Integrating commercially available clinical decision support (CDS) software into an electronic health record (EHR) helps improve the appropriateness of imaging studies ordered by emergency and inpatient healthcare providers, according to a new study published by the Journal of the American College of Radiology. The shift was especially significant for trainees.
A commercially-available CDS tool was integrated into an institution’s EHR in June 2014, displaying “silent” appropriateness scores for six months. After those six months, feedback was “turned on” for adult patients in the emergency setting or inpatient setting for two years. The authors retrospectively looked at how CDS feedback impacted order appropriateness.
The CDS tool scored more than 6,000 studies during the “pre-intervention silent mode period,” finding that 11 percent were low utility, 24.5 percent were marginal utility and 64.5 percent were indicated. The tool scored more than 16,000 studies when feedback was turned on, finding that 5.4 percent were low utility, 12.6 were marginal utility and 82 percent were indicated.
“Locally developed or institution specific CDS software that has been implemented at several institutions across the United States has demonstrated that ordering patterns can change, with reductions in the number of low utility examinations ordered after implementing such systems,” wrote author Arun Krishnaraj, MD, MPH, associate professor of radiology and medical imaging at the University of Virginia in Charlottesville, and colleagues. “Our study demonstrates that similar results can be achieved with commercially available CDS and that the impact of these efforts can vary by provider type and imaging modality.”
Looking at those provider types, the authors found that low-utility studies ordered by attending physicians decreased from 10 percent to 6.7 percent. Low-utility studies ordered by trainees decreased from 10.8 percent to 4.8 percent.
The number of low-utility CT studies, meanwhile, decreased from 7.4 percent to 2.5 percent. Low-utility MRI studies decreased from 19 percent to 7.4 percent, low-utility ultrasound studies decreased from 11.3 percent to 3.8 percent and low-utility PET/NM studies actually increased from 15.3 percent to 29.4 percent.
“There was no improvement with NM studies,” the authors wrote. “The reasons for this are not known although the authors speculate that it may reflect that the content within our version of the CDS application that pertains to NM is not robust enough to demonstrate a difference.”