Published in the March 2008 issue of Today’s Hospitalist
Gravity, for better or worse, used to determine which patients Adam K. Horeish, MD, saw first. Regardless of who needed to be discharged or how backed up the emergency department was, Dr. Horeish would start at the top of Banner Baywood Medical Center in Mesa, Ariz., and work his way down each of the seven floors. Like many hospitalists, he wasted a lot of time and energy backtracking to follow up on lab results, answer pages or track down a specialist.
All that changed when the hospitalists decided to pilot a geographic distribution model, a concept that is also called a unit-based or geographic assignment.
At Banner Baywood, geographic distribution means that one hospitalist is assigned to a 26-bed med/surg unit; physicians assigned to that unit see only those patients. The unit-based location eliminates travel and phone calls, not only making hospitalists more productive, but freeing up more time to talk to patients and family members, consultants and nurses.
Banner’s experiment with geographic distribution has been so successful that the hospital expanded the model to other units. It is in place as a two-month rotation in five of the hospital’s eight units.
Although he was initially skeptical, Dr. Horeish now can’t say enough about the positive benefits of being in one place all day. Along with having easy access to patients and colleagues, he says that the lengths of stay in those units have dropped by a day.
Just as important is the new sense of ownership he feels toward patients. “Before,” Dr. Horeish explains, “it would just be a patient. Now I feel like it’s my patient.”
Despite such glowing reviews, the idea of unit-based assigning is often a tough sell, particularly among physicians. Many hospitalists are drawn to hospital medicine, after all, because they don’t want to be corralled in one place all day. Instead, they want to be free to roam the hospital and see a diverse patient population.
Physicians also worry that given the vagaries of where and when beds become available, they may get stuck with sicker patients and more discharges or admissions. The fear is that hospitalists working in a specific unit could be saddled with a heavier workload than their free-ranging colleagues.
There are also concerns about how unit-based assigning may affect continuity of care if patients transfer in and out of different units while hospitalized.
While those are legitimate concerns, most physicians who have tried some form of geographic admitting say that equity issues work out over time. They also point out that they have no intention of going back.
“The ‘wandering hospitalist’ needs to stop,” says Larry Spratling, MD, Banner Baywood’s chief medical officer. “It’s like the horse and buggy “and it needs to go the same way.”
How it works
Smaller hospitals where physicians don’t have that much ground to cover may not benefit from the unit-based model, which can take one of several forms.
In some hospitals, as many medicine patients as possible are assigned to a single floor where hospitalists can spend much of their day when they’re not visiting patients in other units, such as the ICU or telemetry. But some hospitals like Banner Baywood take a more radical approach and assign hospitalists to a particular unit.
A growing number of physicians see these types of assignments as a logical corollary of the hospitalist model, which is designed to centralize care and maximize the time available to patients and staff.
“If you really believe the role of hospital medicine is to meet the needs of patients, you have to ask, ‘Where are the patients specifically located in relation to attending physicians?’ ” says Mark V. Sheffield, MD, assistant physician-in-chief for Kaiser Permanente, which has introduced the geographic concept at two hospitals in Northern California. “Why not have hospitalists work at one site instead of caring for patients at multiple sites?” That same principle, he points out, is the model used for ICU medicine, labor and delivery, and pediatrics.
The concept of geographic assignments is also getting a closer look as an adjunct to quality improvement. “Most ICUs wouldn’t dream of having core staff floating around five different units,” says Don Goldmann, MD, senior vice president of the Institute for Healthcare Improvement. The solution, he says, is to create a microsystem to improve the system of care under your control.
Convenience and efficiency
Physicians working in geographic units list what they say are the model’s many benefits.
Topping that list are convenience and efficiency. By being in one place, hospitalists don’t have to be paged to talk to families or consultants, and they’re nearby for unexpected crises.
Physicians working the geographic units at Kaiser, for instance, have seen their number of pages plummet by 90%. “It’s amazing how much more you can get done when you’re not spending time going to separate locations in the hospital,” Dr. Sheffield says.
Not surprisingly, nurses like having a doctor available all day without having to call for orders and wait for callbacks. In fact, improving communication and relationships with nursing was a major factor that drove the 400-bed Lynchburg General Hospital in Lynchburg, Va., to create a hospitalist unit. Nursing turnover on that unit is the lowest by far of any unit in the hospital, says Jim Pittard, MD, a hospitalist with the multispecialty Medical Associates of Central Virginia Inc.
