基于变压器的深度学习,通过多变量时间序列分析来早期检测重症监护室获得的血液感染
Jiang-Chen Peng1, Jia-Rui Liang2, Ming-Li Zhu1
1Department of Critical Care, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
Digital health
|January 9, 2026
概括
深度学习变压器模型使用时间序列数据准确预测重症监护室 (ICU) 的血液感染 (BSI). 这种先进的模型显示了早期检测和临床决策支持的前景,优于传统方法.
科学领域:
- 医疗信息学 医疗信息学
- 医疗保健中的人工智能
- 临床预测模型临床预测模型
背景情况:
- 血流感染 (BSI) 是重症监护室 (ICU) 中死亡的重要原因之一.
- 传统的机器学习模型在处理高维患者数据中的复杂时间依赖性方面面临挑战.
- 深度学习,特别是变压器模型,为医疗保健中分析时间序列数据提供了一种新的方法.
研究的目的:
- 开发和评估一个深度学习变压器模型,用于预测ICU获得的BSI.
- 通过使用时间序列电子健康记录数据,评估模型在早期发现BSI方面的表现.
- 为了比较变压器模型的有效性与传统的机器学习方法,如CatBoost.
主要方法:
- 利用来自MIMIC IV数据库的电子健康记录,重点关注ICU入院后48小时收集的血液培养.
- 应用合成少数人过量采样技术 (SMOTE) 用于数据集平衡.
- 在24小时内提取每小时的生命体征和实验室测量,在0,12和24小时设置预测窗口.
- 使用歧视 (AUROC,AUPRC) 和校准指标评估模型性能,使用SHAP进行特征重要性分析.
主要成果:
- 包括2408名患者,其中149人 (6.2%) 在ICU获得了BSI.
- 变压器模型在所有预测窗口中显著超过了CatBoost.
- 在24小时预测窗口中,变压器实现了0.918的AUROC和0.915的AUPRC.
- 由SHAP确定的主要早期预测因素包括葡萄糖,二碳酸盐,平均血压,温度和血液尿素.
结论:
- 深度学习变压器模型显示了预测ICU获得的BSI的巨大潜力.
- 该模型利用时间序列数据的能力使其成为一种有价值的临床决策支持工具.
- 通过先进的深度学习技术,早期发现BSI是可行的,这可能会改善患者的治疗结果.
相关概念视频
01:30Pulse rhythm
Pulse rhythm refers to the pattern of pulsations within specific intervals, offering valuable insights into the regularity or irregularity of the heart's beats as observed through the pattern of pulsation within specific intervals. A regular pulse exhibits a consistent heart rate with uniform waveforms and pulsation force, variations of which can be classified as normal, weak, or bounding.
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac muscle...
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac muscle...
01:30Equipments Used To Measure Blood Pressure
Direct Method
This invasive approach involves cannulating a peripheral artery. During each cardiac contraction, pressure generates mechanical motion within the catheter, transmitted through rigid, fluid-filled tubing to a transducer. This transducer converts mechanical motion into electrical signals displayed as waveforms on a monitor. An automatic flushing system prevents blood backflow. Due to the potential risk of unexpected arterial blood loss, this method is primarily used in intensive...
This invasive approach involves cannulating a peripheral artery. During each cardiac contraction, pressure generates mechanical motion within the catheter, transmitted through rigid, fluid-filled tubing to a transducer. This transducer converts mechanical motion into electrical signals displayed as waveforms on a monitor. An automatic flushing system prevents blood backflow. Due to the potential risk of unexpected arterial blood loss, this method is primarily used in intensive...
01:18Steps in Outbreak Investigation
In the ever-evolving field of public health, statistical analysis serves as a cornerstone for understanding and managing disease outbreaks. By leveraging various statistical tools, health professionals can predict potential outbreaks, analyze ongoing situations, and devise effective responses to mitigate impact. For that to happen, there are a few possible stages of the analysis:
01:24Automated Microbial Diagnostics
Automated diagnostic analyzers have transformed clinical microbiology by providing rapid and reliable methods for pathogen identification and antibiotic susceptibility testing. Among these systems, the Vitek 2 is widely used because it automates the traditionally labor-intensive processes of microbial identification (ID) and antibiotic susceptibility testing (AST), delivering standardized and timely results that are essential for effective patient care.Microbial Identification with ID CardsThe...