Shifting views on shiftwork

The University of Pittsburgh Medical Center developed a new system to maintain night coverage in the ICU under resident work-hour limits.

Where: A 28-bed medical/surgical intensive care unit (ICU) at the University of Pittsburgh Medical Center (UPMC).

The issue: Maintaining night coverage of the ICU under resident work-hour limits.


After residents' schedules were restructured to meet work-hour limitations, leaders of the ICU had the usual concerns about increased handoffs and shrinking teaching time. “We always worry that it's going to affect the educational milieu,” said Lillian L. Emlet, MD, MS, assistant professor of critical care medicine at UPMC.

Since the institution of duty hours, the ICU's critical care fellows had been working a traditional schedule with overnight call every fourth night. That entailed many handoffs and a new person on every night.

In an effort to improve the safety of those handoffs and provide a little more consistency in staffing, UPMC developed a two-part experimental intervention.

How it works

In the first part of the new system, the ICU fellows were trained in structured signout. They received a two-hour educational session on signout (taught by Dr. Emlet) and access to a special signout feature in the electronic medical record.

For the second part, the fellows worked a shiftwork schedule of 12-hour shifts on a circadian design: forward cycling shifts with short strings of nights. The shifts included an hour of overlap between day and night to allow time for a structured signout. The new and old schedules were implemented in alternating four- or eight-week blocks. Both schedules allowed fellows at least seven days off per month.

The new system raised some concerns from the faculty. “They were resistant because one fellow was going to have to carry more patients per day,” said Dr. Emlet. The fellows themselves were noncommittal. “They commented that they were going to get seven days off no matter how you look at it,” Dr. Emlet said.


The new schedule and old schedule blocks were compared in a study published in Critical Care Medicine last December. Dr. Emlet and colleagues found that the systems were pretty much the same on many outcomes: no significant difference in patient mortality, fellow lecture attendance or patients' families' satisfaction. “From a first glance, it might be OK,” she said of the new schedule.

The shiftwork system performed even better than OK on a few measures. Length of stay on the unit was significantly shorter at 8.4 vs. 5.7 days (“I hope it's related to the fact that they were being watched very carefully on handoffs,” said Dr. Emlet) and faculty and nurses reported that they preferred the new system by a substantial margin. Additionally, there was no difference in the rate of readmissions back to the ICU.

Since nurses work a similar schedule, it's no wonder they would like the new one, according to Dr. Emlet. “Having the schedule posted very clearly at every nursing work station, it was very easy for them to figure out who was on call. At most large hospitals, it's really hard to figure out who the night coverage or the cross- covering team is,” she said. “It was more surprising that the doctors liked it so much, because they had the greatest resistance.”


Among those actually working the new schedule, there was a little more division. “The fellows really were split,” said Dr. Emlet. “Night float, not everyone in medicine loves….There were some people who were biologically better suited to doing shiftwork in general.”

After the study period, the new system also faced numerical challenges. “That year I was lucky enough to have many fellows,” said Dr. Emlet. “Our fellowship size varies slightly year to year. We do not do this particular model any longer, because we have also increased the number of ICUs we go to, and [the shiftwork schedule] has not been sustainable due to the expansion of our educational program.”

Words of wisdom

The system might be sustainable for other hospitals, she noted, and she encouraged others to conduct more experiments with it. “With four people [per ICU] you can do it,” she said. “Maybe the next study should look at parsing out for subpopulations and specialized population groups and also severity of illness and complexity to see whether or not there's any difference in mortality.”

Next steps

In the meantime, Dr. Emlet is focusing on the other part of the intervention: signout improvement. “To have one dedicated person to keep watch over the handoff processes…may be a job that actually needs to be created,” she said. “That's where I want to go next. I'm probably going to look at what do we say, and how do we say it.”