Clot Prevention Improved With Computerized Decision Support

Larry Hand

October 16, 2012

October 16, 2012 — The use of a mandatory computerized clinical decision support (CDS) tool significantly improved physician compliance with guidelines on preventing dangerous blood clots that start in deep veins and migrate to the lung in trauma patients, according to a study published in the October 15 issue of the Archives of Surgery.

Elliott R. Hunt, MD, from the Armstrong Institute for Patient Safety and Quality at Johns Hopkins University School of Medicine in Baltimore, Maryland, conducted a retrospective study involving 1599 trauma patients hospitalized for more than a day at Johns Hopkins Hospital between January 1, 2007, and December 31, 2010. The academic medical center is also a level 1 trauma center.

The hospital implemented a CDS-enabled venous thromboembolism (VTE) ordering module within its computerized provider order entry system on December 17, 2007. The trauma center had a specific module that called on physicians to complete computerized checklists of VTE risk factors and potential contraindications to pharmacological interventions for preventing the clots, based on patient electronic health records. The system, using an evidence-based algorithm, then stratified patients according to risk factors and recommended either pharmacologic or mechanical preventive regimens.

Using the patients from 2007, which is before CDS tool implementation, as the baseline group, the researchers analyzed the effect of the mandatory CDS tool use on physician compliance and patient outcomes. They found that compliance with evidence-based VTE prophylaxis rose from 66.2% in 2007 to 84.4% after CDS tool implementation (P < .001). However, during the first-quarter learning period for the CDS tool, compliance dropped to 46.7%.

The proportion of patients receiving drugs to prevent clotting increased from 35.8% in 2007 to 43.4% in 2008 and to 52.6% in 2010 (P < .001). Orders for compression devices remained high throughout, except during for the first quarter of implementation, when the rate of orders fell below 50%.

In 2007, no preventive regimens were ordered for 2.8% of the patients. The proportion increased to 10.4% the next year, with the first-quarter learning period accounting for most of the increase. At the end of 2008, physicians had ordered some type of regimen for 99.1% of patients (P = .13), a number that rose to 99.7% in 2009 and 99.5% in 2010 (P < .001 for each comparison to baseline).

Thirty-five (2.18%) of the 1599 patients experienced VTE events. The proportion of VTE events decreased to 1.25% in 2010 compared with 3.0% at baseline (P = .23). In 2007, 4 patients (1.00%) experienced potentially preventable VTE events compared with 2 patients (0.17%) in the later years, which, the researchers write, represents "an 83% relative risk reduction in preventable harm (P = .04)."

Limitations of the study include the fact that overall VTE event rates were too low throughout the study "to provide sufficient power to detect statistically significant decreases" in events, the researchers write.

In an accompanying invited critique, George C. Velmahos, MD, MEd, PhD, from the Department of Surgery at Massachusetts General Hospital in Boston, also brings up that limitation, writing, "One needs to wonder about the cost-effectiveness of a system that, over 4 years, saves 2 patients from an event of unspecified significance."

Dr. Velmahos adds, "These are laudable results and a clear step in the right direction. At the same time, some healthy skepticism is unavoidable. Like all good studies, this one generates more questions than answers."

In addition to the low number of patients with VTE events, a question arises about whether an algorithm that recommends preventive measures for "almost everybody," when "ambulation is the best method of prophylaxis," Dr. Velmahos writes. The algorithm is also vague on the definition of prolonged immobility, he writes.

Nevertheless, the study researchers conclude, "[W]e demonstrated a dramatic decrease in preventable harm from VTE (1.0%-0.17%; P = .04) for admitted adult trauma patients, including 2 full years with no preventable harm VTE events."

Dr. Haut and one other author have received grant awards from the Agency for Healthcare Research and Quality. Dr. Hunt also receives royalties from Lippincott, Williams, & Wilkins and has given expert witness testimony in various medical malpractice cases, as have several of his coauthors. The study authors report a range of financial relationships with a variety of companies and institutions including sanofi-aventis, Bristol-Myers Squibb, Ortho-McNeil, Eisai, Daiichi-Sankyo, Janssen HealthCare, the Association for Professionals in Infection Control and Epidemiology Inc, the National Institutes of Health, the Robert Wood Johnson Foundation, the Commonwealth Fund, the Leigh Bureau, and the American College of Surgeons. One author also receives book royalties. Dr. Velmahos has disclosed no relevant financial relationships.

Arch Surg. 2012:147:901-908. Abstract