A growing body of work shows that Clinical Decision Support (CDS) systems can predict sepsis T0 through machine learning (ML) and the use of routinely collected electronic health record data (EHR). However, existing ML-based solutions are only as good as the quality and timeliness of the data presented to them. The aim of this FastTrack STTR project is to optimize and validate a-first-of-a-kind vigilant information-seeking (VIS) artificial intelligence (AI)-based CDS (AIVIS-CDS) system that goes beyond traditional pattern recognition, by autonomously generating and submitting nursing assessment and laboratory order sets to reduce prediction uncertainty. Moreover, we aim to implement a new sepsis treatment measure that makes use of HL7/FHIR standards and Causal Inference techniques to provide timely and actionable information to QI teams and nursing managers to enhance compliance with recommended care bundles.
Principal Investigator (PI): Bob Owens
Co-PI: Shamim Nemati
Contract PI: Christopher Josef
UCSD Team: Robert El-Kareh, Robert Owens, Atul Malhotra
Grant: NIH 1R42AI177108-01
Start Date: 07/07/2023
Expected End Date: 06/30/2024