Do resident work hours limits create better physicians?

On a few previous occasions (e.g., Student Debt, Resident Hours, and Primary Care Redux, Jan 2, 2009) I have written about the issue of medical residents’ work hours and the implications that it had, good or bad, for patient care. A recent article in the British Medical Journal (BMJ) by Moonesinghe et al, “Impact of reduction in working hours for doctors in training on postgraduate medical education and patient outcomes: systematic review”, provides the most thorough review to date of studies evaluating the impact of work hours reductions for medical residents in the US and the United Kingdom, where they are even more stringent. They looked at studies examining both “training” (how are the residents learning?) and patient quality and safety.

This was not an easy task; true “meta-analysis” (grouping all the studies together as if they were one and re-analyzing the data) was not possible because they were so different in methodology, issues studied, and quality. In fact, the quality of most of the studies was not terribly high, and they often looked at several different outcomes making it hard to understand whether the changes all went one way, or there were some things that were better and some worse and whether the net result could be stronger (or weaker) if we could focus on only one at a time.

The good news for those who have implemented, supported, and argued for the work hours restrictions is that they seem to have improved the private life of residents, and to have not significantly harmed either the quality of training or patient care, at least insofar as these somewhat contradictory studies seem to indicate. Most of the studies were done in the US, and most were of surgical (28 of 41) or “hospital-based” (e.g., anesthesiology, critical care) trainees, rather than primary care.

Here are some of the results:

Training outcomes:

·    2 papers (both of “low methodological quality”, one of medical residents in the UK and one of surgical residents in the US, which “did not report statistical analyses of the results”) reported an improvement in training outcomes.

·   12 studies found a detrimental association; half from the US (all surgical) and half from the UK.

·    27 studies found no change (20 US and 7 UK)

·    There were also mixed results regarding the quality of “training opportunities” (exam scores, caseload)

Patient outcomes:

·    31 of 34 papers were US

·    4 studies showed improved patient outcomes, including the only randomized controlled trial in the group (note: this was in critical care and coronary care units, where shift work may arguably be more effective)

·    2 studies (in trauma and orthopedics) found increased complications

·    28 studies showed no significant difference

Surgeon and NY Times writer Pauline Chen commented on this article in “Is a well-rested doctor a better doctor?”, April 7, 2011. She describes talking with a surgical resident she knows who is very pleased at how much better her life is than she had feared. “’Training has changed a lot…My life is different than yours was — I have a lot of time outside the hospital.’ She described how she loved her work but was able to sleep at home most nights, go out regularly with friends, stay involved with her church and take an improv class.”  Cool. It is good to know that surgical residents are not acting as slaves, and can have some kind of life outside the hospital.

However, her resident friend was less sure about learning and patient care: “’To be honest, I don’t really know if this is better or worse,’ she said, recounting how she felt she was signing over responsibility for her patients more often than she ever imagined she would, missing key events in their hospital course and even getting dismissed during the middle of a patient’s operation in order to stay within the limits on work hours. ‘Sometimes it seems so counterintuitive to just sign out as if we were shift workers, but this is all any of us know right now…We have nothing to compare it to.’”

So, probably, as Dr. Chen observes, the discussion will go on. Those who believe that working longer hours results in tired physicians and therefore bad patient outcomes will continue to push those ideas; those who (largely, it seems, surgeons) believe that artificial limits on work hours compromise resident learning, thus not only having a negative impact on their current care of patients but, more importantly, their care of patients in the future care because they haven’t had sufficient experience in their residencies. One “side” says “Do you want to be operated on by a surgeon who has been up all night?”, while the other says “Do you want to be operated by a surgeon who is not as skilled because they were coddled in their training and didn’t get sufficient experience?”

“The most important test of success of postgraduate training,” say Moonesinghe and colleagues, “is the professional performance of those who reach the end of it.” They go on to make several suggestions, including, most importantly I believe, that “a consensus should be reached by the medical profession on appropriate measures to assess the quality of postgraduate medical training.” They agree with the assertion (from Temple, et al, “Time for training. A review of the impact of the European Working Time Directive on training 2010”) that “training is patient safety for the next 30 years”. Wise regulation must understand the balance, the risk-benefit of any change, and try to reach the greatest benefit with the least risk.

I would like to comment a bit upon the issues as they relate, in particular, to primary care training. Most of the studies that have been done have been on surgical specialties, which can measure surgical complications, deaths or morbidity, or in anesthesiology or critical care, because these are hospital based and more amenable to shift work. While a very few of the studies were in pediatrics, they also examined hospital work; none were looking at the training of primary care or family doctors. One of the other reasons that the surgical specialties have been so studied is that they have long been those with the longest work hours; thus, they are both the greatest target of reformers and the greatest resistance by current surgeons and teachers who fear that work hours restrictions will jeopardize the skill of their future colleagues.

The irony is that, as in so many areas, when laws or rules are being violated by one group, they are tightened on everyone, and those who were not violators of the old rules find their greater restrictions to comply with while, often, those who were violating it before continue to. In the 1970s when I was in training at Cook County Hospital, we had a resident union (yes!) and were limited to every-4^th-day overnight call. But the surgeons were on more often. If family medicine or internal medicine or pediatrics violated the rule, they were penalized, but the surgery residents were afraid to complain. And so, today, the violations of hours rules in some specialties increases the restrictions for all.

