How AI can boost clinical decision support in emergency medicine

The utilization of artificial intelligence (AI) for clinical decision support at point-of-care is still in its early stages. Despite widespread media attention and an abundance of AI studies, the integration of AI into clinical practice remains uncommon.

Scott Levin, the Senior Director of Research and Innovation at Beckman Coulter and a professor in emergency medicine at Johns Hopkins University School of Medicine, sheds light on this issue. Levin is set to present at HIMSS24 in an educational session titled “Deploying Artificial Intelligence for Clinical Decision Support in Emergency Medicine.” The session will delve into two use cases where AI clinical decision support has been implemented across multiple emergency departments, covering the entire systems engineering success phases: problem analyses, design, development, implementation, and impact analyses.

The primary focus of the presentation will be on the latter deployment phases. Levin emphasizes that the AI tools discussed in the session aim to address challenges in emergency department triage and disposition decision-making. These key decisions often face high variability, bias, and limited prognostic validity.

Attendees will gain insights into the five Agency for Healthcare Research and Quality (AHRQ) systems engineering success phases linked to practical AI clinical decision support examples in the emergency department. Levin stresses the importance of establishing a framework for how AI tools tackle challenges, emphasizing the need for studying clinician interactions with these tools and understanding their impact on decision-making behavior.

Levin notes that it’s still relatively uncommon for AI tools to complete the full cycle, especially those operating at the point of care. He advocates for sharing more examples within the healthcare community to enhance visibility, with the belief that increased exposure to successful cases will benefit patients.

Additionally, the session aims to illustrate ways of studying and mitigating bias using AI. Levin emphasizes the importance of evaluating both AI algorithms and existing clinician decision-making structures for bias. The unique opportunity provided by AI to address biases directly at the point of care aligns with the healthcare community’s ongoing efforts to eliminate disparities in care.