Limiting physician trainees’ work hours and unintended consequences

I’ve been meaning to write about this topic for quite a while but never really found a reason to. Indeed, this one’s been floating around in the back of my mind for a long time. Perhaps one reason is that it’s hard for a surgeon to write about this topic without coming off sounding like an old fart, a curmudgeon, unhappy about change and thinking that a system that was good enough for me must sure as hell be good enough for the current generation of residents. In fact, even after seeing an article that normally would have spurred me to write about this topic more than two weeks ago, I stored it in my folder of bloggable links and forgot about it. But sitting in front of my computer last night, I thought back to my residency and how different it was compared to residency now.

The article appeared in a couple of weeks ago and was entitled Halt the Surgery–It’s Time for My Nap: The downside of requiring young doctors to get more sleep:

Much to the delight of harried young doctors everywhere, an expert panel recently agreed that medical residents aren’t getting enough sleep. Citing evidence that fatigue leads to more medical errors, the Institute of Medicine said last week that doctors in training should not work more than 16 hours without taking a five-hour nap. Though it carries no binding authority, the recommendation of the IOM’s report supplements an earlier rule, passed by the Accreditation Council for Graduate Medical Education in 2003, that limited residents to 30-hour shifts and no more than 80 hours of work each week. Surgical residents may someday soon have to prepare themselves to halt an operation and announce that it’s nap time.

Here’s where it’s time for the old curmudgeon to reminisce about how, back in my day, we used to complain about every other night call because we’d miss half the cases, how we’d show up at 4 AM to pre-round on our patients and like it, and how we’d walk miles to work through the snow every morning uphill both ways. Back when I did my surgery residency, there were no work hour restrictions. None. Work was over when the work was done, and not before. In general, in my residency program, call averaged every third night. True, for some rotations, mainly at the VA Hospital in our program, call was every fourth night, while for other rotations, such as cardiothoracic surgery and trauma, it was every other night. It was not uncommon for me to average 100 to 110 hours a week in the hospital.

There were also rotations, mainly senior rotations, where I did not take call from in the hospital. However, that does not mean I didn’t take call. What it meant was that I was on call every night from home, 24 hours a day, 7 days a week. Rotations where this happened included Transplant and Pediatric Surgery. That might sound better than the other rotations, but in fact it wasn’t. In actuality, I tended to spend more time in the hospital during those rotations. These rotations, I now realize in retrospect, were also the most like being an attending surgeon than the rotations when I was on call on specific nights. Worst of all was one period of time when we were short of senior residents and I got to experience the worst of both worlds. I ended up not only taking general surgery call every third night but being on call for Transplant on the other nights. On more than one occasion, post call I had to go out on an organ donation run. Indeed, I remember one week in particular where I estimate that I probably didn’t get more than four or five hours of sleep total over the course of several days.

Now let me make it clear right now: I’m not romanticizing my experience. At the time I was going through that week with virtually no sleep, I hated it. At the time I was spending 100 to 110 hours a week in the hospital, I hated it. During those rotations where I was on call every other night, I hated it. I may have loved the surgery and medicine, but I hated what it took to learn surgery. There were more than one occasion when I serious–and I do mean seriously–thought of walking into my chairman’s office and quitting. There were times when I feared for my marriage, and there were more than one times when I feared for my health and sanity. At times, I became cranky and irritable (even more so than demonstrated on this blog) to the point where I occasionally behaved in ways that now embarrass me to think back upon them. But I persevered.

Before I get to the topic of work hour limitations, though, let me point out that there were a few positives. First, I learned things about myself that I had never suspected before, the most important of which was that I could endure and tolerate far more than I ever would have dreamt possible. Second, I did learn the importance of continuity of care by observing how disease progressed over time by direct, uninterrupted observation. Finally, I learned how to function under pressure and extreme fatigue. These are not bad things. But are they worth the downside? We may soon find out. Or, more specifically, we may soon find out whether the advocates of work hour restrictions are correct that, if a little work hour restriction is good, more must be better:

The American medical establishment has been slow to give up a hazing ritual that assigns grueling schedules to trainees, with supporters of the schedule arguing that the long hours prime young doctors for the rigors of medicine, expose them to many disease scenarios, and promote continuity of care for patients. Other nations have been quicker to jettison that system. New Zealand limits residents to 72 hours of work each week, while France caps the workweek at 52.5 hours. Danish residents work no more than 37 hours a week. (What a breeze!) Elsewhere in Europe, countries are slowly lowering the work hours of “junior doctors” to comply with the European Working Time Directive, which limits hours for all shift workers. By 2009, junior doctors will work no more than 48 hours a week.

