The pros (and cons) of unit-based staffing
How to get the most out of geographic units
Keywords: Unit-based staffing yields many positives and some surprising drawbacks
by Paula S. Katz
Published in the December 2012 issue of Today's Hospitalist
FOR ALL ITS BENEFITS—like fewer pages and much less travel—the key to making unit-based staffing truly successful is to figure out the right number of patients for each unit. That's one conclusion of a study published in the September issue of the Journal of Hospital Medicine, which examined the impact of assigning hospitalist patients to a localized 32-bed unit in an academic medical center.
Researchers found plenty of good news in terms of productivity and workflow. Compared to historical controls, for instance, the localized unit over the several months of the
trial had .89 more patient encounters and 2.2 more RVUs per day. And when held up to current controls, the unit team likewise had 1.02 more patient encounters and 1.36 RVUs per day and 51% fewer pages. But the localized team also racked up a 10% increase in length of stay (LOS) compared to two other teams that were not localized.
|"We expected length of stay to go down or at least remain the same."|
–Siddhartha Singh, MD, MS
Froedtert & The Medical College of Wisconsin
That was a surprise, says author Siddhartha Singh, MD, MS, associate chief medical officer at Froedtert & The Medical College of Wisconsin in Milwaukee. "We were expecting an increase in hospital efficiency that would be reflected in length of stay," he explains. "We expected length of stay to go down or at least remain the same."
Dr. Singh talked with Today's Hospitalist about how hospitals can determine what he calls the "sweet spot" of localization to make unit-based staffing work.
How were patients assigned to the unit?
The distribution of patients was not randomized, and a certain amount of judgment was involved. The admission group—emergency department physicians, an admitting medical officer who was a hospitalist and a nursing administrative representative—handled patient intake.
Together, they decided admission and patient placement. We made the best effort to make sure no clinical factors were taken into account and to base that decision instead on availability and perceived workload. During the trial, we tried to keep the 32-bed localized unit full.
Were you surprised the hospitalists on that unit were so busy?
We were expecting some change during the course of the trial. In general, the more efficient you get, the more work you get. But even beforehand, the hospitalists were suspicious about why the localized unit was being tried out. They wondered if the hospital was trying to make them work harder, and they shared that apprehension with us. It's not a huge problem, but one you need to manage upfront by building trust.
During the trial, we talked to people staffing the unit to get a sense of how they perceived the workload. As a result, we closed between two and four beds for a portion of the trial period. That way, each hospitalist-physician assistant team ended up seeing up to 14 or 16 patients, about the same as an average patient panel in an academic medical center.
Still, LOS increased. Why do you think that happened?
We had three hypotheses to explain this unexpected finding. First, we did truly increase the workload for the localized teams, thus compromising efficiency. Keep in mind that there are no good ways to measure workload in the inpatient setting. The number of patients per daily census is a poor workload measure because an extremely sick patient will take three to four hours of work a day.
Second, physicians were around these patients more so they could figure out small things going on with them. The problem list kept on increasing and thus they provided more care.
Third, and the most worrying hypothesis, is one of gaming. Teams felt that if patients stayed longer, fewer admissions would flow into the unit.
Which explanation do you think makes the most sense?
Everybody gravitates towards explanation No. 3 because it's the most provocative, but I'm still a little unsure about exactly what happened. Based on discussions and focus groups, it might have been a combination of all three.
What's the best way to avoid these problems?
If you're localizing units, try to avoid 100% localization. Don't have all patients cared for by a particular team on one nursing unit, and don't have all patients on a single unit cared for by only one team. It decreases efficiency.
Why is that?
If you put a hard cap on the total number of patients a team can care for, you create a setting where teams can game the system to decrease admissions. Teams may allow patients to stay longer to avoid new admissions.
Another issue: Most teams cannot maintain a level admitting capacity throughout the day, so a unit may end up having empty beds until localized teams start admitting. Say a team is admitting to a hypothetical unit from 7 a.m. to 3 p.m. Suppose that 10-bed unit has eight patients and two beds open, but it's 7 p.m. and patients are waiting in the ED. The localized team is not accepting patients, so those two beds have to remain open until the team is admitting. That's a waste of resources.
What's the ideal percentage?
After the project ended, we decided 100% localization was not good. So we shoot for a 70% localization effort. We don't know if that is the right number or not, and we're being slowed down by a new patient placement module in our CPOE that was implemented last August. It will take us up to six months to adapt.
What else did you discover on follow-up?
We did four focus groups of nurses who staffed the unit. They reported numerous anecdotes on the amount of attention patients were getting compared to nonlocalized units where teams typically round once a day for five or 10 minutes with each patient. With localized teams, the patient is enveloped by the team and sees physicians and PAs many more times a day.
So is unit-based staffing worth the effort?
My sense is that most results of the study were good. The number of pages per day decreased by half, the amount individuals had to walk fell dramatically, and the number of patient encounters by billing data increased. So there was better workflow and higher productivity.
What other advice do you have for hospitals looking to implement localized units?
Create a balanced scorecard that looks not just at LOS or the number of pages received, but at clinical outcomes and patient satisfaction. And if you do that, share your knowledge. Publish results and inform the hospitalist community, which is thirsty for such innovations—especially if they're reported in a way can be translated to other institutions.
Paula S. Katz is a freelance health care writer based in Vernon Hills, Ill.