Where: University of California at San Diego (UCSD) Medical Center, a 325-bed hospital.
The issue: Reducing the incidence of venous thromboembolism (VTE) by increasing the use of prophylaxis.
Physicians at UCSD suspected that although VTE prophylaxis is effective, it was being widely underused at their hospital. The VTE Prevention Team, led by Gregory A. Maynard, ACP Member, confirmed their assertion that prophylaxis rates were low by doing a random audit of inpatient charts. In the 2005 assessment, they found that only about 50% of patients were being protected against clots. To get that number up, the team of pulmonary specialists, hospitalists and others developed an institution-wide prophylaxis protocol. The project was funded by the Agency for Healthcare Research and Quality.
How it works
First, the team members built a consensus about the standards for assessing VTE risk and choosing appropriate prophylaxis. Then they created a standardized tool that separates patients into three tiers—low, medium or high VTE risk.
A healthy ambulatory patient would be low risk, whereas a hip fracture or spinal cord injury patient would be high risk, for example. The tool (see Table), incorporated into UCSD's electronic medical record, then prompts physicians at each admission or transfer to assess the patient's VTE risk.
After patients are categorized, the tool suggests appropriate prophylaxis options, such as heparin or compression stockings. “Each level of VTE risk was linked to a menu of acceptable prophylaxis options,” said Dr. Maynard.
The team found that more than 80% of their patients fall into the medium risk category, for which pharmacologic prophylaxis was recommended. Overall, about 95% of the assessed patients are candidates for pharmacologic prophylaxis when it's not contraindicated.
“There is no gold standard for adequate prophylaxis for any given patient, so achieving consensus on a VTE prevention protocol and a risk assessment model that the medical staff would buy into was a hurdle,” said Dr. Maynard.
To overcome this, the VTE Prevention Team held a series of meetings with the institution's divisional leaders. “Sometimes gaining support took showing them an anecdote of a patient who failed their preferred mode of prophylaxis,” he said. The team also began using digital imaging to screen for VTE and pulmonary embolisms (PE) daily. “This opened everyone's eyes to the actual number of VTE and PE that were occurring in the medical center. It was a much larger number than most casual observers would guess because individual physicians don't see their own patients [have] many VTEs or PEs during the year,” said Dr. Maynard.
Another barrier was the lack of a prospectively validated risk-assessment model in the literature. “To overcome this, we originated one,” said Dr. Maynard. The team looked at the existing research and came up with a model they thought would work. “The ultimate validation was putting that model into place in the form of an order set and reducing VTE and PE,” Dr. Maynard said.
When UCSD measured prophylaxis rates in 2005, they ranged between 50% and 70%. By the end of 2006, the use of clot prevention hovered at 80%.
“It wasn't until 2007 when we were consistently above 90% that we had a real impact on our patient populations and their risk for VTE or PE. And now in 2008 we've been at 97% for the entire year,” said Dr. Maynard. Continuing analysis of the program found that the moderate-risk group saw the largest reduction in hospital-acquired clots, and overall rates of clot events have dropped by about 35%.
The program was so successful that the VTE prevention protocol has become the foundation of a VTE Prevention Collaborative, which was organized by the Society of Hospital Medicine (SHM) and so far includes 30 medical centers. As part of the collaborative, Maynard connects via phone and email with hospitalists across the country who are dedicated to increasing VTE prophylaxis at their facilities.
- The simpler the risk assessment model the more likely it is to be integrated seamlessly into the workflow. “An ideal VTE risk assessment model does not require the user to add up points [to ascertain the patient's risk level],” said Ian H. Jenkins, ACP Member, a hospitalist and member of the VTE Prevention Team. “Also, the model has to be convenient, but still maintain an adequate level of accuracy and detail.”
- Measurement drives improvement. “If it's done correctly with sampling and/or automation, and if you use digital imaging, like we did, you can get good data without a burdensome amount of work,” said Dr. Maynard.
How patients benefit
Of course, UCSD patients benefit from the reduction in PE and VTE, which is the leading cause of preventable death in hospitalized patients. Patients at other hospitals are now benefiting from the spread of the protocol.
Dr. Maynard said there's still work to be done as long as hospital- acquired VTE has not been eradicated. “There are still a lot of cases where there's a relative contraindication to pharmacologic prophylaxis, so clinicians justify not putting these patients on prophylaxis because of that. But the risk of clots is probably higher than the risk of bleeding from pharmacologic prophylaxis, so we plan to focus more on those patients,” he said.
The team also continues to collect data on VTEs and use that knowledge to further refine the system. “When patients develop a clot, we investigate why,” said Dr. Jenkins.
Words of wisdom
- “Don't reinvent the wheel. Use the VTE Prevention Collaborative tools on the SHM Web site. We put a lot of information into the toolkit that people can use to achieve the same results much more quickly than we did,” said Dr. Maynard.
- “If you involve and solicit the advice of everyone who may have their workflow altered, you'll have allies rather than people who you are trying to make compliant,” said Dr. Jenkins.
- More information about the VTE prevention protocol is available through the Web site of the Society of Hospital Medicine.