Study: Sleepy Doctors Make More ErrorsOct 27, 2004 | AP From prescribing overdoses to sticking a tube in the wrong vein, doctors-in-training made one-third more serious mistakes during typically long shifts than they did during "short" 16-hour ones, a Harvard study found.
At the same time, those first-year interns were wired up with electrodes to measure how often their sleepy eyes rolled, and they ended up nodding off more than five times a night during long shifts.
Together, the findings suggest that recently imposed limits on how many hours new doctors can work do not go far enough, the researchers said.
The studies were the first to measure the real-life toll that sleep deprivation takes on interns' medical judgment. The results were reported Thursday in the New England Journal of Medicine.
"There are currently more than 100,000 physicians-in-training in the United States, most of whom work these kinds of 30-hour shifts on a regular basis," said Dr. Christopher P. Landrigan, who led the study on medical errors.
Since July 2003, interns at U.S. hospitals have been limited to a four-week average of 80 hours a week. Also, they cannot work with patients for more than 24 hours straight, though six hours can be tacked on at the end for paperwork and classes.
"These long shifts are perhaps more hazardous than the number of hours in the work week," Landrigan said.
The two studies involved 20 interns and were conducted in the cardiac and medical intensive care units at Harvard-affiliated Brigham and Women's Hospital in the year before the new limits took effect.
Each intern spent three weeks in one unit, working at least 24 hours on every other shift, and three weeks in the other unit, with no more than 16 hours per shift. Doctors were hired from the outside to watch them work and note any mistakes.
During the longer shifts, the interns made five times as many diagnostic errors, such as missing the bull's-eye rash that showed that Lyme disease was causing a patient's heart problem. They made 36 percent more significant medical errors of all kinds.
There was no difference in the number of patient deaths and the average length of hospital stays, largely because other staffers often found and corrected the mistakes, the researchers said.
"This is testimony to the system of checks and balances we have in place already," said Dr. Anthony Whittemore, the hospital's chief medical officer.
Nurses noticed when one intern ordered 10 times the correct dose of a drug to raise blood pressure, and when another miscalculated a patient's fluid intake and missed symptoms of fluid retention.
But a tranquilizer overdose was not noticed until it caused dangerously low heartbeat and blood pressure. And one patient's lung collapsed because a tube being inserted into an artery punched a hole in the space around the lung, letting in air.
The interns also wore sleep monitors on and off duty, with electrodes attached to their heads. They nodded off about 5.5 times a night on long shifts, compared with 2.6 times overnight during the short ones. They got almost six more hours of sleep a week between short shifts, said Dr. Charles A. Czeisler (pronounced SIZE-ler), who led that study.
Because of the findings, Brigham and Women's has cut interns' hours to 12 in surgery and 18 in medicine, and plans more changes, Whittemore said.
Changes so far have cost the hospital $500,000, on top of nearly $1.9 million for those ordered by the Accreditation Council for Graduate Medical Education, which set the new rules.
Dr. Jordan Cohen, president of the Association of American Medical Colleges, said the current limits are "very much a work in progress" and further cutbacks in ICUs may be needed.
One of the interns who participated in the studies, Dr. Aaron Kesselheim of Cherry Hill, N.J., said he could not remember any specific errors he might have made. "A lot of intern year is a blur because everything is so crazy," he said.
Kesselheim said there was a downside to shorter shifts: He could not attend some teaching sessions in the hospital, and he had fewer chances to do procedures he needed to learn. Also, he could not see patients' full response to his treatment.
"Those first 24 hours are very important, seeing how patients respond to different management, trying to make a diagnosis," he said.