The problem with applying these rigid rules to primary care is that it is not shift-based. While residents spend time on inpatient services, the core of family medicine training is the continuity clinic where those doctors-in-training follow their own patients. It is important to be able to do this, to show up for your office hours to see your patients who are expecting to see you, even if you were up much of the night delivering the baby of one of your patients. If that happened every day, it would be a big problem, but it doesn’t. Yet there is a “zero tolerance” for work hour violations by the Accreditation Council for Graduate Medical Education (ACGME), so the program would be cited. Rigid cut-offs, indiscriminately applied, are a bad idea.

Residents should have work hour restrictions; they shouldn’t be on call every third night and up all night. They should get days off, should get time to spend with friends and family or sleeping. But the restrictions need to regulate hours in a more global fashion: hours in a week, days off in a week, average or typical number of hours off between shifts. They should not be counting minutes (and they currently do!), not create automatic violations for each instance in which, say, a resident returns to clinic after only 9 ½  instead of 10 hours off. They also should be specialty-specific, examining the character of the specialty’s practice, not to allow exploitation but to make them appropriate to how the specialty is practiced.

Some fields, like ER and critical care, work well with shifts. Some, like most hospital work that characterizes internal medicine and pediatrics training, generally work pretty well with “night floats” and days off, as long as there is careful attention paid to information transfer at the shift changes (“hand-offs”). Surgery may require longer shifts with more days off. And family medicine needs to allow residents to occasionally deliver their babies at night without canceling all the patients on their schedule the next day, as long as it is not the everyday norm.

We can do this rationally. We can have training that both provides time for the non-work lives of residents and good training for their careers; that ensures quality care of their patients now and in the future.

Student Debt, Resident Hours, and Primary Care Redux

HAPPY NEW YEAR! May 2009 be a big improvement!

The December 18, 2008 issue of the New England Journal of Medicine includes Perspectives on 3 topics that have been previously addressed on this blog: Medical student debt (Dec 14), resident duty-hours (Dec 3, 9), and the future of primary care (Dec 11).

The piece on “Medical Student Debt – Is there a Limit?” by Robert Steinbrook presents data on the extraordinary rise in tuition, and debt, among medical students, most surprisingly in the public medical schools. “For the current academic year, tuition, fees, and health insurance at private medical schools range from $15,278 (for Texas residents) or $28,378 (for non-residents) at Baylor University to $51,969 at Tufts University in Massachusetts and $52,236 at Temple University (for nonresidents of Pennsylvania – state residents are charged at $43,232.” While tuition rates at private medical schools are generally higher than at public, the non-resident tuition at public schools is about the same as that of the privates, and the rate of rise (percent change) in the last 10 years at public schools has far exceeded that at private schools (100% vs 50% increase). Indebtedness ranged from an average (high is different) of $80,000 to $163,000 at public schools, and $70,000 to $182,000 at private schools. Some schools give significant tuition scholarships, but others are more challenged: Stanford’s endowment allows it to give a far larger number of scholarships relative to loans than does, say Drexel. More important, the article points out that the high debt burden may discourage lower-income students from applying to medical school, and to enter specialties with higher income potentials. “It is not surprising that a recent analysis showed a ‘strong direct correlation’ between higher mean salary in a specialty, such as orthopedic surgery or radiology, and the percentage of residency positions filled by US graduates.” The piece says that there is no easy solution, and probably there is not. But most countries have very low medical school tuitions, but require national service of their graduates.

“Revisiting Duty Hour Limits – IOM Recommendations for patient safety and resident education” by John Iglehart, discusses that topic in a balanced way. It points out the acknowledgment in the IOM report that “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.’” Igelhart also notes that “the 2003 limits on duty hours have resulted in an increase in handoffs of patient care between physicians – transitions associated with increased risks to patient safety.” I have discussed this at length, but I did note that this article includes a table with a recommendation I had missed – that internal and external moonlight be counted against the 80-hour per week limits. I have no difficulty with that conceptually, as it makes perfect sense – what is the point of limiting work hours in the residency if residents can moonlight for extra money in an unrestricted fashion? – but I wonder about the legal ability of program directors to restrict the moonlighting activities of their residents in their off hours.

“The Future of Primary Care – the Community Responds” involves a followup to a series of opinion pieces and a roundtable discussion with Drs. Thomas Bodenheimer, Barbara Starfield, Katharine Treadway, Allan Goroll, and Thomas H. Lee that appeared in the November 13, 2008 issue. The comments, and responses from the roundtable participants, are salient and generally useful. Several writers noted the role of physician assistants, and one (Paul Lombardo) states that “Patients, and the U.S. health care system as a whole, would be better served if the content of and level of primary care education were better matched to the needs of patients. The physician assistant (PA) model of medical education, with its emphasis on physician-physician assistant teams, needs to be expanded.”

These are all thorny, and not unrelated, issues. What is the relationship between resident work hours and physician’s assistants? Well, someone has to do the work. Since residents, even with the 80-hour restriction, work twice as many hours as do physician’s assistants, for about half the salary, and have a greater scope of practice, it would be incredibly expensive for hospitals to replace resident labor with that of physician’s assistants, not to mention physicians. As hospitals complain about the “cost” of resident education, this needs to be kept in mind; they are much better at accounting the cost than the benefit. Even if a hospital closes its residencies because it assesses the costs are greater than the benefit, this usually includes the fact that the residents care for many medically indigent patients, and you can be sure that the hospitals are planning to no longer care for them at all, not to pay someone else to do it. This, of course, again decreases access for the most needy.

I have repeatedly said that the nucleus of a solution is a comprehensive national health program, which includes a single-payer and a system that is tasked with ensuring the health and access to quality health care of all Americans. With such a system, addressing issues such as resident work hours, medical student debt, and the composition of the physician (and NP and physician assistant) workforce could be feasible; without it, they all remain insoluble because they all depend upon each other.