"A state of institutional denialism"

Over a quarter century ago, a young woman was admitted to a New York hospital with fever and agitation. She never walked out. Libby Zion died while under the care of he primary care doctor and two medical residents. The exact cause of death was never identified, but the case led to a forced examination of medical residents' work hours. This was driven largely by Zion's father who felt that his daughter had been killed by inexperienced, poorly supervised, and overworked resident physicians.

"You don't need kindergarten," he wrote in a New York Times op-ed piece, "to know that a resident working a 36-hour shift is in no condition to make any kind of judgment call -- forget about life-and-death."

It was largely thanks to Zion's tireless work that in 1989 a bill was passed in New York State limiting resident work hours and requiring senior physicians to be physically present in the hospital. But though you might not need kindergarten to recognize this problem, you do need data. That came later.

Medical residents have traditionally worked long hours, especially in their first ("intern") year. In fact, they used to "reside" in the hospital, and were universally young, male, and single. Now, graduating medical students are about 48% female, compared to just over 26% in 1982 (although age hasn't changed much, which sort of surprised me).   The Libby Zion law limited resident work hours to 80 hours per week and 24 hour shifts.  During my internship in Chicago, we would typically work about 32 hours in a row on call and post-call, and call took place every fourth night,  which has long been typical for internal medicine residencies.  

In 2003, the Accreditation Council on Graduate Medical Education (ACGME) instituted the first national work hour limitations for residents. These limitations looked very similar to those imposed by NY state. These work hour limitations required significant changes to how hospitals and residencies were run.  Hospitals can only support a certain number of residents, and they count on these residents and the care they provide.  Hospitals have had to reduce the number of patients cared for by residents, and has led to an increase in so-called mid-level providers (physician assistants and nurse practitioners).  And residencies had to find ways to accomplish the same or similar amount of work with the same personnel but with significant time constraints.  

Many of these changes involved a more toward "shift work" and night float systems, where residents worked shifts of limited hours throughout a 24 hour day, handing off patients to the next shift.  This creates its own problems for both patients and residents.  There are concerns that shift work may lead to a disruption in continuity of care, since patients are being "handed off" potentially several times a day.  Also, residents are not supposed to be performing functions that are primarily "service" rather than educational.  During the day, residents can break away from clinical duties for educational conferences, but a 11pm-7am shift is all service.

These, and the urgent questions about the safety of both patients and residents were addressed in a comprehensive report released in 2009 by the Institute of Medicine, part of the National Academies.   While it makes sense that long sleep-free work hours might be dangerous to both patients and residents, knowing the data allows us to make proper, evidence-based decisions about these potential problems.

Resident Safety

As medical educators, we have a duty to our residents to ensure not only their education, but their well-being, at least as it relates to work. It is conceivable that long, sleepless work hours may have adverse health effects.  The 2009 IOM report summarizes some of the evidence for fatigue-related injury.  Much of this evidence is readily available through PubMed.  Needle stick injuries, for example, are a relatively common problem and there is evidence that these are related to fatigue.  There is also good evidence that medical residents have an elevated risk for falling asleep at traffic lights and being involved in motor vehicle accidents.  And these data are not new.

Patient Safety

Data on patient safety isn't new either.  A name that pops up again and again in this research is Charles A. Czeisler. He published a study in the New England Journal of Medicine in 2004 showing fairly convincingly that first-year residents in the ICU are at risk of committing significantly more medical errors when working extended shift vs. less onerous ones.  That's just one good study of many.

Individual errors are inevitable, but as a phenomenon, errors can be reduced significantly, often through simple systems fixes.  One of these fixes is the implementation of reasonable resident work hours.


