Extracting language information from clinical notes using large language models

Open Webpage

Lingfei Qian, Na Hong, Yujia Zhou, Qianqian Xie, Ruey-Ling Weng, Pitchaya Chairuengjitjaras, Xinsong Du, John Lian, Gad A. Marshall, Suzanne V. Blackley, John Laurentiev, Yakeel T. Quiroz, Tae Youn Kim, Nicole Adams, Michelle L. Dossett, Li Zhou, Hua Xu

Published: 01 Jan 2026, Last Modified: 04 Nov 2025Int. J. Medical Informatics 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Developed an LLM-based NER pipeline to extract nuanced patient language status from clinical notes.•Cross-site validation (YNHH, MIMIC) showed GPT-4o strong zero-shot accuracy and LLaMA3 robust generalization.•Enables scalable, fine-grained language data extraction to support equitable, language-focused healthcare research.