Imagine yourself stuck in the hospital. Would you rather your doctors be well-rested, with a limit on how many hours they can work? Or would you rather they work longer shifts, seeing you through the critical hours of your illness and with fewer handoffs of your care? That's the choice being reexamined after a study published in March in the The New England Journal of Medicine found that longer shifts for medical residents were just as safe as shorter shifts. The results, which support an earlier study that also found no association between shift length and patient safety, have led some physicians to suggest that the issue of how long residents should work has now been "laid to rest." University of Pennsylvania's Dr. David Asch, one of the study's lead investigators, said in a statement that despite concern about lengthy shifts for residents, "they really don't seem to have an effect on any important domains." Sixty-three internal medicine residency programs participated in the latest study. Half of the programs adhered to limits on how long residents could work. First-year residents were restricted to 16-hour shifts, while more senior residents could work up to 28 hours. For the other half, there was no limit on how long residents could work. The researchers focused on patients' mortality rates. They found that the number of patients who died within 30 days of admission to the hospital was similar between the two groups. Though the study didn't examine medical errors directly, the implication is that the rate of medical error was also similar. The findings are reassuring, given what we know about the effects of sleep deprivation on cognitive performance. Some studies have found that sleep deprivation can lead to a similar level of impairment as heavy drinking. Concerns about exhausted physicians in training have been around since at least 1984, when a college freshman named Libby Zion died in a New York hospital from an error made by a sleep-deprived resident. After a high-profile investigation, New York state instituted the first limits on how long physicians in training could work — no more than 24 hours in a shift, and no more than 80 hours in a week. These limits were extended to all residency programs in 2003 and strengthened in 2011. It seemed like a prudent change. As many have argued, we don't let our pilots fly without adequate rest. But some physicians challenged the limits, pointing out that we had no direct evidence to support the change. It had simply never been studied. Shift limits could have unintended consequences. For example, shorter shifts mean more handoffs between physicians, during which important information can sometimes be lost. Does the latest study mean that strict limits on shift length should be rolled back, since they've not been proven to help patients? Or is there more to the story than medical errors and patient safety? Just as sleep deprivation has been shown to impair cognition, so too has it been found to dampen empathy for others. Studies show that a single night of sleep deprivation interferes with our ability to perceive emotions in others and compromises our ability to empathize. Longer periods of sleep deprivation reduce measures of emotional intelligence and interpersonal functioning. One of the darkest moments of my medical training came at the end of a 28-hour call in the ICU. Firefighters had rescued a woman from a house fire, badly burned and barely alive. She was at the threshold of death when she arrived. Her suffering must have been unbearable, but I'm ashamed to say that all I could think about was how long her death paperwork would keep me awake. I don't think I'm alone in finding that sleep deprivation warps my ability to provide compassionate care. In 2002, researchers at the University of Pennsylvania surveyed residents before and after their first year of training. The prevalence of chronic sleep deprivation soared from 9% to 43%. Emotional exhaustion rose from 8.5% to 68%, and measures of empathy dropped 10%-15%. We can't say how these changes affected patients, but it's no stretch to imagine that they were probably harmful to the patient-physician relationship. They were certainly harmful to the residents — the prevalence of depression increased from 4.3% to nearly 30% by year's end, and burnout leaped from 4% to 55%. Skeptics about shift limits may point out that both groups in The New England Journal of Medicine study got a similar amount of sleep, when averaged over time. But participants in the group without shift limits were nearly 2.5 times more likely to be dissatisfied with their amount of time for rest and with their overall well-being. They were also more likely to report that patient safety and the quality of their education had suffered as a result of the longer hours. In 2017, after the publication of an earlier study showing no association between shift length and patient safety, the governing body of graduate medical education backtracked on shift limits. Maximum shift length for first-year residents was increased from 16 to 24 hours, and more shift-to-shift variation was permitted for senior residents, as long as the 80-hour weekly cap was maintained when averaged over four weeks. It was an important development, according to Asch and his colleagues, because it seemed that educational policy would finally be based on evidence rather than opinion. Moving toward evidence-based policy in medical education makes sense. After all, as Asch pointed out, we would never approve a new drug without strong evidence supporting its use. But it's my view that using these studies to justify further loosening work-hour restrictions doesn't make sense. This new study provides strong evidence that shorter shifts are just as safe for patients as longer shifts. Given what we know about the effects of sleep deprivation on emotional capacity and residents' well-being, why would we relax work limits again without proof that doing so would cause no harm? Clayton Dalton is a resident physician at Massachusetts General Hospital in Boston. 5/10/19 A previous version of this story incorrectly said the recent New England Journal of Medicine study was published in April. It came out in March. Source
