The Institute of Medicine just released its long-awaited report on trainee duty hours. It is well researched and balanced, and its recommendations appropriately reflect what we know vs. what we believe. Now the fun begins.
Let’s start with a little background, some of it drawn from my book Understanding Patient Safety:
Let’s be honest. Traditional resident schedules – on call every third night, staying up for 48 hours in a row, and working 120 hours per week – were both inhuman and immoral. The Days of the Giants view that such training was needed to “turn boys into men” (before women became the majority of medical students) was machismo garbage. This was a hazing ritual formed when people believed that one should sacrifice one’s life on the Altar of Medicine, perpetuated because all of our egos are such that we said, “Well, that was brutal, but just look how great I turned out – so that must have been a good system!” And, because housestaff labor is easily the cheapest in the building (what intern hasn’t done this math – my own 1983 internship salary of $17,600 translated into about $4.50/hour, less than I made caddying), what began as a rite of passage quickly morphed into an economic imperative. Having fallen asleep at the wheel once or twice driving home during my internship, I have little sympathy for those who wistfully long for the Days of Yore.
Beginning with the famous Libby Zion case at New York Hospital in 1984, the public and media have pressured “the system” to fix the problem of long trainee hours. A 1989 New York State regulation limiting duty hours to 80 per week was largely ignored, and no other state followed suit for over a decade. But the overarching pressure to improve patient safety, which began with the IOM’s 1999 report, To Err is Human, was enough to give the Accreditation Council for Graduate Medical Education (ACGME) the courage to gore this particular sacred cow, and to withstand the subsequent mooing.
In 2003, the ACGME, which accredits the nation’s 7,800 training programs, decreed that residents’ hours would be limited to 80 a week, with no shifts longer than 30 hours. Both numbers were completely arbitrary – there is no research that helps tell us the “right” number of hours per week or per shift. In fact, the research on sleep deprivation as it pertains to resident performance is surprisingly mixed. While it is well appreciated that 24 hours of sustained wakefulness results in performance equivalent to that of a person with a blood alcohol level of 0.1% – legally drunk in every state – studies have shown that tired radiology residents made no more mistakes reading x-rays than well-rested ones, and sleepy ER residents performed physical examinations and recorded patient histories with equal reliability in both tired and rested conditions.
That said, most folks find this to be one of those issues in which common sense trumps evidence-based medicine – pointing to the tongue-in-cheek BMJ piece challenging EBM zealots to participate in a randomized trial of jumping out of an airplane with and without parachutes (since the value of parachutes has never been subjected to evidence-based scrutiny). On this one, I agree: given the substantial evidence of the harms of sleep deprivation, the burden of proof should be on those defending the old schedules, not on those proposing more humane variations.
Several studies have examined the impact of the 2003 ACGME regs. It is fair to say that the jury remains out. The studies generally show no real effect on clinical outcomes or patient safety, and significant concerns have been voiced by both faculty and residents regarding unintended consequences. But the pressure to do more from a wary public remains, and there have been studies that have convincingly demonstrated that shorter shifts in the ICU environment lead to fewer errors.
When the ACGME regulations first came out, programs did what they always do with regulations they don’t like – they tried to skirt them. The ACGME did something clever in response – it fired two shots over the academic bow, placing two of the most prestigious programs in the country (Yale Surgery and Hopkins Medicine) on probation. The message was clear: we’re not screwing around. That said, this week’s IOM report was critical of what it deemed lax enforcement of the existing standards, calling for unannounced surveys, periodic audits, and stronger protections for whistleblowers. I think they were right to do so.
Programs responded to the 2003 duty hours regulations in a number of ways. When the rules hit, I was virtually certain that our residency at UCSF would go to a Night Float-on-Steroids system, sending the on-call team home at 10pm, having the nights covered by a fresh crew, and handing those patients back to a new team in the morning. But that’s not how it turned out.
One of the great things about UCSF is that our residents rotate through three separate hospitals, so we tried three different strategies to see what worked best. And the Night Float/Send The Primary Team Home idea proved to be a disaster – we couldn’t get housestaff to leave the hospital soon after admitting a desperately ill patient (that damn professionalism), so they were getting home in the wee hours of the morning, leaving them well over the hours limits and exhausted the next afternoon.
Surprisingly, the favored system was a robust Day Float system. In it, our teams continue to stay overnight, admitting all patients till about 2 am, after which a night float takes new non-ICU admissions. When I arrive for attending rounds in the morning, my team is there along with a freshly scrubbed day float resident. We hear about all the patients together, and then team rushes for the doors, the goal being to be out by noon. The day float resident and the attending then spend the post-call afternoon finishing up the plan, notes, etc. It works pretty well.
With that background, let’s turn to this week’s IOM report. Although there was considerable trepidation that the IOM would recommend severe additional limits in duty hours (most other industrialized countries limit resident hours to 50-60 per week), the report recommends relatively mild modifications to the existing regulations (they’re summarized here). The biggest one is a requirement for a minimum sleep period of 5 hours in any 24-hour work period, with a maximum shift length of 16 hours. If we keep the scaffolding of our present UCSF system, this will mandate that the on-call team takes no new admissions and doesn’t cross-cover its own patients overnight; instead they’ll have to have a complete handoff and a beeper-less interlude from about 2am-7am. That seems pretty do-able, especially considering the fact that we were girding for much more radical restrictions on hours.