Responses in the literature and in doctors' lounges have been tangential and almost intentionally obtuse. A colleague of mine at another institution has opined that the medical profession is in a state of "institutional denialism" about the effect of long hours on safety and performance.  I don't think that is unfair.  The evidence on this has existed for years, yet we've made only cosmetic adjustments to our training programs.  Even the latest ACGME rules (which take effect in July 2011) fail to address the most significant implications of the problem.  The work hour limitations they mandate will very likely help, but there is a larger systemic problem.  Medical training is lengthy and expensive.  If we're going to cut back on hours, we need to re-evaluate whether the new hours are sufficient to meet educational needs.  If not, we are going to have to find a way to fund longer training programs and to fund the debt-ridden trainees who will spend extra years not paying their educational debt.    Quick fixes, even smart ones, aren't going to do the trick.  

The Libby Zion case that eventually led to the new work rules was over a quarter century ago.  How long will it take us to create real, comprehensive solutions?


Fisman, D., Harris, A., Rubin, M., Sorock, G., & Mittleman, M. (2007). Fatigue Increases the Risk of Injury From Sharp Devices in Medical Trainees: Results From a CaseâCrossover Studyâ¢â Infection Control and Hospital Epidemiology, 28 (1), 10-17 DOI: 10.1086/510569 

Trinkoff, A., Le, R., GeigerâBrown, J., & Lipscomb, J. (2007). Work Schedule, Needle Use, and Needlestick Injuries Among Registered Nurses â¢â Infection Control and Hospital Epidemiology, 28 (2), 156-164 DOI: 10.1086/510785

Steele, M., Ma, O., Watson, W., Thomas, H., & Muelleman, R. (1999). The Occupational Risk of Motor Vehicle Collisions for Emergency Medicine Residents Academic Emergency Medicine, 6 (10), 1050-1053 DOI: 10.1111/j.1553-2712.1999.tb01191.x

Barger LK, Cade BE, Ayas NT, Cronin JW, Rosner B, Speizer FE, Czeisler CA, & Harvard Work Hours, Health, and Safety Group (2005). Extended work shifts and the risk of motor vehicle crashes among interns. The New England journal of medicine, 352 (2), 125-34 PMID: 15647575

Landrigan CP, Rothschild JM, Cronin JW, Kaushal R, Burdick E, Katz JT, Lilly CM, Stone PH, Lockley SW, Bates DW, & Czeisler CA (2004). Effect of reducing interns' work hours on serious medical errors in intensive care units. The New England journal of medicine, 351 (18), 1838-48 PMID: 15509817 

Gaba DM, & Howard SK (2002). Patient safety: fatigue among clinicians and the safety of patients. The New England journal of medicine, 347 (16), 1249-55 PMID: 12393823

Nuckols TK, Bhattacharya J, Wolman DM, Ulmer C, & Escarce JJ (2009). Cost implications of reduced work hours and workloads for resident physicians. The New England journal of medicine, 360 (21), 2202-15 PMID: 19458365


More like this

Part of the problem is that we need to start from scratch, not tweak what we have done before. Programs need to identify the crucial components for education (for which we have no data) and figure out how to make those happen within the constraints of work hour limitations. We probably also need to shift from our inpatient training setting, but that's another story.
Of course, no one seems to care about the attending physician who works 30+ hour shifts, or the effect of being awakened every 1-2 hours overnight on performance the next day. The powers that be feel that experience will trump fatigue. It won't always.

I doubt the educational benefit of long hours, so I don't see cutting back on resident zombie time will affect their education. And learning how to do a good hand-over after 12 h might be more beneficial than getting to see a new patient for the first 24 h but missing half the important stuff doing a hand-over after a 32 h shift.

The powers that be feel that experience will trump fatigue. It won't always.

The military always prided itself on how soldiers could train themselves to get by on minimal sleep. Sleep deprivation in the military is one of those tropes going back time out of mind -- consider the stories of running the troops into the ground until they can fall asleep standing in ranks.

Then they actually, like, studied it and found out that the first thing to go is the ability to realize how badly you're doing. "I only need a couple of hours sleep before I fly" is about the same as "I drive better after a couple of drinks." Now they have strict rules on getting the troops adequate sleep.