Unfortunately, working less comes with a big price tag. Countries that have imposed shorter work hours for residents have faced steep staffing shortages as well as questions about the quality of their medical training.

Let’s review the reasons commonly put forward for work hour restrictions. First and foremost, it is put forward as a means of reducing medical errors, a highly worthy goal. Certainly, there is evidence that fatigue makes the likelihood of errors higher. At the risk of being too anecdotal, these tend to be errors of omission rather than commission, as in preferring sleep to doing what needs to be done. Certainly operative performance decreases as well; as have many surgical residents before me, I have come very close to falling asleep during an operation, although, I will note, only when was holding retractors or not otherwise the primary or secondary surgeon in the procedure. The prototypical case used to further this viewpoint is that of Libby Zion, although a good argument could be made that it was inadequate staffing and supervision more than prolonged work hours, where junior residents took care of her with little or no attending input. The senior resident was never wakened when Zion deteriorated, and the attending was not notified.

But do work hour restrictions actually enhance patient safety and decrease the likelihood of medical errors? For all the confidence that its advocates present, the record is actually quite mixed. At best, it’s a wash. We have retrospective analyses full of confounding factors that suggest perhaps an improvement in patient outcomes, but the changes are small at best. Certainly there has been no large scale improvement in patient safety due to a decrease in fatigue-related medical errors attributable to the 80 hour work week. That much, at least, can be said with confidence.

The reason, very likely, is the law of unintended consequences. While it is quite possible that fatigue-related medical errors may have decreased, it is also quite possible that errors related to other factors, including systemic factors, may well have increased. The reason is that decreasing resident work hours by necessity harms continuity of care. Patients have to be “handed off” to different residents, and the shorter the work hours the more frequent the hand off. Another worry is how well the next generation of physicians will be trained, particularly in procedure-intensive specialties like surgery, where quantity counts; i.e., practice makes perfect. It will become more difficult for surgeons to achieve adequate training under then new system:

Proficiency in the operating room notoriously demands long hours, and one-third of orthopedic surgical residents were deprived of training in the operating theater because of shorter work hours, according to a 2002 survey by the British Orthopedic Association. “To become a competent surgeon in one fifth of the time once needed either requires genius, intensive practice, or lower standards. We are not geniuses,” wrote the authors of an article published in the British Medical Journal in 2004. “That many senior house officers arrive at posts halfway through their rotations without any real competence in operative skills as basic as suturing and tying knots is therefore unsurprising,” they noted.

As a counterpoint, there is evidence that the 80 hour work week doesn’t significantly decrease operative experience among general surgery residents. Of course, the 80 hour work week is only a 20-25% decrease from previous work loads. Will there continue to be no decrease in caseload if hours are decreased to 56, which indications suggest will eventually be the next step? Again, we can look to our European colleagues, who are further along in this process:

As European countries approach a 2009 deadline for fully implementing a 48-hour workweek for doctors, critics have renewed their arguments. In November, a study published jointly by the Royal College of Anaesthetists and the Royal College of Surgeons suggested that medical education in the United Kingdom would need an overhaul in order to maintain certain training standards while complying with reduced-hour rules. Testifying before the U.S. Institute of Medicine’s committee on residents’ work hours, Dr. Bernard Ribeiro, former president of the Royal College of Surgeons of England and an outspoken critic of shorter work hours, urged members to consider the implications of reducing residents’ hours: British residents today perform 25 percent fewer procedures than they did before the regulations began to take effect, he said.

From my point of view, there are two compelling arguments for work hour restrictions. The first is education. The purpose of residency is education, to train the next generation of physicians. Residents should not be considered cheap labor, which they all too often were in the past. Indeed, when I was a medical student, residents and medical students put in all the IVs and drew all the blood aside from one morning blood draw a day. Although all that practice did make me very, very good at putting in IVs by the time I finished medical school (a skill that served me well as an intern, in contrast to a lot of fresh interns these days who seem unable even to put an IV), it was scut work, pure and simple. There was virtually no educational value after I achieved a certain level of proficiency and, from my perspective, using us to put in IVs was exploitation of medical students and surgery residents as cheap labor, pure and simple. Indeed, the article cites a report that shows just how much it would cost to replace this cheap labor:

In Europe, where thousands of physicians were needed to fill vacancies created after residents scaled back their hours, hiring additional personnel cost an estimated 1.75 billion Euros. Exceeding the 48-hour-a-week allotment “is the rule rather than the exception” in Portugal, noted researchers in a 2004 British Medical Journal article. The United Kingdom needed an estimated 15,000 additional doctors to staff the National Health Service to comply with the Working Time Directive, which applied to junior doctors for the first time in 2000. In 2004, the BBC reported that the NHS was facing a “staffing crisis” brought on by shorter hours for residents.


