中国食品药品监管Issue(5)：92-101,10.DOI:10.3969/j.issn.1673-5390.2025.05.009
大语言模型在药品不良事件报告自动化处理的应用及思考
The Application of Large Language Models in the Automation of Adverse Event Handling
摘要
Abstract
With the growing use and the increasing variety of pharmaceutical products,accurately and timely monitoring and identifying adverse drug events is critical for detecting and evaluating potential safety signals,enabling timely risk control measures to ensure public health.Traditional manual methods for identifying such events are costly and inefficient,whereas Artificial Intelligence(AI)provides promising new solutions.This article discusses the application of Large Language Models(LLM)in automating adverse event reporting.Compared to traditional Natural Language Processing(NLP)models,LLM demonstrate stronger capabilities in language understanding and generation.They can extract key entities,evaluate causal relationships,and perform text classification in medical narratives,thereby improving monitoring efficiency.Despite challenges such as data privacy,annotation quality,and model interpretability,ongoing optimization and fine-tuning of LLMs are expected to further advance AI applications for drug safety.关键词
自然语言处理/大语言模型/药品安全/不良反应监测/不良事件报告自动化处理Key words
natural language processing/large language model/drug safety/adverse reaction monitoring/automated adverse event reporting分类
信息技术与安全科学引用本文复制引用
许晋,周益,栾国琴,黄天娇,胡骏..大语言模型在药品不良事件报告自动化处理的应用及思考[J].中国食品药品监管,2025,(5):92-101,10.基金项目
2024年中国药品监督管理研究会研究课题(编号13) （编号13）
