AI/Data Science

Challenges in Sepsis Prediction and Prognosis

Concurrent Session Moderator(s)

• CC

  Christopher Colombo, MD, MA, FCCM

  Associate Professor
  The Geneva Foundation
  Gig Harbor, Washington

  Disclosure(s): No relevant financial relationship(s) to disclose.

Multiple models exist for predicting both the likelihood of sepsis in specific patients and the risk of adverse outcomes. Some of these models, including the National Early Warning Score (NEWS) and its variants, are well validated and used widely but have limitations in both positive and negative predictive value. Newer predictive models based on machine learning, biomarkers, and other sophisticated constructs may have greater discriminative properties but are less practical to implement. This session will cover the state of the art in sepsis prediction, evaluate the role of machine learning and artificial intelligence, and assess the impact of novel biomarkers.

Presentations:

• Physiologic Scoring Systems for Sepsis Prediction
  

  Concurrent Session Faculty: Charlotte A. Thomas, DNP, AGACNP-BC, CCRN – Scripps Mercy Hospital
• Machine Learning Models for Sepsis Prediction
  

  Concurrent Session Faculty: Rishikesan Kamaleswaran, PhD – Emory University School of Medicine
• Beyond Procalcitonin: Emerging Biomarkers for Sepsis
  

  Concurrent Session Faculty: Michael S. Niederman, MD, FCCM – Weill Cornell Medical College