Clinical decision support significantly reduces use of CT for pulmonary embolism

Clinical decision support significantly reduces the use of CT for pulmonary embolism among hospitalized patients, according to new research published Monday. 

Computed tomography pulmonary angiography is the go-to diagnostic tool for suspected PE, given its high accuracy. However, CTPA’s ease of use and availability has resulted in a doubling of exams over the last 20 years, with positive yields for pulmonary embolism as low as 3.8%.  

To address this concern, scientists at Mass General Brigham have implemented a clinical decision support system to aid referrers while ordering imaging. Assessing two years’ worth of data unearthed promising results, experts detailed in the Journal of the American College of Radiology [1]. CT pulmonary angiography yield was significantly higher when physicians followed the support system’s recommendations (18.3% positive for PE) versus when they overrode it (14.2%). 

Across over 340,000 hospital admissions, there was a 7.4% relative decrease in CTPA use after referrers started consulting clinical decision support. 

“Even when based on high-quality evidence, CDS-only interventions are unlikely to optimize high-cost radiology examination ordering behavior,” lead author Amita Sharma, MBBS, with the Department of Radiology at Mass General in Boston, and co-authors cautioned. “Supplementing CDS with multifaceted interventions—including performance feedback reporting using individual, group, and site-specific data, academic detailing, or timely peer-to-peer consultation when attempting to override CDS recommendations—may be needed to enhance the impact of CDS on high-cost imaging ordering behavior.” 

The retrospective study covered the 12 months leading up to clinical decision support implementation (2021-2022) and the 12 months that followed (2022-2023). Experts deployed the system across nine academic and community hospitals, covering a total of 3,853 beds. A multidisciplinary panel at Mass General Brigham designed the CDS tool over a two-week period, implementing it systemwide on June 1, 2022, in response to the global contrast shortage. A CTPA order triggered the intervention, while a patient’s positive D-dimer blood clot test in the past two days allowed an order to proceed. 

Sharma and co-authors noted significantly higher yield in the 2,429 CTPAs ordered post-intervention. However, across 5,372 exams during the entire study period, there was no difference in pulmonary embolism yield before (15.2%) versus after (16.5%) implementation. They noted poor adherence to D-dimer testing recommendations and high incidence of overriding CDS, emphasizing the need for additional changes to address low-value referrals. A negative D-dimer test was “strongly” associated with negative CTPA yield, with 0 of 60 turning up positive for PE. The study only identified four canceled CTPAs, suggesting the clot test does not influence ordering behavior. 

“The effect of D-dimer testing in our group would be enhanced if clinicians were more aware of the importance of identifying negative D-dimers,” the authors noted. “Clinicians' belief in the lack of specificity of D-dimer testing, particularly in hospitalized patients, may deter them from checking D-dimer. Although the number of negative D-dimers is likely small, the high sensitivity should encourage clinicians to cancel CTPA requests in this group.”