Automating clinical phenotyping using natural language processing
To our knowledge, this is the first study to explore computable phenotyping algorithms based on clinical narrative text for these complex tasks, where prior inter-annotator agreements ranged from 0.54 to 0.98. There is no statistical evidence for a difference to the performance of human experts on this task. Our findings underline the potential of LLMs for computable phenotyping and may support large-scale cohort analyses from electronic health records and streamline chart review processes in the future.
Authors
- Bernhard Y. Renard (ORCID: https://orcid.org/0000-0003-4589-9809)
- Florian Borchert (ORCID: https://orcid.org/0000-0003-1079-6500)
- Jellyana Peraza (ORCID: https://orcid.org/0000-0002-3732-6683)
- Julian Hugo (ORCID: https://orcid.org/0000-0003-3355-1071)
- Linea Schmidt (ORCID: https://orcid.org/0009-0006-3014-4058)
- Susanne Ibing
- Allison A. Marshall
- Judy H. Cho
- Erwin P. Böttinger
Institutions
- Mount Sinai Health System (US)
- Hasso Plattner Institute (DE)
- University of Potsdam (DE)
- Icahn School of Medicine at Mount Sinai (US)
Publication Details
- Journal
- Communications Medicine
- Published
- 2026-01-14
- DOI
- https://doi.org/10.1038/s43856-025-01337-0
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Joachim Herz Stiftung
- Georgia Clinical and Translational Science Alliance
- Hasso Plattner Foundation
- National Institutes of Health
- National Center for Advancing Translational Sciences