@inproceedings{wang-etal-2025-towards-adapting,
title = "Towards Adapting Open-Source Large Language Models for Expert-Level Clinical Note Generation",
author = "Wang, Hanyin and
Gao, Chufan and
Liu, Bolun and
Xu, Qiping and
Hussein, Guleid and
Labban, Mohamad El and
Iheasirim, Kingsley and
Korsapati, Hariprasad Reddy and
Outcalt, Chuck and
Sun, Jimeng",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.626/",
doi = "10.18653/v1/2025.findings-acl.626",
pages = "12084--12117",
ISBN = "979-8-89176-256-5",
abstract = "Proprietary Large Language Models (LLMs) such as GPT-4 and Gemini have demonstrated promising capabilities in clinical text summarization tasks. However, due to patient data privacy concerns and computational costs, many healthcare providers prefer using small, locally-hosted models over external generic LLMs. This study presents a comprehensive domain- and task-specific adaptation process for the open-source LLaMA-2 13 billion parameter model, enabling it to generate high-quality clinical notes from outpatient patient-doctor dialogues. Our process incorporates continued pre-training, supervised fine-tuning, and reinforcement learning from both AI and human feedback. We introduced a new approach, DistillDirect, for performing on-policy reinforcement learning with Gemini 1.0 Pro as the teacher model. Our resulting model, LLaMA-Clinic, can generate clinical notes comparable in quality to those authored by physicians. In a blinded physician reader study, the majority (92.8{\%}) of individual evaluations rated the notes generated by LLaMA-Clinic as ``acceptable'' or higher across all three criteria: real-world readiness, completeness, and accuracy. In the more challenging ``Assessment and Plan'' section, LLaMA-Clinic received the same score as the notes authored by physicians. We highlight key considerations for future clinical note-generation tasks, emphasizing the importance of pre-defining a best-practice note format, rather than relying on LLMs to determine this for clinical practice."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2025-towards-adapting">
<titleInfo>
<title>Towards Adapting Open-Source Large Language Models for Expert-Level Clinical Note Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hanyin</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chufan</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bolun</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qiping</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guleid</namePart>
<namePart type="family">Hussein</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohamad</namePart>
<namePart type="given">El</namePart>
<namePart type="family">Labban</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kingsley</namePart>
<namePart type="family">Iheasirim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hariprasad</namePart>
<namePart type="given">Reddy</namePart>
<namePart type="family">Korsapati</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chuck</namePart>
<namePart type="family">Outcalt</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jimeng</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Proprietary Large Language Models (LLMs) such as GPT-4 and Gemini have demonstrated promising capabilities in clinical text summarization tasks. However, due to patient data privacy concerns and computational costs, many healthcare providers prefer using small, locally-hosted models over external generic LLMs. This study presents a comprehensive domain- and task-specific adaptation process for the open-source LLaMA-2 13 billion parameter model, enabling it to generate high-quality clinical notes from outpatient patient-doctor dialogues. Our process incorporates continued pre-training, supervised fine-tuning, and reinforcement learning from both AI and human feedback. We introduced a new approach, DistillDirect, for performing on-policy reinforcement learning with Gemini 1.0 Pro as the teacher model. Our resulting model, LLaMA-Clinic, can generate clinical notes comparable in quality to those authored by physicians. In a blinded physician reader study, the majority (92.8%) of individual evaluations rated the notes generated by LLaMA-Clinic as “acceptable” or higher across all three criteria: real-world readiness, completeness, and accuracy. In the more challenging “Assessment and Plan” section, LLaMA-Clinic received the same score as the notes authored by physicians. We highlight key considerations for future clinical note-generation tasks, emphasizing the importance of pre-defining a best-practice note format, rather than relying on LLMs to determine this for clinical practice.</abstract>
<identifier type="citekey">wang-etal-2025-towards-adapting</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.626</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.626/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>12084</start>
<end>12117</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Towards Adapting Open-Source Large Language Models for Expert-Level Clinical Note Generation
%A Wang, Hanyin
%A Gao, Chufan
%A Liu, Bolun
%A Xu, Qiping
%A Hussein, Guleid
%A Labban, Mohamad El
%A Iheasirim, Kingsley
%A Korsapati, Hariprasad Reddy
%A Outcalt, Chuck
%A Sun, Jimeng
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F wang-etal-2025-towards-adapting
%X Proprietary Large Language Models (LLMs) such as GPT-4 and Gemini have demonstrated promising capabilities in clinical text summarization tasks. However, due to patient data privacy concerns and computational costs, many healthcare providers prefer using small, locally-hosted models over external generic LLMs. This study presents a comprehensive domain- and task-specific adaptation process for the open-source LLaMA-2 13 billion parameter model, enabling it to generate high-quality clinical notes from outpatient patient-doctor dialogues. Our process incorporates continued pre-training, supervised fine-tuning, and reinforcement learning from both AI and human feedback. We introduced a new approach, DistillDirect, for performing on-policy reinforcement learning with Gemini 1.0 Pro as the teacher model. Our resulting model, LLaMA-Clinic, can generate clinical notes comparable in quality to those authored by physicians. In a blinded physician reader study, the majority (92.8%) of individual evaluations rated the notes generated by LLaMA-Clinic as “acceptable” or higher across all three criteria: real-world readiness, completeness, and accuracy. In the more challenging “Assessment and Plan” section, LLaMA-Clinic received the same score as the notes authored by physicians. We highlight key considerations for future clinical note-generation tasks, emphasizing the importance of pre-defining a best-practice note format, rather than relying on LLMs to determine this for clinical practice.
%R 10.18653/v1/2025.findings-acl.626
%U https://aclanthology.org/2025.findings-acl.626/
%U https://doi.org/10.18653/v1/2025.findings-acl.626
%P 12084-12117
Markdown (Informal)
[Towards Adapting Open-Source Large Language Models for Expert-Level Clinical Note Generation](https://aclanthology.org/2025.findings-acl.626/) (Wang et al., Findings 2025)
ACL
- Hanyin Wang, Chufan Gao, Bolun Liu, Qiping Xu, Guleid Hussein, Mohamad El Labban, Kingsley Iheasirim, Hariprasad Reddy Korsapati, Chuck Outcalt, and Jimeng Sun. 2025. Towards Adapting Open-Source Large Language Models for Expert-Level Clinical Note Generation. In Findings of the Association for Computational Linguistics: ACL 2025, pages 12084–12117, Vienna, Austria. Association for Computational Linguistics.