Funny how we as a society care more when it's an airplane at stake than a patient.

I doubt the educational benefit of long hours, so I don't see cutting back on resident zombie time will affect their education.

One of the first things to go with sleep deprivation is the ability to learn. We've known this for decades (although football still takes priority in secondary education.) Flip over to Bora's blog for mountains of references on the subject.

Someday someone involved in medical education may actually read the literature on how people learn. Of course, that would only be relevant if the neuropsychology research applied to med students.

By D. C. Sessions (not verified) on 29 Jun 2010 #permalink

One major issue that is implicit in this debate is money: If each resident can only work 80 hours per week where they used to work 120 (yes, that is a realistic number) then you need someone to pick up the slack. That can be more residents, attendings (who don't have limits on their working hours but who are both more expensive and more likely to quit if asked to work 120 hours per week), more PAs and APNs, or even, to some extent, more RNs. But in the end someone, who will have to be paid, is going to have to do the work. Patients can not be left without care just because the clock ran out for one particular resident.

So, how much more are you willing to pay in health insurance and/or hospital bills to get more alert doctors?

So, how much more are you willing to pay in health insurance and/or hospital bills to get more alert doctors?

That's a lottery-ticket question. People are really, really bad at guesstimating risk of rare events.

It's also somewhat misleading, in that the cost of MDs in hospital care is pretty low on the list when you sort out where money goes. Increasing the number of residents by a factor of 2 isn't going to anywhere near double the per-capita spending on health care.

Finally, in (partial) balance to my admittedly harsh views of the traditional system, I'll point out that there are known ways of reducing the consequences of errors in any system [1]. Checklists and review are pretty high on the list, and although they have some costs of their own they're not remotely as expensive as failures.

For instance, nurses may not be qualified to make medication decisions but (at least where the ex worked) they were very strongly expected to double-check meds orders and on memorable occasions caught potentially catastrophic errors. The same can be said for a great many other situations.

There's a world of difference between a system which demands that there be no errors in the first place and one which accepts that errors are inevitable and puts controls in place to keep them from resulting in failures. What we have now, to a frightening degree, is a system which rejects the idea of guarding against inevitable errors while promoting conditions which practically guarantee them.

[1] I distinguish between "errors," where bad information is injected into a process, and "failures," where uncorrected bad information leads to irreversible adverse consequences.

By D. C. Sessions (not verified) on 29 Jun 2010 #permalink

Increasing the number of residents by a factor of 2 isn't going to anywhere near double the per-capita spending on health care.

I agree, but it's not going to be free and given the current hysteria about anything that might increase the cost of health care I'm surprised no one noticed the little problem that if people work less and the same amount of work needs to be done you need more people to do it. Also, more residents=need for more attendings to teach the residents, arguably a longer residency (which I kind of think should be done anyway: it's ridiculous to try to learn all of internal medicine in 3 years), more medical students to become residents some day (and all of them require attendings to teach them, etc), and so on. In theory, this could also lead to an excess of doctors, but I've yet to see that look like a problem.

Also, more residents=need for more attendings to teach the residents, arguably a longer residency (which I kind of think should be done anyway: it's ridiculous to try to learn all of internal medicine in 3 years), more medical students to become residents some day (and all of them require attendings to teach them, etc), and so on. In theory, this could also lead to an excess of doctors, but I've yet to see that look like a problem.

This sounds like an apprenticeship program under stress. I wonder what it will take to force a rethink that goes to the basics of learning?

I don't know if more granularity would help or not, but other fields (including law) work on something closer to "the interns are trained by the junior residents, who are coached by the senior residents, who are ..." I get the impression that the balance between "training" and "supervision" shifts during the program, whatever that implies.

Longer residencies wouldn't have to be so daunting if the pay and working conditions weren't so ugly. One way or another it's a very different economic model.

