Communications of the ACM
Health Researchers Use AI to Better Predict Hepatitis C Treatment Outcomes
Machine learning also allowed the researchers to distinguish a host of conditions associated with treatment failure, including tobacco and alcohol use, diabetes, high blood pressure, and certain non-prescription medications used to treat acid reflux disea
Credit: University of Florida Health
Algorithms developed by University of Florida Health (UF Health) researchers employ artificial intelligence (AI) to forecast outcomes from treatments for hepatitis C.
Four algorithms were trained on almost 5,000 patient samples from a national hepatitis C registry, and can generate treatment-failure predictions that outperform multivariable logistic regression methods.
Another 1,631 patient samples separately validated the algorithms' efficacy.
Some of the research was executed with HiPerGator, Florida's most powerful supercomputer.
The most accurate algorithm was the gradient boosting machine, which distinguished itself by being able to identify patients at highest risk of treatment failure and segment them into different risk groups. UF Health's Haesuk Park described the algorithm as "the first AI model developed to predict direct-acting antiviral treatment failure," calling it "a good foundation for future research."
From UF Health
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA
No entries found