Compliance and factuality of large language models for clinical research document generation
Journal of the American Medical Informatics Association · 2026
DOI: 10.1093/jamia/ocaf174 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the American Medical Informatics Association, 2026. Verfasst von Zifeng Wang und 7 weiteren Autor:innen. Beteiligte Einrichtung: Keiji AI, Seattle, WA, 98115, United States. Hauptthemen laut MEDLINE: Biomedical Research, Clinical Trials as Topic, Consent Forms, Guideline Adherence. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NSF.
Abstract (Originalsprache)
Objectives
Large language models’ (LLMs’) performance in high-stakes, compliance-driven settings such as drafting clinical research documents remains underexplored. This study aims to build a benchmark and an evaluation framework for assessing LLMs’ compliance and factuality in generating informed consent forms (ICFs) from clinical trial protocols. Materials and Methods We introduce InformBench, a benchmark comprising 900 clinical trial documents, and propose an evaluation framework grounded in regulatory guidelines and site-specific consent templates. We assess LLM performance on transforming trial protocols, often hundreds of pages, into concise, patient-facing ICFs. Additionally, we design InformGen, a retrieval-augmented, human-in-the-loop pipeline aimed at improving generation quality.
Results
Baseline LLMs such as GPT-4o achieved only 70%-80% compliance and exhibited factual errors in 18%-43% of cases. In contrast, InformGen substantially improved outputs, achieving nearly 100% regulatory compliance and over 90% factual accuracy, as validated by 5 domain-expert annotators.
Discussion
The study reveals critical limitations in current LLMs for clinical research document drafting, particularly in regulatory sensitivity and factual grounding. Our results highlight the need for domain-specific benchmarks and structured evaluations to support safe deployment in real-world clinical research workflows.
Conclusion
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.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Keiji AI, Seattle, WA, 98115, United States.
- Centre for Medical Informatics, Usher Institute, University of Edinburgh, Edinburgh, EH3 9DR, United Kingdom.
- Carle Illinois College of Medicine, University of Illinois Urbana-Champaign, Urbana, IL, 61801, United States.
- Siebel School of Computing and Data Science, University of Illinois Urbana-Champaign, Urbana, IL, 61801, United States.
- Medical Research Collaborating Center, Seoul National University Bundang Hospital, Gyeonggi, 13620, Republic of Korea.
ORCID-Profile: Zifeng Wang, Junyi Gao, Jimeng Sun
Förderung
- NSF
Zitieren
APA
Wang, Z., Gao, J., Danek, B., Theodorou, B., Shaik, R., Thati, S., Won, S., & Sun, J. (2026). Compliance and factuality of large language models for clinical research document generation. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocaf174
Vancouver
Wang Z, Gao J, Danek B, Theodorou B, Shaik R, Thati S, et al. Compliance and factuality of large language models for clinical research document generation. Journal of the American Medical Informatics Association. 2026. doi:10.1093/jamia/ocaf174
BibTeX
@article{Wang2026,
author = {Wang, Zifeng and Gao, Junyi and Danek, Benjamin and Theodorou, Brandon and Shaik, Ruba and Thati, Shivashankar and Won, Seunghyun and Sun, Jimeng},
title = {Compliance and factuality of large language models for clinical research document generation},
journal = {Journal of the American Medical Informatics Association},
year = {2026},
doi = {10.1093/jamia/ocaf174}
}
EUVIMED