Researchers test the impact of CDS on ordering CTA pulmonary embolism studies in the ED
New clinical decision support system (CDS) mandates are on the horizon with an aim to reign in unnecessary imaging. But how feasible and clinically effective are these systems?
Authors of an Oct. 30 study published in Academic Radiology sought to answer those questions by identifying and piloting a CDS for CT angiography (CTA)-pulmonary embolism (PE) imaging studies which were thought to be overutilized in their institution’s emergency department (ED).
Alexander Goehler, MD, PhD, with Brigham and Women’s Hospital in Boston, and colleagues created an algorithm that combined established risk scores to recommend patients for PE work-up. It was then integrated into their Epic Radiology Information Ordering System, and piloted from April 1 through Oct. 31, 2015.
Of the 853 CTA PE studies completed, 8.2 percent (70 studies) were positive for PE. Goehler et al. found the algorithm “accurately stratified studies by recommendation." In total, 11.2 percent of patients for whom the algorithm recommended further testing had a PE, compared to the 2.6 percent in the group the algorithm recommended for no further testing.
The study revealed 15 percent of providers ordered more than 60 percent of all PE studies. Among those who ordered 10 or more, the mean compliance rates with algorithm recommendations was 45 percent, but ranged from 15 to 68 percent. Compliance and positivity rates achieved 10 percent correlation. The most common reason for not following algorithm recommendations were clinical concern and turnaround time, the authors noted.
“This study demonstrates the feasibility of implementing a clinical decision support for the ordering of CTA to assess for pulmonary emboli in the ED, both in terms of the algorithm performance and in terms of Epic's best practice advisory (BPA) as a vehicle for delivery of a CDS in the ED,” Goehler and colleagues wrote.
Despite the positive findings, the researchers found “many underappreciated challenges” to rolling out a CDS. They noted if adequate infrastructure is not in place to support implementation, its clinical impact and cost-effectiveness could suffer.
“It still remains unclear how radiology as a discipline will be affected by attempts to reduce costs via reforms such as bundled payments but a mandated CDS will allow for a comparison of its impact on both healthcare quality and costs across different institutions with similar patient compositions,” the authors concluded.