When Kevin Volpp was a resident at Brigham and Women’s Hospital in Boston, he once worked nearly 40 straight hours. It was nothing out of the ordinary.

For medical residents throughout the country, such 40-hour stretches and 100-hour workweeks weren’t at all uncommon until new national requirements, enacted in 2003, finally capped work weeks at 80 hours and limited shifts to a maximum of 24 consecutive hours.

The change was made to no small amount of protest. But two new studies authored by Volpp, an assistant professor of medicine and health care systems at Wharton and a staff physician at the Veterans Affairs (VA) Medical Center in Philadelphia, have begun to answer some of the questions surrounding the impact of the new regulations.

By analyzing more than 8 million patient hospitalizations in the Medicare system and more than 300,000 in the United States VA system, Volpp and co-principal investigator Jeffrey Silber, professor of pediatrics at Penn, found that while the new regulations significantly improved mortality outcomes in the VA hospitals, they did not affect rates among the Medicare patients. Volpp offered a number of theories as to why the effects differed in the different environments.

“Among the non-VA hospitals, only 30 percent are teaching,” Volpp says. “You might expect a larger effect in the VA because more of them are teaching hospitals. Also, the work intensity within the VA is probably lower. The residents are less overtaxed.”

Concerns over sleep-deprived residents making too many errors led to the regulations, but fears remain that shorter shifts will lead to less continuity for patients—therefore increasing mistakes.

“While there’s a strong basis of the relationship between sleep depravation and cognitive performance, there also was a real concern that the fact that each person would be working fewer hours meant there would be more transitions of care,” Volpp says. “For each patient in intensive care there have been studies saying 180 decisions are made per day. If you’re the one taking care of that patient, you’re making all those decisions.”

The VA study followed all patients admitted to acute-care VA hospitals from July 1, 2000, to June 30, 2005.

The Medicare study followed all patients admitted to acute-care non-federal hospitals during the same time period.

Both studies focused on patients admitted with principle diagnoses of acute myocardial infarction, congestive heart failure, gastrointestinal bleeding, or stroke; or general, orthopedic, or vascular surgery patients. The main outcome measure for both studies was mortality within 30 days of hospital admission.

While there were no significant relative increases or decreases in mortality for either medical or surgical patients in the post-reform years among Medicare patients, the VA system did find about an 11 percent improvement in mortality for patients in hospitals in the 75th percentile of teaching intensity as compared to hospitals in the 25th percentile of teaching intensity.

Volpp is the first to admit that while studies have provided some important information, more research is needed before any concrete conclusions can be reached.

“Overall, I would say what this suggests is the existing duty hour standards need further refinement,” he says. “I think one of the problems is they still allow residents to work 30 hours in a row. That’s a big problem because there’s lot of evidence that the rate of error is highest between 2 and 6 a.m. I think more work needs to be done designing and testing different ways of structuring work hours and seeing if that makes significant difference in changing patterns of outcomes.

“We really need to better understand what impact this has had on the quality of training of physicians. A lot of people are worried that compressing the amount of time will shortchange the amount residents will learn and change their basic approach to practicing medicine. They are going to be the next generation of physicians taking care of all of us. If the quality of their training is shortchanged, the quality of care is changed for years and years. It’s a very important question.”

Originally published Oct. 4, 2007.