I wonder how this tradeoff goes in the UK, since they seem to manage a high quality of care for a very low cost compared to the USA.

By D. C. Sessions (not verified) on 29 Jun 2010 #permalink

As a starting 3rd year I'm excited that the new work hours will be implemented in time for them to affect me. But I can't stop thinking about my student loans and how they will be eating up my financial well being for 10-20 odd years. I noticed that for the loan that will pay for my third year (~$65,000) will cost me three times as much to pay back because of the interest (~$110,000). I might yet me swayed into general practice as an internist but right now I can't figure out how I'll be able to afford to.

Last week I finished my last week of inpatient service as an attending. In there I had a 50+ hour stretch of being on call. One night I got about an hour of sleep (delivering babies will do that to you). The second night I got maybe 6-7 hours with one 30 minute nap stuck in there. Thankfully, my residents were all experienced and dealing with patients who were on our service for a long time, had chronic issues, and didn't require me to think too much because there was no way I was thinking or even contributing to care in a substantial function other than physically being there. Having attendings without work hour restrictions is problematic in and of itself, particularly since we are supposed to be the last line of defense, the ones who are supposed to be sure that the residents are properly supervised, and the ones with the ultimate responsibility for the patient's care. If we're sleep-deprived, we aren't doing our jobs. Yet, there's very little done to limit our hours.

The problem with debt from a medical education is huge. When I entered a private medical school in the mid-1990's, I remember a talk where the med school projected that students were going to leave med school with about $100K in debt and that they were concerned about what that much debt will do to the students. Fast forward about 10-15 years later. My sub-intern is going to the state school on in-state tuition and will graduate with about $200K in debt. That's a mortgage. She wants to go into family medicine, but already she's worried about her ability to pay back that debt. Residents are paid a pittance for how much work that they do. My residency paid very well comparatively, but still my fellow residents struggled. It's not pretty. I am extremely privileged in that all my education was paid for and I left school without debt. I know this and I am hoping I can provide the same to my children.

I believe that some of the differences between the US and European nations is that the US doesn't subsidize medical education as much and that fewer of our counterparts in other nations leave medical school with the same levels of debt. I could be wrong, but that is my impression.

Extending residency definitely will give more time to get more training and supervision and learning time. I think that's why I ended up in academic medicine, because I felt I needed more time to learn and get comfortable being a family doctor. However, residency salaries, the extra time training prolongs the time that someone has debt that is increasing. Admittedly, once someone gets out of residency, she/he can start to pay back debt... assuming they don't want to get a house. However, if one goes into primary care it will take a lot longer than if one was to specialize, particularly in one of the specialities that are heavy into procedures. If you want to do academic medicine, then your pay is even less, though you can get some nice other perks. If you want to do the National Health Service Corp or engage in any other debt forgiveness programs, then that is one way to manage the debt. however, life can throw you a curve ball along the way and make such an agreement a lot less appealing than when you first signed-up.

Which is a long way of saying, yeah, the whole system is messed up. I know how I'd consider fixing the problem (as well as fixing health care), but I don't see anyone having the necessary political will to do so because it would be too socialist. I also think that there still is an underlying sense of "Suck it up and deal. I had to do this in residency so you have to as well". Does it make sense? No. Will it ever change? Maybe. I'm not going to hold my breath.

Anyone arguing against limiting work periods for physicians can ask themselves: are you fine with getting on an airplane with your pilots doing 36-hour shifts?

It would be nice if doctors had something like lexis-nexis where they could search a database for evidence then spend hours honing it into something meaningful for one particular patient. And to do this from home in a comfy chair with a glass of wine on the table beside you... why yes! Lawyering is so much like doctoring!

I'm not saying the system of training doctors doesn't need change or even expressing an opinion on whether it should be tweaked or re-thought from the bottom up. I'm just saying that one should be very very careful in thinking that what works in one profession will work in another.