通过基于提取的语言模型分类来提高医院出院安排的预测
William R Small1,2, Ryan J Crowley1, Chloe Pariente2
1NYU Grossman School of Medicine, New York, NY USA.
npj health systems
|January 12, 2026
概括
早期识别需要熟练护理设施 (SNFs) 的患者通过一种新方法得到了改进. 这种人工智能方法,即基于提取的语言模型分类 (ELC),总结了漫长的医疗笔记,提高了SNF排放的预测准确性.
科学领域:
- 临床信息学 临床信息学
- 自然语言处理自然语言处理.
- 医疗保健预测分析 预测分析
背景情况:
- 有效的护理过渡规划需要早期识别出院到熟练护理设施 (SNFs) 的患者.
- 预测性临床信息往往分散在漫长的入院史和身体 (H&P) 笔记中.
- 现有的语言模型面临着长长的文件,杂的数据和缺乏透明度的挑战.
研究的目的:
- 开发和评估基于提取的语言模型分类 (ELC) 用于预测SNF排放.
- 评估来自ELC的AI风险快照是否改善了与H&P原始文本相比的语言模型性能.
- 为了减轻语言模型在处理广泛的临床文档时的局限性.
主要方法:
- 开发了ELC以将H&P提炼成结构化数据和简洁的AI风险快照.
- 追溯地比较了使用H&P原始文本,截断的笔记,结构化提取数据和AI风险快照的九种语言模型.
- 使用接收器操作特征曲线 (AUROC) 下面的区域和精度召回曲线 (AUPRC) 下面的区域评估模型性能.
主要成果:
- ELC显著减少了输入数据长度 (AI风险快照中位数为141个令牌,原始H&P中位数为2120个令牌).
- 平均AUROC和AUPRC在使用ELC衍生的预测器时在模型中得到改善.
- 在AI风险快照上微调的Bio+Clinical BERT实现了最高的AUROC0.851.
结论:
- ELC有效地构建和总结H&P笔记,克服语言模型令牌长度限制.
- 人工智能风险快照提高了预测性能和可解释性,与临床评估保持一致.
- ELC代表了一种有前途的方法,可以改善SNF释放预测,并促进护理过渡.
相关概念视频
Discharge Summary Forms
1.2K
The discharge summary is crucial as it enables a smooth transition from a healthcare facility to a patient's home or another care setting. This critical document facilitates seamless continuity of care, ensuring patients receive the necessary support and attention.
Here's a detailed look at the key components and guidelines for preparing a discharge summary:
Here's a detailed look at the key components and guidelines for preparing a discharge summary:
1.2K
Classification of Illness
8.5K
The meaning of illness is individualized to each person who experiences an alteration in health. In contrast, disease is a medical term indicating a pathological change in the structure and function of the body or mind. It is a condition that has specific symptoms and boundaries.
An illness is a response to a disease in which the person's level of functioning is changed compared with a previous level. The general classification of illness includes acute and chronic.
Acute illness is severe...
An illness is a response to a disease in which the person's level of functioning is changed compared with a previous level. The general classification of illness includes acute and chronic.
Acute illness is severe...
8.5K

