AI Model Identifies Undetected Bird Flu Exposure Risks in Emergency Departments

AI Tool Revolutionizes Bird Flu Detection in Emergency Departments

Researchers from the University of Maryland School of Medicine have developed a groundbreaking artificial intelligence (AI) tool that can rapidly scan electronic medical records to identify patients at high risk of H5N1 avian influenza, or "bird flu" exposure. This innovative application of generative AI, detailed in a study published in the journal Clinical Infectious Diseases, has the potential to significantly enhance public health surveillance efforts 1

1

2

3

Source: Newswise

Study Methodology and Findings

The research team utilized a generative AI large language model (LLM) to analyze 13,494 emergency department visits across the University of Maryland Medical System (UMMS) in 2024. The AI model focused on adult patients presenting with symptoms consistent with early H5N1 infections, such as acute respiratory illness or conjunctivitis 1

1

2

3

Key findings from the study include:

1. The AI model flagged 76 cases mentioning high-risk exposures for bird flu, such as working as a butcher or on a farm with livestock.
2. After human review, 14 patients were confirmed to have had recent, relevant exposure to animals known to carry H5N1.
3. The AI review process required only 26 minutes of human time and cost just 3 cents per patient note, demonstrating high efficiency and scalability 1

   1

   2

   2

   3

   3

   .

AI Performance and Implications

The LLM (GPT-4 Turbo) showed strong performance in identifying animal exposure mentions:

• 90% positive predictive value
• 98% negative predictive value

These results were based on an evaluation of 10,000 historical emergency department visits from 2022-2023, before bird flu was circulating in U.S. livestock 1

1

2

3

Dr. Katherine E. Goodman, the study's corresponding author, emphasized the critical role this AI tool could play in public health infrastructure: "With H5N1 continuing to circulate in U.S. 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" 1

1

2

3

Related Stories

AI-Powered PandemicLLM Revolutionizes Infectious Disease Forecasting

AI Advances Promise to Transform Pandemic Preparedness and Response

AI Models Predict COVID-19 Severity and Treatment for Hospitalized Patients

AI Tool Predicts Virus Outbreak Hotspots, Enhancing Pandemic Prevention Efforts

Current H5N1 Situation and Future Implications

Source: Medical Xpress

The urgency of this research is underscored by the current H5N1 situation:

• Since early 2024, H5N1 has infected over 1,075 dairy herds across 17 states.
• More than 175 million poultry and wild birds have tested positive during this outbreak period.
• By mid-2025, there were 70 confirmed human infections and one fatality in the U.S., according to the CDC 1

  1

  2

  2

  3

  3

  .

Researchers suggest that large language models could be used prospectively to alert healthcare providers in real-time, prompting more vigilant questioning about potential exposure to infected animals, targeted testing, and infection control measures 1

1

2

3

Future Research and Applications

The research team plans to test the large language model for prospective surveillance and deployment within electronic health records. This could lead to faster real-time identification of high-risk patients, especially critical as the respiratory virus season approaches 1

1

2

3

Dr. Anthony Harris, a study co-author, highlighted the potential broader impact: "This method has the potential to create a national network of clinical sentinel sites for emerging infectious disease surveillance to help us better monitor newly emerging epidemics" 1

1

2

3

As AI continues to evolve, its application in public health surveillance and emergency medicine represents a promising frontier in the fight against emerging infectious diseases.