question I posed at the beginning of this post. My daughter, Emily, is having her first baby sometime in the next week or two. When I asked her what she would do if the doctor delivering her baby told her he hadn’t had any sleep for 25 hours, she replied: “I’d ask for a nurse-midwife.”

In part 2 of this post, I will explore why so many hospitals are clinging to long shifts—and discuss the fact that, at some hospitals, residents work even longer shifts than the ACGME allows. At a conference last summer, one resident explained how, at his hospital, residents regularly fill out fake time sheets. While ACGME monitors hospital compliance it does not make unannounced visits. In too many cases, residents work long hours without adequate supervision.

In its report, IOM suggests that ACGME itself needs tighter external supervision—perhaps by the Centers for Medicare and Medicaid or the Joint Commission.

Finally, in Part 2, I’ll explain what Kevin M.D. means when he says that shorter shifts “plug only one hole in a slice of [Swiss] cheese”—and why I agree.”

10 thoughts on “Sleep-Deprivation: Can Residents “Learn” to Function with Less Sleep?

  1. It’s pretty frightening to think of the sleep deprivation that some of these residents are suffering from, and how it affects their work. I think that placing limitations on how long their shifts are is beneficial.

  2. I do agree with all the objective data showing sleep deprivation leads to serious (or lethal errors). I don’t have any data to offer about this, other than one, thirty-year-long anecdote.
    I practice pediatric critical care, a specialty notorious for odd hours. I don’t know why, but it seems to be the case that the sickest kids arrive in the PICU in the middle of the night.
    I’ve worked many, many 24 hour stretches (or more) over the years. Sometimes, at least at the attending level, there’s no other way. I agree that one can’t “learn” to do this. But what I have noticed, both in myself and in colleagues who have done this for as long as I have, is that all of us have the ability to get extraordinary benefit from a nap of an hour or so. If you give me a couple of those in a 24 hour stretch I have no troubles. I think this is just an inherent, constitutional thing — not something I learned. Colleagues I’ve talked to about it feel the same way. Critical care may attract such folks because we can fall asleep in 2 minutes, pass rapidly into deep sleep, and awaken quickly.
    Residents, of course, should not be expected to do this — they’re apt to hurt somebody and they won’t learn much, either.
    But this brings up a fundamental problem that specialty boards have avoided: how long should residency be? Like many physicians my age (60), I was on call every other night in the first year and every third night for the remainder of a 3 year pediatric residency. Critical care fellowship had even more on-call than that. The simple fact is that if you cut on-call time in half, you will also severely cut resident experience. Yet the total length of residency has not changed in a half-century.
    The results are already quite apparent for any specialty that involves procedures. Today’s senior residents have just done less stuff. If, for example, your doctor is a newly graduated general surgeon and you need something mildly complicated or out of the usual done to you — not rare, just a bit unusual — your surgeon may have only done a few of whatever procedure it is. In my own field, I had done hundreds of endotracheal intubations and central venous catheters, plus dozens of full-code cardiopulmonary resuscitations, before I was let loose on the public. That is no longer the case.
    So the answer is not to abuse residents and endanger patients. But something has got to change in how we train them. We can make residencies and fellowships longer. That will cause pain and expense. We can start tracking young physicians at an earlier point so they are directed to their ultimate careers earlier, but that brings with it other problems.
    But the growing body of these sorts of data compel us to do something.

  3. And texting while driving gets the headlines.
    Do any studies indicate that adrenaline ever gets low? Seems to me it’s time to review the old POW studies as well. This looks like low-hanging fruit for journalists.

