Published in the June 2010 issue of Today’s Hospitalist
When hospitalist John Kelly, MD, wore a pedometer to work one day last year at Albuquerque’s Presbyterian Hospital, he clocked five miles. During that time, his pager went off several times an hour, sending him running off in different directions. Because he was busy covering so much ground, he often missed the daily discharge planning rounds that nurses and social workers had recently launched.
Talk to Dr. Kelly, and the idea of moving to a unit-based staffing system in which hospitalists see nearly all of their patients in one geographical location is a no-brainer. Dr. Kelly is one of 40 hospitalists at Presbyterian, a program that cares for nearly 90% of the 500-bed hospital’s medical inpatients. Faced with that kind of volume, he says, the group is simply too big “and too busy “to have hospitalists covering patients in beds throughout the campus.
But one physician’s idea of a no-brainer is another’s no-go. A few years ago, geographic hospitalist staffing began floating to the top of many administrators’ to-do lists. The concept seemed to go hand in hand with other goals like improved patient satisfaction, teamwork and quality of care.
Some hospitalist groups responded with an all-or-nothing approach to unit-based staffing, and ran smack into problems including unfair physician workloads, scheduling complications and new bottlenecks popping up even as old ones were being eliminated. Those types of problems doomed many unit-based staffing experiments.
As a result, many hospitalists now are using the word “hybrid” to describe the type of unit-based staffing they are putting into place. As hospitalist groups roll these units out, they are treating some, but not all, of their patients in specific units “and putting a high premium on flexibility when it comes to patient assignments and physician workloads.
Settling for 50%
A good example of a hybrid approach can be found at Gundersen Lutheran Medical Center in LaCrosse, Wis., where hospitalists are launching a unit-based rounding system on a limited basis this spring. The goal from the start is for hospitalists to see 70% of their patients at any given time on one unit, with the remaining 30% on just one other unit.
By visiting other hospitals that have put unit-based staffing in place, explains hospitalist director Mary Frances Barthel, MD, “we learned that it’s pretty difficult to have your goal be 100% unit-based.”
At Northwestern University Hospital in Chicago, the hybrid model calls for some hospitalist teams to have between 85% and 90% of their patients on a single unit.
But some hospitalists have only 50% of their patients localized on one floor, and that’s as good as they’re going to get, says Kevin O’Leary, MD, associate chief of Northwestern’s division of hospital medicine. Those are floors where hospitalists don’t have enough available beds for their patients, and have to share the beds that are available with other services.
“You can’t make the perfect the enemy of the good,” says Dr. O’Leary, who notes that Northwestern launched its unit-based staffing program in January 2008. Starting with 15% of your patients in one place and ending up with 50% “is much better than running around to seven units.”
Targeting problem units
At St. Mary’s Health Center in St. Louis, a community teaching hospital, IPC’s academic hospitalist program takes another type of hybrid approach. For two years, it has maintained one dedicated 30-bed unit where all eight physicians in the group have some patients.
“We all see other patients throughout the hospital,” explains Philip Vaidyan, MD, the group’s director. “But we all spend more time in that one unit, building relationships and developing a sense of ownership.”
The IPC group at St. Mary’s plans to build on that success as it grows, but intends to maintain only a “regional” approach within the hospital. After the group hires more doctors this summer, Dr. Vaidyan says, several hospitalists will cover about 50 patients who will be spread throughout a few units in the hospital’s east wing, while several other physicians will cover 70 patients on the different floors in the west wing. The plan is to divide hospitalists between one side of the hospital or another “but not, he says, to try to localize any one physician’s patients to only one or two units.
And at Virtua Memorial in Mount Holly, N.J., another type of hybrid model focuses on the needs of patients.
Hospitalist director Erik DeLue, MD, MBA, explains that unit-based staffing occurs on only specific floors that have been chosen because they have a history of lower patient satisfaction and longer lengths of stay. Those floors are challenging for hospitalists, Dr. DeLue says, because they tend to house patients who have medical and social issues, making them more of a challenge for discharge.
The thinking was that these are the patients who would benefit the most from better coordination between doctors, nurses, case managers and social workers. Multidisciplinary rounds were seen as the fix for those problems “and unit-based staffing was the means to achieve that end. (See “Which should come first?,” below.)
