This is unpublished
Dr. Jason Coult
August 21, 2023

Jason Coult taps AI to improve cardiac arrest survival

Research by Dr. Jason Coult uses machine learning algorithms to improve resuscitation efforts during cardiac arrest.
Scroll for more
arrow icon
Back to top

Research assistant professor Dr. Jason Coult (General Internal Medicine) is a biomedical engineer focused on improving cardiac arrest survival through clinically-informed development of resuscitation technology. Earlier this spring, Coult completed a postdoctoral fellowship with his current supervisor Dr. Thomas Rea to prototype proof-of-concept machine learning algorithms that evaluate a patient’s prognosis during cardiac arrest.

Coult’s latest algorithm expands on this work by predicting shock-refractory ventricular fibrillation during resuscitation. This collaborative project demonstrates potential to warn rescuers in advance of repeated failed defibrillation attempts, allowing preemptive interventions targeted at improving patient outcome. The method was published in Circulation with collaborators from cardiology (Dr. Peter Kudenchuk), emergency medicine (Drs. Betty Yang and Heemun Kwok), applied math (Dr. J. Nathan Kutz), bioengineering (Dr. Patrick Boyle), and King County Public Health (Jennifer Blackwood).

Coult recently received a joint grant from UW Population Health Initiative and Co-Motion to optimize this algorithm for potential real-world use in defibrillators. The grant seeks to accelerate clinical implementation of the method through technology transfer to defibrillator manufacturers.

Coult’s future research will focus on leveraging the group’s expanding cardiac arrest dataset to build state-of-the-art deep learning models. He believes that emerging data-driven methods can produce “smart” defibrillator algorithms with accuracy sufficient for clinical use, and that inclusive engineering approaches can ensure equitable performance of these algorithms across subgroups that are typically underrepresented in cardiac arrest datasets.

These proposed smart algorithms would continuously analyze the cardiac arrest electrocardiogram during resuscitation to determine a patient’s real-time actionable phenotype. If successful, this approach has the potential to transform resuscitation from a static, one-size-fits-all approach to a dynamic, individualized strategy that leverages real-time understanding of patient physiology to guide treatment and improve cardiac arrest survival.

Coult is a Washington State native who was homeschooled until college. Growing up in a musical family, he thought that he wanted to pursue a career in music and started college as a piano major. It only took a few STEM classes during his undergraduate coursework to pique his interest in bioengineering, ultimately setting him on a course to pursue resuscitation science research. Jason still enjoys playing piano in his free time.

Coult will be presenting about his team’s ongoing work at the Division of Cardiology Grand Rounds on January 5.