What may prove to be a bigger deal is the new requirement that housestaff have “immediate access to an in-house supervising physician” – which I interpret to mean 24-hour in-house attending coverage, most likely by hospitalists. Although we have some moonlighters in the house overnight, we don’t yet have faculty hospitalists. But the tea leaves are clear: it is time to start planning for around-the-clock hospitalist coverage at teaching hospitals.
Efforts to cut duty hours raise a number of questions and concerns, which I’ll separate into five buckets: 1) handoffs, 2) costs, 3) do people really sleep when they’re off?, 4) practice makes perfect, and 5) the culture of medicine. Let’s tackle them each briefly.
First, handoffs. Until 2003, our handoffs were haphazard, on the fly, and completely unsystematized. Early on, we recognized that the 80-hour workweek was markedly increasing the number of handoffs – our own Arpana Vidyarthi found that resident handoffs increased by 40% after the 2003 regulations. Like so many other aspects of the safety field, we essentially had a squeezing balloon phenomenon: one fix (better rested residents) was traded for a new safety hazard (more handoffs). In my own judgment, patient safety worsened in the first couple of years after the 2003 rules because the handoff hazards trumped the advantages of rested trainees. It was only after we developed standardized sign-out systems that the balance became more favorable, and the new IOM report calls for even more attention to such systems. That said, there are few days when I don’t hear our nurses complain about paging the resident and hearing, “I really don’t know that patient very well. I’m just covering.” (That’s assuming that they can figure out which resident is covering at that particular moment, an immense challenge unto itself.)
The second issue is cost. The new IOM report estimates that the cost of implementing the new standards will be $1.7 billion nationally – including the hiring of about 6000 mid-level providers (NPs, PAs) and 5000 hospitalists. I don’t doubt it: the 2003 regs were the Hospitalist Full Employment Act. At UCSF, while early efforts to deal with duty hour reductions focused on residents covering for themselves coming off non-call electives (didn’t work and was wildly unpopular), they soon shifted to using NPs and PAs (worked sometimes, but some patients were simply too complex and some providers were too expensive and inefficient) and ultimately to using hospitalists. Of our 42 faculty hospitalists, I’d estimate that about 12 FTEs are here because of the need to replace resident bandwidth on a variety of services. The new restrictions are likely to increase the need for additional coverage, and thus the costs. The reason that the IOM blinked when it came to cutting the hours down to 60 must have been partly due to these cost considerations, especially in an era in which many teaching hospitals are struggling to break even.
The third concern is whether housestaff really sleep when they’re off. Remember, these are young people with significant others, hobbies, laundry, and debts. Not surprisingly, there is some evidence that they don’t use the time out of the hospital to sleep, and the IOM weighed this in choosing to keep the weekly hours at 80. As John Iglehart observes in his excellent editorial in this week’s NEJM, “Although some might propose further reductions in total duty hours, the report notes, ‘evidence suggests it is an indirect and inefficient approach given the moderate correlation that exists between resident duty hours and sleep time.’”
The fourth is Practice Makes Perfect. Particularly in surgery and other procedural specialties, there is real concern that trainees may not be handling enough cases to become fully competent. There are few data to support this concern, and one has to believe that some of the work that residents put in during hours 80-110 in the old days were not highly educational (not to mention safe). But I’ve met many surgical program directors who are quite convinced that their graduating trainees are not prepared to operate independently – both because trainees are doing fewer cases and because of the enhanced supervision that is chipping away at the trainee autonomy necessary to develop clinical instincts and judgment.
Which brings us to the final concern (and my greatest worry): the culture of training. When the 2003 ACGME regulations came out, New England Journal editor Jeff Drazen and Harvard policy maven Arnie Epstein wrote that that traditional residency schedules,
. . . have come with a cost, but they have allowed trainees to learn how the disease process modifies patients’ lives and how they cope with illness. Long hours have also taught a central professional lesson about personal responsibility to one’s patients, above and beyond work schedules and personal plans. Whether this method arose by design or was the fortuitous byproduct of an arduous training program designed primarily for economic reasons is not the point. Limits on hours on call will disrupt one of the ways we’ve taught young physicians these critical values . . . We risk exchanging our sleep-deprived healers for a cadre of wide-awake technicians.
Therein lies the tension: legitimate concerns that medical professionalism might be degraded by “shift work” and that excellence requires lots of practice and the ability to follow many patients from clinical presentation through work-up to denouement, balanced against concerns about the effects of fatigue on performance and morale. Getting this balance right will be the central challenge of medical education over the next decade.
In my view, the IOM is to be commended for thoughtfully reviewing the issues and developing a set of recommendations (likely to be embraced by the ACGME) that seem quite sensible and balanced.
So let us old fogies cast aside the warm afterglow of our residency experiences and admit that we’ve blocked out the memories of the bone-crushing fatigue, the errors caused by the immoral mantra of “see one, do one, teach one”, and the all-consuming fear that we would crash and burn, with nary a safety net in sight. Once we get over romanticizing the past, we can start figuring out how to work within these sensible limits on hours and supervision requirements to create a more perfect system for both our trainees and our patients.