Home Feature Theory and practice: a look at how to marry the two in...

Theory and practice: a look at how to marry the two in quality improvement

September 2007

Published in the September 2007 issue of Today’s Hospitalist

Talk to a quality improvement expert about strategies to improve patient care, and you’ll inevitably hear a lot about the theory and principles that you need to follow to make these projects work. What’s not always clear, however, is how hospitalists can successfully put those principles into practice.


Related article: Can midlevels take-up the slack ?

At a presentation at the American College of Physicians’ annual meeting earlier this year, two quality improvement experts did just that: They married theory of quality improvement with its practice in a hospitalist group.

Jeffrey Greenwald, MD, a hospitalist and director of the hospital medicine unit at Boston Medical Center, outlined the key building blocks that physicians can use to improve quality over the long run. He discussed strategies and common mistakes made in quality improvement initiatives.

Co-presenter Gregory Maynard, MD, chief of the hospital medicine division at University of California, San Diego (UCSD) Medical Center, described how he and his colleagues applied those principles to an initiative to reduce the incidence of venous thromboembolism (VTE) at his center.

Together, the two presenters gave an in-depth look at how the theory behind quality improvement initiatives can be used in real-world scenarios. Here’s a look at what they said.

Theory: Identify the problem

At the outset of any quality improvement effort, Dr. Greenwald said, you need to identify the problem so you know what you’re trying to fix.

That means defining the problem as specifically as possible and performing the necessary due diligence to find out what resources are available. You can use resources like standards or regulations from groups including the Joint Commission. Another good source of information is guidelines and recommendations from quality improvement organizations like the Institute for Healthcare Improvement.

“If there is a successful product or process out there,” Dr. Greenwald said, “steal it and tailor it to your institution and your local culture.”

That said, you also need to find preliminary data from your own institution when defining a problem.

Even if the data come in the form of anecdotes, Dr. Greenwald explained, you need local information to make sure you’re moving in the right direction to identify what you want to target.

Practice: Calculate VTE-related mortality

According to Dr. Maynard, preliminary research found that VTE accounts for 200,000 hospital-acquired deaths every year, 80,000 of which are preventable. “VTE is the most common cause of preventable hospital death, with big mortality, big morbidity and big cost,” he said.

And while physicians know that VTE prophylaxis can cut the incidence of VTE by as much as 60%, studies show that providers get pharmacologic prophylaxis on board less than half the time.

When Dr. Maynard set about collecting data from his own institution to define the problem, he found that morbidity from VTE was in fact an issue. “We had a lot of people developing hospital-acquired DVT and PE right under our nose,” he noted.

When the UCSD project began, 150 patients per year were developing hospital-acquired VTE, with one-third to one-half of those being potentially preventable. “They were not on what we thought was adequate prophylaxis,” Dr. Maynard explained.

While physicians were interested in preventing VTE based on clinical considerations, Dr. Maynard quickly realized that the hospital stood to benefit financially. The costs of caring for VTE at the hospital, it turns out, ran as high as $1 million a year.

Theory: Identify key players

According to Dr. Greenwald, another key step is assembling a committee to tackle the quality improvement project. He warned, however, that you need to think carefully about who those committee members should be.

“You have to ask yourself, ‘Who does this have an impact on at the hospital level, the clinical level and the patient level?’ ” Dr. Greenwald said.

When convening a committee, he added, keep in mind the following caveat: Never mistake seniority for first-hand experience. While you need buy-in and some participation from senior administrators, those aren’t the people you need dominating a committee.

Practice: Think beyond hospitalists

The team that Dr. Maynard helped assemble included himself as well as several other hospitalists; a pulmonary/critical care physician; a radiologist to help the team use the hospital’s digital imaging system; a pharmacist; a quality improvement data nurse; a nursing supervisor; and two front-line nurses. The committee also established an administration liaison with the chief medical officer.

Dr. Maynard emphasized the need in this project to make sure that the committee had representation by many subspecialists, including surgeons, orthopedists, and physicians in the trauma and burn units. “If they’re not part of your core team,” Dr. Maynard said, “you at least have to reach out to them and make sure you’re talking to some of their leaders.”

