University of Washington (UW) researchers have developed an artificial intelligence (AI) system that uses patient data to predict whether patients are at risk of abnormally low blood oxygen (hypoxia) during surgery.
The Prescience system also provides users with real-world explanations to support and explain its predictions.
In collaboration with physicians, UW's Su-In Lee and colleagues trained Prescience on about 50,000 patient files, so the program could analyze data such as patient age and weight to calculate the likelihood of hypoxemia prior to surgery.
The system also uses real-time data during surgery to predict when patients are in danger of hypoxemia, and a new AI model helps Prescience provide doctors a concise description of the prediction's underlying factors.
In simulations, Prescience guided anesthesiologists through sample scenarios, boosting doctors' ability to diagnose hypoxemia before surgery by 16% and during surgery by 12%; the system’s findings helped doctors successfully differentiate between patients at risk of hypoxemia and those not at risk 80% of the time.
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA