Generative AI and unstructured audio data for precision public health
In this study, transcribed videos about personal experiences with COVID-19 were used for variant classification. The o1 LLM was used to summarize the transcripts, excluding references to dates, vaccinations, testing methods, and other variables that were correlated with specific variants but unrelated to changes in the disease. This step was necessary to effectively simulate model deployment in the early days of a pandemic when subtle changes in symptomatology may be the only viable biomarkers of disease mutations. The embedded summaries were used for training a neural network to predict the variant status of the speaker as "Omicron" or "Pre-Omicron", resulting in an AUROC score of 0.823. This was compared to a neural network model trained on binary symptom data, which obtained a lower AUROC score of 0.769. Results of the study illustrated the future value of LLMs and audio data in the design of pandemic management tools for health systems.
Authors
- Hannah Huth (ORCID: https://orcid.org/0000-0002-2856-4794)
- Bradford J. Wood (ORCID: https://orcid.org/0000-0002-4297-0051)
- Ulaş Bağcı (ORCID: https://orcid.org/0000-0001-7379-6829)
- Lindsey Hazen (ORCID: https://orcid.org/0000-0003-4013-8226)
- Anna Christou
- Jocelyne Rivera
- David A. Clifton (ORCID: https://orcid.org/0000-0002-9848-8555)
- Miranda Song (ORCID: https://orcid.org/0000-0003-4448-277X)
- Yaël Bensoussan (ORCID: https://orcid.org/0000-0002-1635-8627)
- Thanh-Luan Le (ORCID: https://orcid.org/0000-0001-9827-1113)
- Adam Landa
- Robert Morhard
- Alec Peltekian
- Ming Li (ORCID: https://orcid.org/0000-0002-1446-4493)
- Hang Nguyen (ORCID: https://orcid.org/0000-0002-1843-5104)
- Ashley Shin (ORCID: https://orcid.org/0000-0002-0531-9040)
- James Anibal
- Jacqueline Brenner
- Veronica Daoud
Institutions
- Northwestern University (US)
- National Institutes of Health (US)
- The George Institute for Global Health (GB)
- University of South Florida (US)
- United States National Library of Medicine (US)
- McCormick (United States) (US)
- Oxford University Clinical Research Unit (VN)
- National Institutes of Health Clinical Center (US)
Publication Details
- Journal
- npj Health Systems
- Published
- 2025-06-02
- DOI
- https://doi.org/10.1038/s44401-025-00022-7
- Citations
- 1
- Primary Topic
- Personal Information Management and User Behavior
- Type
- article
- Field-Weighted Citation Impact
- 4.17
Funders
- Colgate-Palmolive Company
- American Association for Dental, Oral, and Craniofacial Research
- Wellcome Trust
- National Institute for Health and Care Research
- Royal Academy of Engineering
- University of Oxford
- National Institutes of Health
- Genentech
- Engineering and Physical Sciences Research Council
- National Cancer Institute
- National Institute of Allergy and Infectious Diseases