Clinical Integration of AI into a Radiology Clinical Decision Support Tool to Minimize Workflow Disruptions

Monday, August 9 at 12:10 PM - 12:40 PM PDT
Wynn, Bandol 2
UNC Health recognized that implementation of a CMS-mandated radiology clinical decision support mechanism would likely cause provider workflow disruptions due to extra clicks and redundant information required for imaging order entry completion. To minimize data entry by clinicians, UNC Health was the first healthcare system to deploy the novel technology for AI-enabled indication selection available in CareSelect Imaging. During implementation, the UNC Health ISD Department determined that the CareSelect project should be a clinical stakeholder-led initiative with ISD support. UNC ISD project leads learned the existing radiology order entry process for ED providers. They then met with System ED Leaders to educate them on the CMS-mandate, and to prepare them for the new CareSelect workflows and AI interface. Technical solutions were developed within the EMR to silence the clinician-facing alert in emergency cases as allowed by CMS. Clinician education was developed to drive workflow compliance and to improve the AI clinical indication match rate. UNC Health EDs have emerged as the CareSelect national leader in successfully driving PAMA-compliant, non-interruptive ED provider workflows.


Danielle Degennaro
Clinical Imaging Applications Developer
UNC Health
Charul Haugan, MD, FACEP
Medical Director of Clinical Optimization: UNC Health
Emergency Medicine Physician: UNC REX Healthcare