Using AI and NLP Analytical Innovations to Improve SDoH Data Capture and Care Management Services at CareSouth Carolina

CareSouth Carolina, one of the largest FQHCs in South Carolina, has committed to the use of advanced, AI-enabled data analytics to optimize its capture of SDoH data and the use of this data to drive improved outcomes and impact of its care management programs. This presentation will demonstrate the process and the lessons learned through the implementation of AI analytics for SDoH data and care services optimization at CareSouth, with support from the SCPHCA. The targeted areas where AI has been used to enhance the analytics and services are as follows: Use of AI analytics to aggregate and combine SDoH survey data with demographic, medical, and behavioral health data to create more accurate predictive risk models for populations with significant social and healthcare disparities served by CHCs; Use of AI models to discover which SDoH risk factors are contributing the most to the risk for future complications and costs; Use of natural language processing (NLP) applied to unstructured EHR care coordination notes to completely and accurately identify SDoH risk factors and services needed to address them; and Use of AI models for services impactability analysis to identify the care-enabling services most likely to reduce health risks and costs.