Mount Sinai uses federated learning AI technique to predict COVID-19 disease progression

New York City-based Mount Sinai Health System researchers used federated learning, an artificial intelligence technique that protects patients' privacy, to analyze EHRs and better predict how COVID-19 patients will progress, according to a Jan. 27 news release.

The federated learning technique has the potential to create machine learning models that go beyond a single health system's information without compromising patient privacy. Using the model and data from EHRs at five separate Mount Sinai hospitals, the researchers were able to predict mortality in COVID-19 patients.

"Federated learning is gaining traction within the biomedical space as a way for models to learn from many sources without exposing any sensitive patient data," Benjamin Glicksberg, PhD, genetics and genomics assistant professor at the Icahn School of Medicine at Mount Sinai, said in the news release. "In our work, we demonstrate that this strategy can be particularly useful in situations like COVID-19."

Researchers compared the performances of the federated model against ones built using data from each hospital separately, referred to as local models. After training their models on a federated network and testing the data of local models at each hospital, the researchers found the federated models outperformed local models at most hospitals and showed stronger predictive capabilities.

More articles on artificial intelligence:
HHS releases AI strategy, establishes AI council
How UVA Health uses AI in its remote monitoring tech
RCM roles beat out IT as top users of AI in hospitals, report finds


© Copyright ASC COMMUNICATIONS 2021. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.