使用大型语言模型进行产后出血的零射击可解释的表型化
Emily Alsentzer1, Matthew J Rasmussen2, Romy Fontoura2
1Division of General Internal Medicine and Primary Care, Brigham and Women's Hospital, Boston, MA, USA.
NPJ digital medicine
|November 30, 2023
概括
大型语言模型 (LLM) 可以使用临床笔记准确地表型产后出血 (PPH) 患者,识别比当前方法更多的病例,而不需要注释数据.
科学领域:
- 医疗信息学 医疗信息学
- 计算语言学 计算语言学
- 临床的表型化 临床的表型化
背景情况:
- 准确的患者表型在医学中至关重要,但往往受到广泛注释数据的需求的限制.
- 大型语言模型 (LLM) 显示出适应新任务的潜力,只需很少或没有特定任务的培训.
研究的目的:
- 评估公开可用的LLM,Flan-T5的性能,用于使用电子健康记录 (EHR) 释放笔记来表型产后出血 (PPH).
- 评估LLM识别可解释PPH亚型的细粒度概念的能力.
主要方法:
- 使用Flan-T5,一个大型语言模型,分析了271,081个EPH释放笔记用于PPH表型化.
- 提取了与PPH相关的24个细粒度概念,并将LLM性能与患者识别和亚型分析的标准索赔代码进行了比较.
主要成果:
- 在PPH表型化中,Flan-T5实现了高保真性 (PPV为0.95),比索赔代码识别的患者多47%.
- 该LLM管道在PPH亚型中表现出卓越的性能,特别是在子宫,异常胎盘和产科创伤方面.
- 该方法允许可解释的表型和高效的算法更新,随着临床指南的发展.
结论:
- 像Flan-T5这样的LLM提供了一种强大,快速和可解释的方法,用于使用EHR笔记进行临床表型化.
- 这种基于LLM的方法与传统的基于索赔的方法相比,显著改善了PPH识别和亚型.
- 没有手动注释数据的表型能力为临床研究和精准医学开辟了新的途径.
更多相关视频
03:14Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
594
05:47Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems
Published on: June 13, 2025
232
相关概念视频
01:07Teratogenicity
The ability of a drug to produce structural deformations and functional abnormalities in the developing embryo or the fetus is called teratogenicity, and the drug producing this effect is known as a teratogen. Teratogenic effects include stillbirth, miscarriage, intrauterine growth restriction, and neurocognitive delay. A teratogen may affect the embryo at different stages of development, which is important in determining the type and extent of the damage. During blastocyst formation, the early...
01:25Modeling with Differential Equations
Population dynamics can be described mathematically by considering the population size P(t) as a function of time. The rate of change of the population is then represented by the derivative of P(t). A simple assumption is that the rate of growth is proportional to the size of the population itself. This leads to an exponential growth model, where the population increases rapidly without bound. While this is a useful first approximation, it does not reflect realistic long-term...