Development and Verification of a Digital Twin Patient Model to Predict Specific Treatment Response During the First 24 Hours of Sepsis

Lal, Amos MBBS; Li, Guangxi MD; Cubro, Edin MSc, MBA; Chalmers, Sarah MD; Li, Heyi MD; Herasevich, Vitaly MD, PhD, FCCM; Dong, Yue MD; Pickering, Brian W. MB, BAO, BCh, FFARCSI; Kilickaya, Oguz MD; Gajic, Ognjen MD, MSc, FCCP, FCCM

This group of investigators developed a digital twin model of critically ill patient with a causal artificial intelligence approach using directed acyclic graphs (DAGs)(which can model various different kinds of information). This methodology was used to predict the response to a myriad of treatment strategies during the first 24 hours of sepsis. A hybrid approach of agent-based modeling and simulation, discrete event simulation, and Bayesian network was designed to simulate the treatment effect both across multiple stages as well as with interrelationships of seven organ systems. The agreement between the observed and expected response ranged from fair to good (kappa coefficient of 0.41-0.65). This causal AI model can mitigate the lack of transparency in the “black-box” AI models that involve complex analytics. In short, this digital twin model in the ICU setting is a refreshing causal AI model (compared to associative AI models) that promises significant impact for patient management with an individualized clinical trial strategy.

View the article click here

Anthony Chang, MD, MBA, MPH, MS
Founder, AIMed
Chief Intelligence and Innovation Officer
Medical Director, The Sharon Disney Lund
Medical Intelligence and Innovation Institute (mi3)
Children’s Hospital of Orange County