Critical Path Network: ED diversions reduced by tight monitoring

Borrowed processes allow real-time reaction

Using a combination of techniques that benchmark best practices from other industries, Roger Resar, MD, pulmonologist and change agent at the Luther Midelfort-Mayo Health System in Eau Claire, WI, has gained tighter control on monitoring patient flow and, based on preliminary results:

  • Reduced emergency department (ED) diversion hours from 12% to less than 2%.
  • Slashed the cost of diversions from about $250,000 a month to less than $30,000.
  • Increased from 23% to 40% the percentage of patients who were put to bed within one hour.

In addition, since nurses were given greater control over processes, staff vacancy rates dropped significantly and satisfaction rose, Resar says.

Resar says that a trip to Boston in January 2001 helped set things in motion. At that time, he and his CEO participated in an IHI-sponsored session about hospital flow. "They set up a group of experts to talk about what we might start to do to improve the problem," he recalls. "Our problem was not as acute as the one they had in Boston, but there were still times when the ER shut down, operations were cancelled, and so on."

Resar and his CEO eventually met Eugene Litvak, MD, professor of health care and operations management at Boston University, an expert in the field, and also involved one of their senior vice presidents in charge of nursing in the process.

"What we heard from that group of experts tweaked our interest, and we felt we could probably use some of their ideas," Resar notes. "So we came back and tested some of these ideas."

One realization that emerged early on was that the hospital had no good system for measuring flow through the organization. "We measured bodies that went through in a month or a year, but we had no real-time measurement of flow," he says.

"Every industry, every airport, any kind of large organization that deals with production measures flow at various points; they know where their bottlenecks are and what to do in contingencies. In hospitals, we don’t do that at all," he says.

Benchmarking these other industries really opened Resar’s eyes.

"My CEO, myself, and several vice presidents went to a local power company that serves a large area of the Midwest," he recalls. "Their office has a huge control board with a panel that runs all around a huge room. They can tell at any time almost down to a single telephone pole how much power is going through, how much is needed, and how much should be changed if there is an increase in demand. It struck me that at any given time in a given hospital, you can’t tell what’s going on; nobody knows."

Another instructive visit was made at the Minneapolis airport’s hub for Northwest Airlines. "They really need to understand flow," Resar notes. "If you notice, airplanes very seldom circle airports anymore before landing. That’s because you can’t take off from Boston, for example, until you have a landing slot in Minneapolis. In hospitals, we just take patients without knowing where they will land.’"

The first thing Resar and his frontline staff did was set up a measuring system that involves all patient floors of the hospital.

"We started out very small, using a pilot area, and did it all on paper before we moved to an electronic [computerized] system," he explains.

The system is based on the stoplight’ concept, designating red, orange, yellow, or green states of patient flow. "We set it up so that each unit would report on a regular basis, or if something changes drastically, what their color was," says Resar. "If you tell someone you’re having a red’ day, you really don’t have to go into a long description for other departments to understand you."

A given unit’s color is determined by using a measurement tool — a paper assessment anchored in several objective measures, the end result being the reporting of a color. This is done by the frontline staff.

By April 2001, Resar says, everybody in the hospital was using the system, which was accessed through its intranet.

One of the unique aspects of the changes instituted at Luther Midelfort, and a key to its success, was a mechanism to ensure action. "You can’t have a measurement system if people can’t act on it," Resar explains. "The first action we put into place was that when a unit reached a point, usually red, which in the frontline staff’s estimation meant they could not accept another patient, they capped the unit." This policy, he says, was called the Capping Trust Policy.

"If you were a frontline staffer and assessed your unit as completely saturated, you were allowed to cap your unit for safety reasons," Resar observes. "This was what Litvak was talking about; the limitation of elective procedures so that you could end up smoothing the artificial variability," he says.

A lot of admissions to hospitals are artificial variations, Resar asserts. "What if I get a call from a doctor at another hospital who says he’d like to send a patient over? In my former life, I would have sent him over. Now, I think about the airport. If I said yes without knowing there was a bed available, the patient ends up in the [emergency department] with no landing spot [and] waits for a bed."

Now, when a patient can’t go on a unit, the only way for physicians to have him admitted is to go through an admissions coordinator, who functions very much like an air-traffic controller. "We are a level 2 trauma center; we will bring in a patient if it is an emergency," Resar notes. "But if the patient is 100 miles away, has pneumonia, and needs tests done, they can wait to come in in a couple of hours, not right away," he explains.

The measurement initiative, Resar notes, "allowed the Capping Trust Policy, allowed us to shut down, and forced us to think about how to handle people coming in."

Another key element of the new system is what Resar refers to as upstream evaluation for downstream resource use. It addresses certain facts you must have in hand before scheduling an elective procedure. "If I’m a surgeon doing a procedure that requires an ICU [intensive care unit] bed and a ventilator after surgery, and if I have four operations on Monday and all will require a breathing machine for at least two days, those four will totally plug up the ICU," he observes.

"Instead, I might want to schedule just two procedures for Monday and two for Wednesday, and fill in the operating time with patients who don’t need breathing machines. We never did that until now. You can’t schedule procedures willy-nilly without realizing what the downstream effects will be," Resar says.

The new system has had a profound effect on nursing morale, he says. "In 2001, we had a nurse vacancy rate running somewhere around 8% to 10%," he recalls. "About six months after we started the Capping Trust Policy, it went down to 1% to 2%, and each unit shows the same dramatic change."

Resar has no doubt there is a connection.

"When frontline nurses get empowered to have a say-so in their work, it improves morale considerably," he asserts.

He also notes that despite the new system, admissions did not drop. "We were able to hire more nurses, but the other thing is this: We found out these nurses and other folks who work in our hospital are dedicated people. We don’t see them slacking. Yes, when it gets to the point where they just can’t take any more patients, they shut down for a few hours. But they are not taking advantage of the policy."