UVA students developing AI to help spot early sepsis

CHARLOTTESVILLE, Va. (WVIR) - Two undergrad students at the University of Virginia have been working on a project that allows artificial intelligence to spot early sepsis.

Louisa Edwards and Zach Boner are using their knowledge of computer science and mathematics to help medical professionals in the future by bridging two different UVA schools.

“Usually, as undergrads, we wouldn’t have a lot of interaction with the School of Medicine. Usually, that’s mostly med students,” Edwards said.

“Our data is taken from patients who have episodes of a disease called bloodstream infection, which is a precursor to sepsis. The way that that might work in the real world is like doctors and nurses will enter vital signs into the electronic health record, as well as some lab tests if they’re taken, and then we have a model that sits on top of the electronic health record and uses the information that’s entered in to try to make predictions about risk,” Boner said.

They say AI supports doctors in a number of ways.

“A lot of potential power in terms of recognizing patterns in data and be able to alert doctors and kind of guide their decision making, and just give them more information and help them synthesize patterns and data and apply to certain patients,” Edwards said.

“If a patient, for example, complains of shortness of breath, and the CoMET (Continuous Monitoring of Event) score, which shows respiratory instability as one of its factors, if that score is high, then the doctor is probably more concerned about that patient. If that score is low, then maybe they start them on oxygen. They do some kind of simpler interventions,” Boner said.

Do you have a story idea? Send us your news tip here.

Copyright 2023 WVIR. All rights reserved.