  4. I clearly recall the first time I worked a 16 hour shift (worked a double) during a snow emergency when the day nurses could not get to the hospital. I went from my 11p-7a shift and worked the 7a-3p shift. By noon my brain was completely fried, and I had a hard time concentrating on the MAR while passing meds. I was truly frightened by this; I knew I was not operating at my best, and I’d only been on for 13 hours.
    I slept from 3pm until 10pm, then worked another 11p-7a shift. It was all I could do to stay awake.
    Since then I’ve worked plenty of 16 hour shifts as a nurse, but exhaustion always sets in by around hour 14 or so, and I am not firing on all cylinders, and driving home was a nightmare.
    Perhaps the real answer to the problem of resident hours is this: make them salaried and have them work 12 hour shifts, 3 or four days a week, like nurses do and increase the length of their residencies from 3 years (which is about average I think) to 5 or six years.
    Certainly I knew a lot more as an ER nurse after working 3 years than I did at 1 year, and 6 years as opposed to 3.

  5. Chris, Accunurse
    Chris– You write: “What I have noticed, both in myself and in colleagues who have done this for as long as I have, is that all of us have the ability to get extraordinary benefit from a nap of an hour or so. If you give me a couple of those in a 24 hour stretch I have no troubles.”
    You add: “I think this is just an inherent, constitutional thing — not something I learned. Colleagues I’ve talked to about it feel the same way. Critical care may attract such folks because we can fall asleep in 2 minutes, pass rapidly into deep sleep, and awaken quickly.”
    Chris, this makes great sense. Some people are able to fall into deep sleep quicklly and benefit from a nap. (Though I’m skeptical about the doctor who claims that after 15 minutes he’s completely refreshed.)
    And I suspect that, as you suggest, a self-selecting group of physicians choose specialties that are as demanding as yours.
    But this is not always the case.
    It would be extremely helpful if in med school, mentors and teachers talked about the fact that different people have different constitutions– some have more stamina than others.
    This is something that they should tell med students to think about when picking a specialty.
    I’m afraid that a long macho tradition in medicine has taught students that, in the past, the Great Men (“the gods of medicine”) possessed extraordinary stamina (or learned it) and that those who can’t “measure up” should feel shame.
    In truth, physicians who are not well-suited to doing 14-hour surgeries can do as much (or more) good as other doctors..
    Medicine is not football or soccer. It is not a “manly sport.”
    On the other hand, Chris, as you say, your specialty does require a particular constitution. Since you are working with critically ill children, these are not “elective surgeries” that can be posponed.
    Meanwhile, some of these procedures are incredibly difficut. I can’t imagine how the lead surgeon could “hand off” the patient to another surgeon 10 hours into a 14 hour procedure.
    Your idea of lenghtening the residency so that young doctors can get more
    experience (rather than having them work long hours when memory and the
    abililty to consolidate knowledge is impaired) seems, to me, spot-on. (I would love it if you would write a post about this.)
    But I don’t think we can ask residents to work on a resident’s salary for an extra two years. I would suggest that in those extra two years, they should be viewed as “apprentices” to a senior surgeon, working closely with that surgeon who will treat them as a colleague (albeit a junior colleague.)
    The senior surgeon should be present at all of the potentially complicated surgeries where they work together– with the senior surgeon serving as the lead surgeon, teaching the apprentice– or observing the apprentice, ready to jump in if problems develop, or something unusual happens.
    Today, the experienced surgeon often is not present when the resident operates, even in potentially complicated cases. The patient thinks that “his” surgeon operated on him–when, in fact, he wasn’t in the room.
    Over time, we might want to shift some of the income that we now pay to the best-paid surgeons (say, 10% of that money) and use it to lift pay for their apprentices.
    For senior surgeons there would be a tradeoff that could offset the financial loss: apprentices who had two extra years could take some of the pressure off senior surgeons. During a long surgery, a senior surgeon could step back, and let the apprentice take over. He would need to remain present–and alert–but he could take a break from the stress of being lead surgeon.
    We also should remember that in some fields we do far too many unnecessary elective surgeries. If we used comparative efffectiveness research to careully eliminate many of the angioplasties, back surgeries and orthopedic surgeries that provide little or no benefit to the patient (while exposing those patients to all of the risks of surgery), both Medicare –and the system as a whole–
    would be able to use that money to pay the apprentice surgeons the income they need and deserve.
    Those salaries would make it possible for them to start a family, buy a home, and begin paying down med school debts.
    I don’t think these longer residencies would be necessay in all or even most specialties.
    But in some high-risk, extremely complicated areas such as yours (as well as organ transplants;
    deliveries when the mother and/or baby are at risk; and probably neurosurgery . . . ) your ideas open a important door . . .
    Accunurse– Yes, this is frightening.
    We need a better system– as Chris suggests. See what you think about his proposal . . .