Geographic units can also lead to a new level of team-building among providers. Before Regions Hospital, a HealthPartners facility in St. Paul, Minn., started geographic units in 2004, the hospital held multidisciplinary discharge rounds twice a week. Hospitalists at the 427-bed facility, however, couldn’t participate because the teams would have had to page 11 different doctors.
Today, daily team rounds are held in each unit at the same time every morning, with physicians, nurse and case managers, physical therapists, and pharmacists reviewing patient progress, family communication plans and pending tests. According to Burke T. Kealey, MD, assistant medical director of hospital medicine for Health Partners Medical Group, “We saw huge gains in teamwork, camaraderie and efficiency.”
Boosting throughput and quality
Patient satisfaction scores can also benefit. At Banner Baywood, patient satisfaction scores for physicians in the first geographic unit were 99%, compared to 15% hospital-wide. Patient satisfaction scores related to discharge were likewise 99% from patients in that unit vs. 27% hospital-wide.
According to Dr. Spratling, Banner Baywood’s geographic units continue to post dramatically improved throughput. Because 71% of discharge orders are written by 9 a.m., the hospital has been able to eliminate the previous 3 p.m. discharge “rush hour.” The first unit to adopt the model has also seen 40 more patients per month.
Dr. Horeish says he has experienced that throughput change first-hand: He now sees 20 patients during his 7 a.m. to 4 p.m. shift instead of the 15 he saw under the old “gravity” system, yet he finishes earlier. As a result, he works 15 days per month instead of 20 but earns the same amount of money and takes more time off. It’s no wonder, he points out, that Banner Baywood’s physician satisfaction scores for geographically-based hospitalists were the highest in the hospital.
Kaiser’s Dr. Sheffield, however, warns against using geographic units just to boost the number of patients seen. “That’s not a goal,” he says, “and it hasn’t happened at Kaiser.” Instead, he says that Kaiser sees the units as a means to better focus on patient safety and quality.
Is equity possible?
Geographic admitting also introduces problems that need to be addressed. For one, there’s continuity. Many physicians worry that the number of handoffs will increase if patients need to be transferred in and out of different units.
But hospitalists who work these units say that concern is more theoretical than real. Instead of increasing handoffs, they point out, having a dedicated hospitalist on the unit eliminates many handoffs.
And some hospitals build in a level of compromise to minimize the lack of continuity. At Lynchburg General, for example, a hospitalist who admits a patient in the emergency room or who followed a patient in and out of the ICU will continue to treat that patient, even if the person isn’t in the hospitalist’s unit. “That’s better for the doctor and the patient,” says Dr. Pittard.
Then there are concerns about inequitable workloads. Will hospitalists tethered to a unit that has half its beds suddenly open up be inundated with admissions, while roaming colleagues see fewer?
Most groups address volume concerns by rotating hospitalists weekly, monthly, bimonthly or even daily. In Dr. Pittard’s group, for example, patients are divided among the 14 hospitalists who are then told which one of them is “doing the unit” “the 21-bed hospitalist-only unit that’s been up and running for about a year.
Others tweak the system by adding a floater. At Regions Hospital in St. Paul, for instance, the hospitalists rotate in a “sweeper service,” taking patients on a floor that has been hit particularly hard that day. In addition, Regions pairs hospitalists from two contiguous hospitalist-only units to make sure their workloads are even.
“Equity does not happen on a day-to-day basis,” Dr. Kealey says, “but over weeks instead.”
A similar concern is how to even out the number of admissions from the emergency department to various units. While some meet with admitting staff to come up with a plan, others use queuing theory to devise a more precise system. (See “Geographic units and queuing theory,” below.)
Then there’s the pushback that comes with such a drastic culture change. In one hospital, a few physicians balked at rounding with nurses. At Regions, Dr. Kealey says, hospitalists worried about having to attend a half-hour team meeting every day.
“It took months for them to realize that the meetings were saving them time,” he points out. Each unit coordinator, who would also attend the meeting, would then work to facilitate tests and discharges. “The doctors figured out that they no longer had to call down and beg for a CT.”
Experts say that more research is needed to gauge the model’s impact on continuity of care. They also say that the concept of geographically localized units for hospitalists is still too new to generate anything other than preliminary data on how units affect quality and patient safety.
One hospital is even studying the difference between how many flights of stairs and steps physicians have to travel each day. But most hospitalist groups that have taken the plunge plan to build on their initial results.
At Kaiser, Dr. Sheffield says that gains in teamwork, collaboration and communication produced by the model have cemented its future. The next step is to pair a unit physician manager with the nurse manager to help the rounder.