Theory: Map out workflow

Before you can implement any necessary changes, Dr. Greenwald said, you need to understand your current workflow from the perspective of your front-line users. That means not only collecting baseline data, but creating a “process map” that examines how the group currently practices.

Practice: Identify holes in DVT risk assessment

The VTE improvement team at UCSD helped pinpoint where necessary interventions weren’t taking place. That process revealed, for example, that physicians were failing to do a DVT risk assessment on patients.

Even when physicians did such an assessment, he added, they often disagreed about which agent or therapy patients should receive. And finally, patients who developed a clotting or bleeding risk over the course of their hospitalization weren’t being regularly re-assessed.

“Their platelet count would fall and all their medications would be stopped,” explained Dr. Maynard. “Then their platelet count would come back and be fine, but we’d forget to put them on DVT prophylaxis.”

Theory: Choose specific goals

Dr. Greenwald said it’s important to pick specific, measurable options. One tip: Avoid vague wording like “better” or “improved.” And while you want to be aggressive in improving patient care, he warned against making goals so ambitious that they can’t be achieved.

Practice: Set VTE targets

The UCSD team set two goals. One targeted processes, while the other targeted actual outcomes. Dr. Maynard said that to target care processes, the quality improvement committee created a protocol whose goal was to make sure that within 12 months, 90% of adult inpatients were receiving VTE prophylaxis appropriate for their risk.

In targeting outcomes, the committee said that the percentage of preventable hospital-acquired VTE in adult inpatients needed to fall by 75% within 12 months.

Theory: Identify “and test “the solution

Dr. Greenwald urged the audience to pilot small innovations and test them to see how they are working. (He advised using the PDSA cycle, an acronym for plan/do/study/act.) At this stage of the improvement process, he said, it’s critical to measure defined change, actively seek feedback and customize your project based on what people tell you isn’t working.

“It is not worth doing a process if you cannot measure it in many situations,” Dr. Greenwald said. “Get prospective data along the way, audit the process, get feedback and tweak.”

Practice: Build a workable order set

Because UCSD uses point-of-care decision support, the VTE improvement team incorporated an order set into the computerized physician-order entry system (CPOE). That order set is the default for VTE assessment and prophylaxis. The VTE risk assessment and prophylaxis are integrated into all admission and transfer order sets, which all physicians must navigate to complete their orders.

Dr. Maynard said that while the team came up with a menu of prophylactic options, it identified just one choice for each risk level. “You are narrowing the choice down from three or four acceptable choices,” he admitted. “But you are also eliminating 25 other choices that are not appropriate.”

While Dr. Maynard acknowledged that the idea of limiting treatment options may not sound very appealing to physicians, he assured the audience that physicians get used to such a pared-down menu very quickly.

“It does not really take away much choice from them,” he explained. “It is sometimes a trade-off, but not a huge one.”

A big key to the order set’s success, Dr. Maynard added, was that the team tested it relentlessly before roll-out. As a result, it was overhauled numerous times before going live.

For UCSD, that diligence paid off. After several months of consensus building, the VTE prophylaxis rate rose from 50% to 70%. Then, virtually overnight, as the VTE prevention protocol was integrated into CPOE orders, the prophylaxis rate rose to 90%. It was a stark illustration, he said, of the importance of changing systems, as opposed to simply providing education.

Dr. Maynard also pointed out that before the quality improvement intervention, the hospital had 13 patients per quarter with preventable VTE. After the project, he noted, that rate fell to one patient per quarter.

Theory: Ensure sustainability

When the committee shutters its doors, you don’t want the project to fold. Make sure key stakeholders throughout the hospital “not just the champions or committee chair “own the process. While you should check in on that process from time to time just to make sure it’s running smoothly, said Dr. Greenwald, be flexible with ongoing feedback and consider pegging some performance indicators to sustaining the ongoing change.

Practice: Don’t stop measuring

To keep tabs on the VTE initiative, Dr. Maynard said that the facility keeps a close eye on outcomes. “We usually reduce the interval or the sample size with each measurement, once we reach a target, but we do not stop measuring,” he said. “Sometimes this can be automated, but sometimes it takes a limited sample of manual data collection.”

Phyllis Maguire is Executive Editor of Today’s Hospitalist.