In 2003, when the Accreditation Council for Graduate Medical Education ruled that residents could work no more than 80 hours a week, hospitals were forced to hire additional nurses, technicians, and senior doctors to pick up the residents’ slack. Last week, the IOM committee said its recommendations could cost $1.7 billion a year. The committee justified the expense by saying medication errors and the cost of treating drug-related injuries in hospitals add up to more than $3.5 billion a year.

I have yet to see one whit of evidence that work hour restrictions have decreased medication errors by anywhere near that amount–or, that they have decreased medications errors measurably at all. In this area of the debate, faith all too often wins out over science. However, from an educational perspective, work hour restrictions do appear to allow residents to spend more time studying, thus resulting in improvements in the American Board of Surgery In-Training Examinations taken every year by residents.

The second argument for resident work hour restrictions is simple humanity. Thinking back on my residency, there are times when I marvel that I made it through it all. It was brutal. Even now, nearly 13 years after I finished residency, the tincture of time has not led me to look back with misplaced romanticism on just how brutal it was.

There is one thing that’s completely clear, however, and that’s that, if we keep decreasing resident work hours, the structure of residencies will have to change. They will have to become longer, particularly in residencies that demand a lot of procedural skill. Indeed, the move to decrease resident work hours is conflicting directly with another movement in medical education to document patient care experience, such as operative cases, patients admitted, patients with different conditions and diseases cared for, and the like. More importantly, the ideology–and yes, it is an ideology; the science supporting reducing work hours much further than what they’ve been reduced to is shockingly thin, particularly the claim that doing so will decrease medication and medical errors–requires flexibility. As the article points out, the “one size fits all” approach completely ignores differences between the specialties and could actually end up hindering medical education in some areas:

We all want our doctors to be well-rested, but the IOM’s effort to ease the burden on overworked residents saddles some doctors with recommendations that could hinder their education. Across-the-board guidelines lump together doctors with vastly different skills, sleep needs, and career goals. More flexibility would keep the United States from facing the doctor shortages and training deficiencies seen by other countries. By allowing individual programs to tailor work hours to meet the needs of their residents, the rules could accommodate aspiring physicians for whom shorter shifts are sufficient as well as those surgery residents who may benefit from logging extra hours in the operating room

We are now in the middle of the sixth full year of resident work hour restrictions, and thus have now had sufficient time to see the first couple of classes of surgeons who have trained solely under the new system, as well as more than that for specialties that require less training. There has been a sea change in how general surgery residencies operate, if you’ll excuse the term. Despite that, it’s surprising how little hard evidence there is to support their efficacy. Education and humanity may argue for such restrictions, not to mention a need to make medicine a more family-friendly specialty in order to make it possible for women in particular to be physicians and have children, but the long-ballyhooed and extravagantly promised reductions in medical errors predicted to result from resident work hour restrictions have yet to materialize. Color me cynical when I predict that they probably never will, because of the complexity of the systemic factors that influence rates of medical errors and, of course, to the law of unintended consequences, including increased handoffs and the takeover of these functions by less trained and less motivated workers. Very likely, though, as Kevin MD and DB have mentioned, the next unintended consequence of these work hour restrictions will be the increase in the length of residencies.

I’m wondering how physicians in training will like that. It’s already an amazingly long commitment to become a surgeon, for instance, even more so to become a surgical specialist. At the old hardest of the hard core surgical residencies at Duke, they used to joke about a “decade with Dave” (meaning Dr. David Sabiston, a giant of surgery), who would keep residents until he thought they were good surgeons. Few could tolerate training that long, but it may well become the norm for surgeons. One thing’s for sure, though. We’re about to see just how much society values providing shorter work hours for physicians, because it’s going to cost a whole lot of money to replace that labor and and to extend the number of years most physicians will need to train before they can practice.