Validation of 13 102 International Classification of Diseases, Tenth Revision, Clinical Modification codes using a large language model-based system
In this large-scale evaluation, ICD-10-CM codes exhibited generally high accuracy, though variability existed by position and condition type. A validated LLM system performed comparably to physician review and offers a scalable means to improve coding accuracy. These findings support the potential for integrating LLM-based auditing into routine workflows to strengthen the quality of administrative and research data.
Authors
- Daryl Ramai
- Dilhana Badurdeen
- J. Wang
- Induja R. Nimma
- Rex K. Siu (ORCID: https://orcid.org/0009-0003-4389-5967)
- Vivek Kumbhari (ORCID: https://orcid.org/0000-0002-7772-9006)
- Yan Yan (ORCID: https://orcid.org/0000-0001-6142-4251)
- Yuzhou Wang (ORCID: https://orcid.org/0000-0001-6338-3799)
- Thomas Savage (ORCID: https://orcid.org/0000-0003-4828-5802)
- Yilin Song (ORCID: https://orcid.org/0009-0007-6027-6235)
- Yuting Huang
- Yiming Wang
- Zhichen Li
- Cui Tao
Institutions
- University of Maryland, Baltimore (US)
- Brigham and Women's Hospital (US)
- Hospital of the University of Pennsylvania (US)
- Jacksonville College (US)
- University of Maryland Medical Center Midtown Campus (US)
- St. Johns Center for Clinical Research (US)
- Neurologique Foundation (US)
- Mayo Clinic in Florida (US)
- Xi’an Jiaotong-Liverpool University (CN)
- University of Pennsylvania (US)
Publication Details
- Journal
- Journal of the American Medical Informatics Association
- Published
- 2026-01-14
- DOI
- https://doi.org/10.1093/jamia/ocag008
- Primary Topic
- Medical Coding and Health Information
- Type
- article
- Field-Weighted Citation Impact
- 0.00