  6. If shorter shifts were to be traded off for a longer residency period, it seems essential that the doctor in training should be paid considerably more for the additional training time as it shortens the length of his post training career. It doesn’t matter if you call him an apprentice or a senior resident. To make the economics work for the hospital, though, it needs to be able to bill when senior residents actually treat patients.
    I also wonder with all the technological advances in recent years, there should be more opportunity to use simulators for various aspects of training surgeons. Such training could encompass more surgical circumstances including complications and the training could take place during scheduled and predictable time periods. While the young surgeon also needs to get exposure to real patients as well, the total time in residency could probably be made more productive and efficient from a training standpoint than it is now. Maybe some of the doctors could comment on the value and efficacy of simulator training and to what extent it’s used today.

  7. John B., Barry
    John B– an vivid analogy, and of course you’re right. Thanks very much.
    Barry–Research already shows that “Laparoscopic surgical simulator studies have revealed that, on average, error rates doubled in residents when they performed simulated surgery after a night on call as compared to when they were more well rested”
    Simulator exercises can
    definitely help reveal the problems–and no dobut are an excellent tool to help residents train without working for 28 hours at a time.
    But while this training helps, my guess is that it is not a total substitute for practice performing (or assisting in performing)
    surgeries on flesh & blood
    The level of stress must be quite different.
    I agree that residents who spend an extra 2 years in residency would have to be paid significantly more.
    That said, I suspect that doctors like Chris Johnson, who were drawn to a difficult specialty such as intensive care for critically ill children, would not have to be paid $300,000 or $400,000 a year for those extra years.
    They, like transplant surgeons, are drawn to these specialties because the specialty is so challenging and potentially fulfilling. They realize of the huge amount of good that they might be able to do– even though, at times, they will suffer devastating disappointments.
    Given the risks of what they do, they would probably be grateful for the extra time to train.
    Moreover shorter shifts would allow them to have more normal lives–with less stress–during their early years of training.

