New Delhi: US researchers have developed a new and highly effective generative artificial intelligence (AI) tool to rapidly detect exposure to the H5N1 virus — the cause of avian influenza or “bird flu” — and strengthen national surveillance as the virus continues to spread in other animals.
The study, published in Clinical Infectious Diseases, revealed that the AI tool quickly scans notes in electronic medical records to identify high-risk patients who may have been exposed to H5N1.
Bridging Public Health Gaps
“This study shows how generative AI can fill a critical gap in our public health infrastructure by detecting high-risk patients that would otherwise go unnoticed,” said corresponding author Katherine E. Goodman, Assistant Professor of Epidemiology and Public Health at the University of Maryland School of Medicine.
“With H5N1 continuing to circulate in animals, our biggest danger nationwide is that we don’t know what we don’t know. 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,” she added.
Low-Cost, Fast, and Scalable
The AI tool requires only 26 minutes and costs just 3 cents per patient, showing potential to create a national network of clinical sentinel sites for emerging infectious disease surveillance.
Using a large language model, the research team analysed 13,494 hospital emergency department visits in 2024 from adult patients in urban, suburban, and rural areas. All had acute respiratory illness or conjunctivitis — symptoms consistent with early H5N1 infections.
Identifying High-Risk Patients
Scanning the emergency department notes, the AI flagged 76 cases mentioning a high-risk exposure to bird flu, such as working as a butcher or at a livestock farm. After review, 14 patients were confirmed to have recent exposure to animals known to carry H5N1, including poultry, wild birds, and livestock.
These patients were not tested specifically for H5N1, so infections were not confirmed, but the model successfully identified potential “needle in a haystack” cases among thousands of patients treated for seasonal flu and other respiratory illnesses.