Residency redesign helps patients and pleases doctors

Brigham and Women's Hospital redesigned its residency program to focus on integrated teaching.

Where: The 72-bed medicine ward of Boston's Faulkner Hospital, a community teaching hospital affiliated with Brigham and Women's Hospital.

The issue: Increasing the educational value of residency for trainees while not compromising patient care.


The leaders of the Brigham and Women's residency program had a growing sense that their trainees were getting a less-than-ideal educational experience, with much of their time taken up by patient care to the detriment of learning.

“We felt like our trainees were being consumed by the rapid turnover of patients, the high patient load, and the turnover of the residents with the duty-hour requirements,” said Graham T. McMahon, MD, assistant professor of medicine at Harvard Medical School and physician at Brigham and Women's. “We resolved to see if we could come up with a different way of doing things that might optimize the educational experience while at the same time studying whether it might actually make a difference for patients.”

How it works

The program leaders designed an experimental general medicine service, which they called the Integrated Teaching Unit (ITU). Two teams were created, each with two attendings (one hospitalist and the other an internist or specialist), two residents and three interns. Each team conducted bedside rounds together every morning, and at least one of the physicians conducted additional teaching and reviewed the trainees' progress at the end of the day. The interns also had less frequent call than normal—they were on all night every sixth night, instead of covering until 10 p.m. every fourth night. The patient load for the team was capped at 15.


After a year, the educators compared the new model with the usual system (one resident, two interns, multiple attendings per team) in a study published in the April 8 New England Journal of Medicine. Trainees on the new teams were more satisfied, spent twice as much time in learning activities and had significantly fewer patients at a time (average census of 3.5 patients per intern vs. 6.6).

The satisfaction of both attendings and trainees with some of the changes was also observed anecdotally. “The dual-attending model was very popular with everybody. Going back to bedside team rounds, which has become a vanishing entity in U.S. medicine, was also very popular,” said Dr. McMahon. There was support even for some of the changes that the program designers had not expected trainees to like. “The fact that they were on call through the night, when they were on call, was not as unpopular as we thought at all. In fact, our interns appeared to appreciate that continuity,” he said.

How patients benefit

The really unexpected results appeared when the researchers compared patient outcomes between the two teams. On the ITU service, 26 patients (or 1.4%) died, compared to 48 (2.3%) of those treated normally. “We were personally very surprised that there was such a difference in overall patient survival,” said Dr. McMahon. “When we adjusted for everything from case mix to complexity of the patients or their other factors, it still was highly significant.”

The gains in survival were also accomplished with no loss in efficiency. The differences weren't significant, but length of stay and 30-day readmissions on the ITU service were lower than both the regular service and national averages. “The general concern before we ran the study was that if you cap teams at a census cap—in this case 15—then the teams will have little incentive to discharge patients efficiently,” said Dr. McMahon. “We found the opposite.”

The challenges

Although the project had many successes, not every aspect of the intervention worked out, such as a plan to group patients being cared for by a team in one area. “We worked hard to try to regionalize patient care into the team-based pods and we simply couldn't make that work,” said Dr. McMahon. A planned post-discharge follow-up clinic also didn't work out (see sidebar).

Lessons learned

The results of the experiment indicate that a new model of residency training is possible, and given the patient mortality statistics, maybe necessary. “One of the implications of those two numbers [26 vs. 48 deaths] is that the current standard of inpatient residency care could be improved,” Dr. McMahon said.

The researchers also calculated the costs of the program, and concluded that the extra spending, on attendings' salaries and providers (such as physician assistants) to care for patients who would otherwise be treated by residents, could be mostly offset by cost savings from shorter lengths of stay on the ITU since the hospital was readily able to fill empty beds. “Ultimately, the program was going to cost us, after all the deductions, less than approximately $100,000 per year,” said Dr. McMahon.

“We went out and showed that not only can you do this, but that thoughtful investment in resident education can result in improved patient care, improved resident satisfaction and improved efficiency, all at the same time,” Dr. McMahon said.

Next steps

The program has been expanded from Faulkner Hospital to the main Brigham and Women's hospital, and program leaders continue to track the results. “The data appears to be holding. Even when we took it out of a community hospital and put it in a major academic center, we're starting to see very similar improvements in length of stay and efficiencies,” said Dr. McMahon.

Words of wisdom

“Remember that trainees are there to learn and to work to learn,” Dr. McMahon said.