通讯作者: 邹海欧, haiou5275@163.com

DOI：10.12201/bmr.202406.00001

A Scoping Review of the Application of Large Language Models in Clinical Practice

• 摘要：目的/意义 对大语言模型（LLM）在临床实践中的应用情况进行范围综述,为今后LLM在临床实践中的推广提供参考。方法/过程 明确研究问题后，通过计算机检索PubMed、Embase、万方和CNKI数据库，筛选LLM在临床实践中应用的相关文献，检索时限为建库至2024年4月22日，按照纳入和排除标准筛选文献，并对纳入的文献进行内容提取、汇总和分析。结果/结论 共检索到相关文献4061篇，经筛选后最终纳入32篇文献，分析LLM在临床实践中的应用方向、具体作用、优势、局限性等。研究发现LLM在提供治疗建议、辅助疾病诊断、健康宣教、分析文本影像资料等方面具有应用价值，但在答案正确率、个案化等方面的表现不尽人意。总体来说，LLM在临床实践中展现出显著的潜力，但仍需进一步优化、微调、研究，以把控应用风险，确认适用范围。

  关键词： 大语言模型；人工智能；临床实践；范围综述；ChatGPT; ; ; ; 

   

  Abstract: Purpose/Significance This scoping review aims to examine the applications of Large Language Models (LLMs) in clinical practice comprehensively, providing references for their future implementation in clinical settings. Method/Process The research question was clarified initially, followed by a computerized search in databases such as PubMed, Embase, Wanfang, and CNKI, for literature related to the application of LLMs in clinical practice up until April 22, 2024. Literature was screened according to predefined inclusion and exclusion criteria, with selected studies undergoing content extraction, summarization, and analysis. Result/Conclusion A total of 4061 relevant articles were retrieved, and after screening, 32 articles were finally included to analyze the application directions, specific functions, advantages, limitations, etc. of LLM in clinical practice. It was found that LLM has practical value in providing treatment recommendations, assisting in disease diagnosis, health education, and analyzing text and imaging data. However, performance in terms of answer accuracy and individualization was not entirely satisfactory. Overall, LLM demonstrates significant potential in the field of clinical medical care and nursing, yet further optimization, fine-tuning, and research are needed to control application risks and confirm its scope of applicability.

  Key words: large language models; artificial intelligence; clinical practice; scoping review; ChatGPT

  提交时间：2024-06-07

  版权声明：作者本人独立拥有该论文的版权，预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。
• html

• 图表

• 郑琰莉, 韩福海, 李舒玉, 苏文星. 人工智能大模型在医疗领域的应用现状与前景展望. 2023. doi: 10.12201/bmr.202312.00027

  马傲, 葛小玲. 人工智能大模型在医疗健康领域应用的研究. 2024. doi: 10.12201/bmr.202408.00039

  吕婷钰, 李晓瑛, 刘宇炀, 杜晋华, 李心怡, 罗妍, 唐小利, 任慧玲, 刘辉, 尹浩. 中文医学知识大模型问答语料数据集构建研究. 2024. doi: 10.12201/bmr.202404.00002

  汪亚男, 徐霞艳, 朱月兰. “互联网＋护理服务”质量管理评价指标研究的范围综述. 2024. doi: 10.12201/bmr.202407.00067

  蔡自兴. 智慧医疗的临床应用与技术. 2021. doi: 10.12201/bmr.202012.00013

  司唯, 徐婷, 林家玥, 曹文婷, 朱爱勇. 人工智能技术在老年认知症筛查中的应用现状. 2024. doi: 10.12201/bmr.202407.00048

  车贺宾, 徐洪丽. 构建医学大数据应用实践中临床数据处理流程规范. 2021. doi: 10.12201/bmr.202109.00002

  孙艺铭, 李童童, 段棣飞. 老年人数字健康素养评估工具的范围综述. 2024. doi: 10.12201/bmr.202408.00062

  赵锐, 石秀园, 钟雪然, 任平, 田雪晴, 刘春平, 游茂. 基于卫生技术评估构建医学人工智能技术临床应用评估路径的探讨. 2022. doi: 10.12201/bmr.202203.00005

  方明旺, 郭玲, 黄应德, 苑伟, 高芸艺, 周益, 赵一洋, 帅冰星, 陈相军, 张伟义, 李大江. 探讨ChatGPT在医院管理领域中应用的前景与挑战. 2024. doi: 10.12201/bmr.202406.00003

• 序号	提交日期	编号	操作
  1	2023-12-28	bmr.202406.00001V1	下载

• 公开评论  匿名评论  仅发给作者 提交评论