A retrieval-augmented generation large language model framework for accurate dementia identification from electronic health records
A RAG-based LLM pipeline without structured ICD codes improved dementia ascertainment from EHR data compared with ICD-based rules and keyword-based filtering. This approach can enhance dementia case identification and support patient care, predictive modeling and risk analysis.
Authors
- Hossein Estiri (ORCID: https://orcid.org/0000-0002-0204-8978)
- Baoren Liu
- Richard Yang (ORCID: https://orcid.org/0009-0008-8283-4952)
- Gad A. Marshall
- Ya-Wen Chuang
- Liqin Wang
- Li Zhou
- Shawn Murphy
Institutions
- Brigham and Women's Hospital (US)
- Harvard University (US)
- University of Illinois Urbana-Champaign (US)
- Massachusetts General Hospital (US)
- Taichung Veterans General Hospital (TW)
Publication Details
- Journal
- medRxiv
- Published
- 2026-01-25
- DOI
- https://doi.org/10.64898/2026.01.24.26344477
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Alzheimer's Association
- National Institutes of Health