ARPA-H awards UVM up to $38M to build AI digital twins of ICU patients
The federal Advanced Research Projects Agency for Health awarded the University of Vermont up to $38 million, the largest research award in the school's history, to build AI models that create a digital "twin" of a critically ill patient's immune and inflammatory state from blood samples and physiological monitoring. Led by UVM trauma surgeon Dr. Gary An under ARPA-H's CIRCLE program, the team will use the twin to test candidate treatments virtually before giving them to the patient, with data collection sites at Wake Forest, WashU Medicine and the University of Alabama at Birmingham. The five-year award is structured in milestones; researchers must show by year three that the digital twin can accurately replicate a patient's immune response and predict outcomes to unlock the full funding, and the team believes success could cut ICU length of stay by at least 25%.
Why it's notable: It is a large, concrete federal bet on using AI to personalize ICU treatment decisions rather than just document them, and a positive result could change how intensivists choose therapies for the sickest patients years before any product reaches the bedside.