Deployment of artificial intelligence for point-of-care clinical decision support is in its infancy. Despite media attention and proliferation of AI studies, translation to clinical practice is not commonplace.
Little evidence exists on best practices for deployment, particularly in emergency medicine. Scott Levin knows all about this. He is senior director, research and innovation, at Beckman Coulter, and professor in emergency medicine at Johns Hopkins University School of Medicine.
Two use cases discussed
Levin is scheduled to present at HIMSS24 in an educational session entitled “Deploying Artificial Intelligence for Clinical Decision Support in Emergency Medicine.” In this session, there will be two use cases of AI clinical decision support implemented across multiple emergency departments through the systems engineering success phases: problem analyses, design, development, implementation and impact analyses.
“Emphasis will be placed on the latter deployment phases,” Levin said. “The AI tools address challenges in ED triage and disposition decision making; key decisions that can be fraught with high variability, bias and limited prognostic validity.”
A major learning objective for those who attend the session will be to identify the five Agency for Healthcare Research in Quality (AHRQ) systems engineering success phases linked to pragmatic AI clinical decision support examples in the ED, he noted.
“It is vital for healthcare to have a framework for how AI tools address challenges, are developed, implemented and evaluated for impact,” he said. “This includes studying how clinicians interact with these tools and how it may change their decision-making behavior.
“It is still uncommon for AI tools to make it through this full cycle, especially those that function at the point of care,” he continued. “The more examples the healthcare community can gain visibility to, the better the chances of realizing benefits for patients.”
Mitigating bias in AI
Another objective will be to illustrate ways of studying and mitigating bias using AI.
“This includes evaluating both AI algorithms for bias and status quo clinician decision-making structures that may be biased as well,” Levin explained. “When the latter is present and measurable, AI provides a unique opportunity to address the challenges directly at the point of care.
“This is very important to healthcare today as the community strives to eliminate disparities in care,” he concluded.
The session, “Deploying Artificial Intelligence for Clinical Decision Support in Emergency Medicine,” is scheduled for March 12, 1:15-1:45 p.m. in room W307A at HIMSS24 in Orlando. Learn more and register.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.