Cutting Doctors’ Hours May Not Reduce Medical Errors, Studies Find

Cutting the grueling work hours of doctors-in-training had little effect on reducing patient deaths, according to two large studies.

Death rates dropped in one group of patients in veterans’ hospitals but not in three other groups, the researchers reported.

The results come from what the authors describe as the largest and most comprehensive national look at work-hour restrictions, which were implemented four years ago in an effort to reduce medical errors by tired physicians.

“We were a little surprised,” said Dr. Kevin Volpp, the studies’ lead author and a physician at the Philadelphia Veterans Affairs Medical Center. “We thought that mortality outcomes would improve more consistently.”

The studies appear in this week’s Journal of the American Medical Association.

The new work-hour rules limit doctors-in-training to 80-hour weeks. Critics of the restrictions feared they would hurt continuity of care for patients and create a shift-work mentality among doctors. Others supported the limits, saying they might lead to fewer deaths from medical mistakes.

Before the rules, medical residents often worked 100-hour weeks, with some shifts lasting 36 hours straight. Although the new limits are still roughly double what other full-time jobs require, long hours for doctors-in-training are seen as a traditional trial-by-fire approach that give them necessary, intensive experience.

The two studies included 318,000 VA patients and more than 8.5 million Medicare patients at hospitals nationwide. The researchers looked at deaths that occurred within 30 days of hospital admission in the years before and after the rules went into effect in 2003. They compared death rates between hospitals with large number of residents and hospitals with few residents.

One study examined VA patients treated for a heart attack, stroke, gastrointestinal bleeding or congestive heart failure. It found that two years after the rules were implemented, mortality improved by 11 percent to 14 percent in major teaching hospitals, compared with hospitals with few residents. There was no change in mortality rates for VA surgical patients.

A parallel study of Medicare patients showed no significant changes in mortality for either medical or surgical patients in major teaching hospitals.

For the groups with no change, Volpp said one possible explanation is that more patient handoffs by residents offset the benefits of reduced fatigue. He also noted the work-hour limits may not be strictly enforced at every hospital.

Ingrid Philibert, a senior vice president for the Accreditation Council for Graduate Medical Education, said the lower mortality rate in the one VA group is “a very significant finding” that proves the restrictions work. The council implemented the rules.

As for the groups that didn’t show improvement, she noted that there are many aspects to medical care.

“It would be naive to expect that changing one input would produce a vast difference in outcomes,” she said.

Dr. David Meltzer, of the University of Chicago, interpreted the findings differently. He said the studies give the overall impression “that there just wasn’t any big effect on mortality one way or the other.”

That could mean that work-hour restrictions are not the best way to address doctor fatigue, said Meltzer, co-author of a commentary on Volpp’s studies in the same JAMA issue.

“There may be much better solutions than the one we’ve come up with,” said Meltzer, who has studied doctors’ use of naps.

Philibert said the council is already planning a pilot program this fall that involves naps. She said the two studies will help the council refine the work-hour rules.

Besides limiting work hours, the restrictions also require residents to have at least 10 hours of rest between shifts, and prohibit them from working more than 24 hours straight. Residents can work 30 hours if the additional six hours are for education or to transfer care.

___

On the Net:

Journal of the American Medical Association:
http://jama.ama-assn.org/

Accreditation Council for Graduate Medical Education: www.acgme.org