Next month, Lynchburg General will set up its second hospitalist-only unit. And at Regions Hospital, Dr. Kealey says the model is proving to be an effective recruitment tool.
However, he warns hospitalists who plan to take the plunge to expect some resistance. His advice is to have a clear vision, a strong design and plan, and the flexibility to modify the model over time. “Persevere and fix it,” says Dr. Kealey, “and it will work out.”
Paula S. Katz is a freelance health care writer based in Vernon Hills, Ill.
Geographic units: a cautionary tale?
TALK TO PHYSICIANS who work in hospitalist-only units, and many say that the units improve their sense of autonomy and job satisfaction.
But at the 750-bed Staten Island University Hospital in New York, the lack of equitable workloads that resulted when geographic units were established killed the effort to keep them afloat. While the model was well-received when the units were first launched five years ago, the system just couldn’t guarantee that each hospitalist would get similar numbers of patients over the course of a week or even a month.
Hospitalists assigned to those units grew frustrated watching other colleagues who had a lower daily census and a disproportionate amount of free time, says Aaron L. Gottesman, MD, director of hospitalist services and associate program director of the internal medicine residency.
Another problem? Hospitalists assigned to geographic units on teaching floors felt that they were being treated as residents. The fact that they were always available led to them being asked to do more scut work, Dr. Gottesman says. That in turn resulted in awkward relationships, tension and conflict.
“They became a PGY 1, 2, chief and attending at the same time,” Dr. Gottesman says. “They felt it blurred the lines between resident and attending, and removed the prestige and respect that the attending position deserved.” (Because it can be difficult to maintain a demarcation between resident duties and attendings’, some hospitals that use geographic distribution don’t include teaching floors.)
Efforts to address the discontent included first intermingling, then integrating, the critical care units. After 18 months, however, the hospitalists dropped the project to once again visit multiple units. To make communication easier, the hospitalists now use cell phones internally.
Geographic units and queuing theory
THE MAGIC NUMBER IS EIGHT. That’s the total number of beds that Cleveland Clinic figured it needed to reserve each night in the four units that are staffed by hospitalists to prevent emergency department logjams and enhance patient flow.
Hospitalists arrived at that number a year ago by using a mathematical “queuing” equation that looked at admission and discharge data. A business term that’s getting a lot of buzz in health care, queuing theory is used to calculate how long “customers” “in this case, patients in the emergency department “need to wait in a queue before they can be admitted to a hospitalist-only unit.
The goal is to minimize the risk that the hospital would under- or over-hold beds. By tracking the average percentage of unit beds that would be occupied based on hourly admission and vacancy rates, hospitalists calculated that the number of empty beds they needed to reserve was eight.
Sanjeev Suri, MD, MBA, a hospitalist at Cleveland Clinic who has a background in industrial operations, explains that because 80% of admissions arrive by 2 a.m., the eight beds are held from 4 p.m. until 2 a.m. If those beds are not filled by then, they can be used by another service.
On most nights, between 70% and 80% of those beds are actually used. Each of the four units is staffed by one of the hospital’s 30 hospitalists. A fifth “rotator” helps out in the busiest units or takes admissions if all the units are full.
“It’s the same idea that’s behind those customer service messages that tell you your wait time is going to be eight minutes,” Dr. Suri says. Dr. Suri and his hospitalist colleague, Ali Usmani, MD, studied admissions and discharge data from different floors over time to figure out how many beds to hold.
Applying queuing theory to those data, they found that the average number of patients waiting in the ED for a bed was one, while the average wait time was 1.6 hours. Further, 86% of medicine patients who were admitted were able to be assigned to a hospitalist-only unit.
Using queuing theory allowed the physicians to gauge the risk that throughput for other services will be upset by holding beds in a hospitalist-only unit, Dr. Suri points out. “Not quantifying risk can lead to fear overriding any rational decision-making,” he says. “Using queuing theory and operations data together, like we used in our modeling, put all possible scenarios with associated risk on the table, so you can make a more educated decision.”
While queuing theory can particularly help hospitals that are at capacity and have busy emergency departments, Dr. Usmani admits that queuing strategies come with challenges. Chief among those is the need to train emergency department staff in the queuing system and figuring out how to cope with seasonal variations in patient census.
But hospitals that delve into queuing theory to fill hospitalist-only units will find that it pays off, Dr. Usmani says. “It’s a better metric for workflow.”