U of Maryland creates bird flu-detecting AI: Study

College Park-based University of Maryland researchers created an AI tool that can scan electronic medical records and identify patients at high risk of being infected by H5N1 avian flu.

The team used a generative AI large language model to analyze 13,494 adult emergency department visits at the UM Medical System in 2024, according to an Aug. 25 system news release. Patients included in the analysis sought care for acute respiratory illness featuring conjunctivitis or symptoms such as cough, fever and congestion. 

The AI tool identified 76 records that mentioned a risk factor for H5N1, and researchers confirmed 14 patients had been exposed to animals that transmit avian flu. The patients were not tested for the flu, so infections were not confirmed, but the tool effectively found relevant cases of suspected flu for medical professionals to assess.

“Because we are not tracking how many symptomatic patients have potential bird flu exposures, and how many of those patients are being tested, infections could be going undetected,” lead author Katherine Goodman, PhD, said in the release. “It’s vital for healthcare systems to monitor for potential human exposure and to act quickly on that information.”

Read the full study here. 

The post U of Maryland creates bird flu-detecting AI: Study appeared first on Becker’s Hospital Review | Healthcare News & Analysis.