Compliance and factuality of large language models for clinical research document generation
LLMs offer value in clinical research document generation but must be adapted and rigorously evaluated for high-stakes applications. Our benchmark and framework provide a foundation for improving and assessing LLM-generated outputs in compliance-critical domains.
Authors
- Brandon Theodorou (ORCID: https://orcid.org/0000-0002-8926-2378)
- Zifeng Wang (ORCID: https://orcid.org/0000-0002-3026-9970)
- Seung Won (ORCID: https://orcid.org/0000-0003-1764-7890)
- Riyaaz Uddien Shaik (ORCID: https://orcid.org/0000-0001-8581-3374)
- Jimeng Sun (ORCID: https://orcid.org/0000-0003-1512-6426)
- Benjamin Danek (ORCID: https://orcid.org/0009-0002-8804-0826)
- Shivashankar Thati
- Junyi Gao (ORCID: https://orcid.org/0000-0002-4951-8682)
Institutions
- University of Illinois Urbana-Champaign (US)
- Seoul National University Bundang Hospital (KR)
- Keiju Medical Center (JP)
- The Queen's Medical Research Institute (GB)
- University of Edinburgh (GB)
Publication Details
- Journal
- Journal of the American Medical Informatics Association
- Published
- 2025-09-26
- DOI
- https://doi.org/10.1093/jamia/ocaf174
- Citations
- 2
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 1.31