ICURE:重症监护室 (ICU) 对30天死亡率的风险评估. 开发和评估一个多变量机器学习预测模型,用于瑞典一般ICU的患者
Tobias Siöland1,2, Araz Rawshani3,4,5, Bengt Nellgård1,2
1Department of Anesthesiology and Intensive Care Medicine, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.
Acta anaesthesiologica Scandinavica
|July 22, 2024
概括
一个新的机器学习模型ICURE准确地预测了30天的重症监护室 (ICU) 死亡率. 该模型的性能优于传统的SAPS 3评分,为ICU患者的风险评估提供了更高的准确性.
科学领域:
- 关键护理医学 关键护理医学
- 医疗信息学 医疗信息学
- 医疗保健中的机器学习
背景情况:
- 在入院时估计重症监护室 (ICU) 死亡率有助于患者的护理和资源分配.
- 像SAPS 3和APACHE这样的传统模型在预测准确性方面存在局限性.
- 机器学习 (ML) 为开发先进的死亡预测模型提供了新的潜力.
研究的目的:
- 开发和评估一种基于ML的新型预测模型,用于ICU患者30天死亡率.
- 将新模型的性能与已建立的SAPS 3评分进行比较.
- 为了确定ICU死亡率的关键预测因素.
主要方法:
- 使用瑞典重症监护登记 (2008-2022) 记录了296,344人首次入住ICU.
- 在ICU入院1小时内提取了153个候选预测因素.
- 训练并测试多个ML模型 (例如随机森林,梯度增强,神经网络) 和后勤回归,评估AUC-ROC的性能.
主要成果:
- 最好的ML模型在测试数据上实现了0.884的AUC-ROC.
- 关键预测因素包括年龄,格拉斯哥昏迷表,肌,静脉压和pH值.
- 开发的ICURE模型表现出与SAPS 3相当的性能,变量较少.
结论:
- 与SAPS 3得分相比,ICURE预测模型提供了对30天死亡率的优异预测,用于首次ICU入院.
- ICURE模型为ICU中的风险分层提供了一个更准确,更有效的工具.
- 在临床关怀中,ML方法显著有望改善临床决策.
更多相关视频
相关概念视频
Actuarial Approach
72
The actuarial approach, a statistical method originally developed for life insurance risk assessment, is widely used to calculate survival rates in clinical and population studies. This method accounts for participants lost to follow-up or those who die from causes unrelated to the study, ensuring a more accurate representation of survival probabilities.
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
72
Steps in Outbreak Investigation
121
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:
121