“We started in reverse,” says Dr. DeLue. “Instead of saying that we want to do geographic rounding, we wanted to fix the problems on this specific unit, and geographic rounding was the solution we evolved into.”
The need for flexibility
After working with the model on that one floor, Dr. DeLue adds, patient satisfaction scores rose from the 10th percentile to the 85th. Length of stay improved, and “we found that the nurses were highly satisfied, and our doctors were happy because teamwork and communication improved.”
Hospital administration has also been pleased with those results, according to Ninfa Saunders, PhD, MBA, RN, MSN, Virtua’s executive vice president and COO of health services. “I have been most impressed with the impact of the care coordination project on patients,” Dr. Saunders says. “Having patients at the center makes for an outstanding patient experience within a multidisciplinary team.”
Virtua recently expanded the unit-based system to a second floor, and it has chosen a third. But Dr. DeLue is quick to point out that there are some floors where hospitalists work where unit-based staffing may never materialize “chest pain units, for example “because “care is more protocol-driven and the proximity of the doctors may be less important.”
One key to Virtua’s success with its unit-based model is a looser approach to geographic constraints. Patients transferred off a dedicated unit may be assigned to another physician, or hospitalists working in those units can follow patients to another floor to balance out physician workload.
“Initially, we said that if the patient is on the third floor, you sign out to the other doc, but now we are starting to allow for a little flexibility,” Dr. DeLue says. Patients likely to leave the next day “or people who are well-known to a particular hospitalist “aren’t handed off.
“Or if the census is really high that day, we make some decisions based on staffing needs as opposed to strict continuity,” he adds. “We are not fanatics about making sure that every patient on a unit is cared for by that unit-based doctor.” Paradoxically, says Dr. DeLue, “that can lead to less continuity.”
Making patient workloads even
Opposition to needless handoffs and concerns about inequitable workloads, in fact, are two of the biggest issues that have sunk unit-based staffing efforts. Aaron Gottesman, MD, director of hospitalist services at New York’s Staten Island University Hospital, calls his group’s attempt at unit-based staffing about six years ago a “failed experiment.”
For one thing, trying to be flexible can make it hard to figure out where to draw the line. At Staten Island, for instance, Dr. Gottesman says that his group allowed physicians to follow some patients off a unit for the sake of continuity. What evolved over time were so many exceptions, he says, that “there sometimes were more exceptions than rules.”
An even bigger problem was that hospitalists were unwilling to tolerate daily workload inequities, even though those differences would theoretically even out over time. Making matters worse, the workload “didn’t even out over time,” he says, “but was unpredictable.”
“More importantly,” Dr. Gottesman explains, “the doctors didn’t care whether it averaged out. They didn’t want to have a significantly higher workload on a daily basis. They would rather have had a similar average on a daily basis than on an annualized basis.”
Jesse Wagner, MD, vice chair of the department of medicine and a hospitalist for 12 years at Middlesex Hospital in Middletown, Conn., points out that any efficiency gained with unit-based staffing is lost when one person is overworked while another is underworked.
His hospitalist program is now trying its second iteration of unit-based staffing. Under this model, part of the fix for making sure patient loads are equal in units with geographic staffing is to designate one hospitalist as unit medical director.
While the unit director has most of his or her patients in that unit, directors also do brief daily rounds with a nurse manager on all the unit’s patients to improve communication and collaboration. (Unit directors typically have, says Dr. Wagner, one or two fewer patients to give them time for these daily rounds.)
“It is hard to get one doctor’s patients all on one floor while balancing all the other demands about who has how many patients, who has complex patients, who is doing an admission,” Dr. Wagner says.
The group also accepts the fact that some of each hospitalist’s patients will be on other floors “and it sees that as a plus, helping to guard against isolation. “If it is a med/surg floor, you are not seeing any cardiac patients,” Dr. Wagner points out. “Over time, you do want to see all kinds of patients.”
Making it work
To help ensure that physicians’ workloads are fair and balanced, Northwestern’s unit-based model includes an “overflow hospitalist” whose job it is the week he or she is assigned to that service to see patients throughout the hospital.
