Integrating large language models with human expertise for disease detection in electronic health records
The proposed LLM-based pipeline demonstrated reasonable accuracy and high efficiency in disease detection for multiple conditions. Human expert knowledge can be integrated into the pipeline to guide EHR note analysis without manually curated labels. The method could enable comprehensive real-time disease surveillance using EHRs.
Authors
- Kiarash Riazi (ORCID: https://orcid.org/0000-0002-2291-7405)
- Cheligeer Cheligeer (ORCID: https://orcid.org/0009-0007-0810-3011)
- Elliot A. Martin (ORCID: https://orcid.org/0000-0001-5127-4333)
- Hude Quan (ORCID: https://orcid.org/0000-0002-7848-7256)
- Na Li (ORCID: https://orcid.org/0000-0002-0789-0515)
- Seungwon Lee (ORCID: https://orcid.org/0000-0002-6532-5303)
- Jie Pan (ORCID: https://orcid.org/0000-0001-6398-1756)
Institutions
- Alberta Health Services (CA)
- University of Calgary (CA)
- Calgary Laboratory Services (CA)
- Libin Cardiovascular Institute of Alberta (CA)
Publication Details
- Journal
- Computers in Biology and Medicine
- Published
- 2025-04-08
- DOI
- https://doi.org/10.1016/j.compbiomed.2025.110161
- Citations
- 4
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 9.92
Funders
- Canadian Institutes of Health Research