Reducing Emergency Department Utilization by Targeting Care Interventions With AI/ML
Unnecessary emergency department use is a major source of preventable healthcare costs. As value-based care models such as ACO REACH push providers to full risk-sharing arrangements, more organizations are attempting to educate members and promote utilization of primary care resources, as greater primary care utilization is linked to lower costs and better outcomes. These outreach efforts are necessarily resource- and personnel-intensive, and healthcare providers are turning to AI/ML to better identify members at highest risk of these and other costly, preventable events. This session will discuss how targeting a multidisciplinary, complex care team to intervene with highest risk members using AI/ML led to a reduction in emergency department utilization that exceeded expectations.
Learning Objectives
- Illustrate the process for training, validating, and deploying customized AI/ML predictive models intended to identify and help prevent future costly, adverse events
- Compare the predictive accuracy and explainability of AI/ML models vs. rules-based logic
- Evaluate potential projected ROI as well as the assumptions and approach to measuring ROI
Speakers