  8. Panacea–
    Yes. Thank you.
    I would add that nurses, like doctors should not be working such long shifts.
    It’s interesting that, normally, nurses do not work these “mega-shifts” of more than 24 hours.
    Even in an emergency (a snow storm) you were not working 30 hours…
    Traditionally,doctors have been men and nurses were women.The long shifts for residents really does seem tied up with macho ideas about “hazing” doctors and forcing them to prove that they are “real men”. . .
    In return they receive the
    “real money.” But is it worth it, for anyone— patients, doctors or doctors’ families and doctors’ relationships with their families?
    If people are worried about
    then one would think that they would be at least as worried about “hand-offs” among nurses who are much more intimately involved in the day to day of patient care. (Typically, good nurses know more about the patient than the hospitalist or other caregivers simply because they spend more time with them.)
    My daughter said she would ask for a nurse-midwife if her physician had been working for 24 hours in part because she assumes that it’s very unlikely that nurse-midwives work 25 hour shifts delivering babies.
    In the practice that she has chosen for her pregnancy and delivery, there are 5 nurse midwives, plus Ob-Gyns. She would be perfectly happy to have one of the nurse-midwives “hand her off” to another NM. (She has met each of them several times, and likes 4 out of the 5 very much.
    The 5th is pefectly fine, just not as warm, but seems perfectly competent.)
    At the same time,if Emily runs into major problems, she knows that there is
    an Ob-Gyn from her practice group present at the hospital at all times, 24/7.
    When she first visited the practice, the Ob-Gyn gave her a choice between choosing a doctor or a nurse-midwife. She chose a nurse-midwife because she thought a nurse-midwife would have more time to answer her questions during pregnancy–and more time to spend with her during labor. There is no reason to think it will be a complicated labor, and she will also have a duala.
    But she likes knowing that if, for some reason, she absolutely needs a C-section (her last choice), there will be an Ob-Gyn from a group she trusts, on the site, ready to perform the operation.
    This is a hospital that reports fewer C-sections and inductions, as well as fewer complications during deivery, than any other hospital in NYC.
    It also delives more babies than any other hospital in New York State, and nurse-midwives deliver a large percentage of those babies..
    As a result of going to birthing classes, Emiy has discovered that a great many expectant women and their partners don’t feel the way she does about the people who will be delivering her baby. .
    Very likely, some of these expectant parents are simply very anxious. But I think Emily was right to choose nurse-midwives who
    always call her back to answer questions . . ,
    Their time is less expensive, and they are trained to spend more time listening to and talknig to individual patients.
    Ob-Gyns should be spending their time on the patients who may face complications.
    It would be much better if they could spend 30 minutes (or more) in appointments with these patients while the nurse-midwives who work with them spend 15 minutes in an appointment with Emily –and,at the end of the day, have 5 minutes to return a phone call once in every two weeks in between her twice-monthly appointments when she remembers a question that she forgot to ask.

  9. A couple other thoughts.
    Maggie, you mentioned handoffs. This is an issue bound up with that of shorter shifts for residents, because shorter shifts mean more handoffs of care. It’s well known that such frequent handoffs carry risks, too. There is more opportunity for miscommunication, for important things to get lost. Plus, residents are far less likely to take ownership of the case; they don’t know the patient as well, and they are less invested in advancing the care. They are more likely to just try to get the patient through their shift without rocking the boat. Since it’s often not “their” patient, they don’t dive into the books and medical articles to ferret out what’s happening to the patient. Teams are good, but I’ve seen more than a few situations in which nobody on the team shows much initiative. And patients, although they appreciate the need for well-coordinated team care, also want a doctor to be their doctor, not some amorphous team.
    This issue is also related to the ongoing changes in hospital care in institutions without residents. At least in pediatrics, the American Board of Pediatrics has mandated more and more required outpatient rotations — things that need to be crammed into the same residency period. So residents have increasingly less experience with complicated hospitalized patients. This is one of the engines driving the hospitalist movement — physicians who only practice in the hospital, but are not one of the traditional hospital-based subspecialists. Soon that will be a subspecialty in its own right. On balance, I think hospitalists improve patient care, and I believe studies have shown this.
    But something is lost when a newly minted pediatrician (or internist) is less comfortable caring for complicated, sicker patients. I have definitely seen this happening; these days young, bright, well-trained pediatricians want my advice, or even want their patient transferred to the PICU, because they are justifiably nervous about managing a case of something they’ve only seen a few times during their residency.
    Some are of the opinion that pediatric (or internal medicine) residents should be tracked early in their training toward hospital-based or office-based practice, that it’s unrealistic to become competent in both. But if we do that, what happens to those who want to practice in more remote areas, places without the luxury of hospitalists? Those physicians need to be more self-sufficent. And they are. I deal with pediatricians and family practitioners in places like that who, when they send me a critically ill child, have done all the right things stabilizing the child because they must — there’s no one else. In contrast, young suburban pediatricians and family practitioners, when faced with similar situations, are often clueless about what to do; they’ve not been trained to do it, and never have.

  10. It is really scary to think of the sleep deprivation that some of us are suffering. Look at how it affects our mood for work, play and family. I do agree that there should be such limitations on this long hour shifts.