Harvard students are no strangers to sleep deprivation—the tongue-sagging, chin-dropping, drool-inducing sort that wells up after a good 30-hour stretch without sleep. Fortunately, there is rarely time in an undergraduate’s daily routine when another individual’s life is in his hands. But first-year medical residents—doctors-in-training who are in their first year out of medical school and working in hospitals—are not afforded that luxury. They traditionally work around-the-clock shifts each day and investment banking-worthy hours (without corresponding wages) over the course of each week.

With this in mind, in 2003, the Accreditation Council for Graduate Medical Education (ACGME), which accredits post-medical school training programs in the U.S., implemented new rules that wisely limited residents to work no more than six days per week, 80 hours per week, and 30 hours per shift. But a report published this month by Harvard Medical School (HMS) researchers found that four out of every five residents had violated the regulations at least once in their first year of residency. With lives at stake, hospitals need to continue to work towards greater compliance with the ACGME rules.

Residents’ long hours stem from a worthy goal: a better understanding of the full cycle of emergency care, from a patient’s admission to her discharge. But hospitals take advantage of residents’ long hours as sources of cheap labor. The results, unfortunately, are dangerous, with an increased risk of accidental needlestick injuries, for example, at the end of a long shift. The HMS researchers, led by Assistant Professor of Pediatrics and Medicine Christopher P. Landrigan, also cite an earlier study that found that “human performance” after staying awake for 24 hours is comparable to human performance of those with a blood alcohol content of 0.10 percent; so a resident working a 30-hour shift might, by the end, quite literally be acting drunk.

The good news is that Landrigan’s team found that the new rules were effective in decreasing residents’ marathon work weeks, even if the decrease wasn’t in absolute compliance with the ACGME’s work limits. But for compliance to further improve, hospitals—including Harvard’s throng of teaching hospitals—should continue to introduce institutional changes that make compliance easier for residents. The HMS report, for example, suggests not requiring residents to work up to the very last minute of their scheduled shift, a situation which often leads to residents working overtime when an emergency occurs near the end of a shift. (Unfortunately, residents can do little to improve their own situations, since reporting their own violations could lead to de-accreditation of their own hospitals).

Should self-regulation by hospitals fail to make further progress, there is certainly an alternate possibility: in the U.K., by law, residents work no more than 13 hours per day and 58 hours per week. While we understand the need for intensive training for our next generation of doctors, it shouldn’t be so intensive that they injure their patients, or themselves, in the process. We echo Landrigan’s suggestions that hospitals must commit more resources and energy to reducing the load on their first-year residents, and if hospitals are unable to comply with the ACGME’s rules, then Congressional intervention may be necessary at some future date.