The overflow hospitalist also covers the 10 to 12 patients a day assigned to beds away from the medical units, usually because those units are full. “That’s your running-around week,” Dr. O’Leary says.
At Gundersen Lutheran, the hospital realized it had to take a more systematic approach to staffing and patient assignments to make geographic rounds possible. One big change, says Dr. Barthel, was renegotiating hospital policy to be able to pull nurses off other floors and send them to units with dedicated hospitalists.
Another was setting up a new department “the admissions referral coordination center, which Dr. Barthel is also medical director of “staffed by senior nurses who process all hospital admissions and decide where patients are going to go. “Now, patients 99% of the time are able to go to the right bed,” Dr. Barthel says, “and we can pull staff from wherever to cover them.”
But at St. Mary’s in St. Louis, hospitalists themselves even out each other’s patient load, meeting together every morning and reviewing who will be the attending for which patient.
As for regulating bed assignments, Dr. Vaidyan says that in the early days of having a dedicated hospitalist unit, ED staff in the middle of the night would assign patients to that unit, even when those patients weren’t covered by the hospitalist group.
A memo from the administration to all managers, however, spelled out the goals and potential benefits of the dedicated unit. Now, virtually all the patients assigned to that unit belong to the IPC group.
“In two years, we also went from a census of 60 to 120,” Dr. Vaidyan explains. “We don’t have any problem filling up our beds with our own patients now.”
Deborah Gesensway is a freelance writer based in Toronto who covers U.S. health care.
WHICH SHOULD COME FIRST: geographic assignments or multidisciplinary rounding? Some hospitalists insist that interdisciplinary rounds can’t possibly work if unit-based staffing isn’t already in place, but others contend that multidisciplinary rounds need to be the first priority.
Amy Boutwell, MD, a hospitalist at Newton- Wellesley Hospital in Newton, Mass., and director of health policy strategy at the Institute for Healthcare Improvement, admits that unit-based staffing is possible even without first putting multidisciplinary rounds in place.
But such rounds are superior, she contends, for care coordination and safety. The real payoff of unit-based staffing, Dr. Boutwell adds, is that “regionalization makes interdisciplinary rounding much more likely to happen.”
Unit-based staffing is helpful for not only multidisciplinary rounds, but for other quality improvement goals like increased patient and nurse satisfaction and more face time with patients.
But a leading reason to move to geographic staffing, says Russell Holman, MD, chief operating officer of Cogent Healthcare, is to cut the “tremendous waste and inefficiency” that comes from having “hospitalists with 15 patients spread out over seven units on four different floors.” At the first hospital in which he helped launch unit-based staffing eight years ago, Dr. Holman says that time-motion studies of the hospitalist group found that hospitalists were wasting “between 20% and 30% of their day, simply walking from one unit to the next and answering pages.”
Finding the “killer app”
IF YOU’RE CONSIDERING SWITCHING to unit-based staffing, get ready for a world of pain. That’s according to Russell Holman, MD, chief operating officer at Cogent Healthcare, a national hospitalist company based in Brentwood, Tenn. In his role with Cogent, Dr. Holman has overseen several hospitalist groups that have moved to geographic hospitalist staffing. “This has to be a shared vision among not just the hospitalist group, but hospital administration, the ED, the admitting office, bed control and nursing staff,” Dr. Holman says. “This will change what everybody does.”
The hospitalists at Gundersen Lutheran Medical Center in LaCrosse, Wis., resisted that change for several years. The idea of switching to unit-based staffing was first floated in 2006, says hospitalist director Mary Frances Barthel, MD. What she didn’t realize at the time was how emotionally attached people can be to the way they’ve always done something, even if that way doesn’t make much sense.
“Folks couldn’t see the point,” Dr. Barthel says. “They’d say, ‘Yes, maybe I’ll save some time, but I like getting exercise and going different places during the day.’ You really need to have a good reason to make such a significant change.”
The “killer app” that finally pushed the group to embrace unit-based staffing was multidisciplinary rounds. Once the group agreed that such rounds could help discharge planning, readmission rates and satisfaction scores, “we decided those rounds would be impossible with six teams doing to six different units,” says Dr. Barthel. “They needed a good reason for unit-based staffing, and this seems to be it.”
