Members frequently ask our research team if imaging CDS "works." In other words, does this tool positively affect ordering behavior? A research study from University of Virginia Health System (UVA) sheds new light on the tool's effectiveness.
Results from University of Virginia
UVA implemented CDS in 2014 in the inpatient setting and ED. After six months of a silent roll-out, where ordering providers did not see feedback, the organization turned on alerts. Clinicians could then assess the appropriateness of their orders—and change or cancel them, if needed—at point of entry. While UVA did implement CDS for outpatient orders in June 2017, the health system for this study restricted data analysis to the two earlier care settings: inpatient services and ED.
After using the tool, imaging appropriateness improved in both care settings. In particular, CDS consultation decreased the relative frequency of low-utility imaging orders—or those scoring one to three on the appropriateness scale—from 11% to 5.4%. This trend was even more pronounced among trainees (residents or fellows); in this cohort, the frequency of low-utility orders dropped from 10.8% to 4.8%.
Creating a culture of appropriateness
The strong results advance UVA's overarching goal of using CDS to improve quality. Dr. Cree Gaskin, vice chair of informatics, explained why the health system chose to be early adopters of the tool: "We look at CDS as a way to improve patient care and experience. This tool educates providers at the point of order entry, and, as a result, patients should receive better quality care because they are more likely to get appropriate imaging tests and avoid unnecessary ones."
Plus, UVA's results demonstrate that trainees are especially receptive to the tool's feedback, suggesting that the health system is laying a foundation of imaging appropriateness early in physicians' careers that could have a long-term, positive influence on their ordering patterns.
Lessons from Advisory Board: Strategies to improve CDS effectiveness
While UVA's experience supports the hypothesis that CDS increases imaging appropriateness, it also presents opportunities to improve the tool. For example, the organization saw the biggest improvement in appropriateness scores for MRI, CT, and ultrasound—but no impact for PET and nuclear medicine. There are two potential explanations for this discrepancy: the lack of robust vendor content for PET and nuclear medicine exams, or training gaps for providers ordering those modalities.
To address the first problem, CDS vendors are releasing software updates on an ongoing basis to improve CDS clinical content. At time of publication, UVA's software had already released seven newer versions than what was used during the study time period, including updates from the Society of Nuclear Medicine and Molecular Imaging.
To resolve the second issue, Advisory Board research has found that organizations must proactively engage referring providers. For CDS education, it's imperative that leaders address the pain points of particular provider cohorts while stressing the overall quality benefits of the tool for providers to trust CDS interventions. Education doesn't stop at implementation; programs should distribute performance reports and consider selectively deploying hard stops to improve adherence.
These efforts require significant time and investment for organizations, but they will result in outsized outcomes. In particular, investing in CDS will improve care coordination, enable high quality imaging care, and advance value-based goals.
To help organizations navigate CDS implementation and fully realize the tool's benefits, the 2018 Imaging Performance Partnership National Meeting will have a dedicated session, Beyond the Mandate: Elevating the Value of Imaging Clinical Decision Support, that presents best practices for capturing the value of CDS.