机器学习与传统的回归分析,用于预测ICU中的流体过载
Andrea Sikora1, Tianyi Zhang2, David J Murphy3
1Department of Clinical and Administrative Pharmacy, University of Georgia College of Pharmacy, 1120 15th Street, HM-118, Augusta, GA, 30912, USA.
Scientific reports
|November 10, 2023
概括
预测重症监护室 (ICU) 液体过载是具有挑战性的. 像XGBoost这样的机器学习模型和传统方法在识别疾病严重程度和药物复杂性等关键预测因素方面表现相似.
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:
- 液体过载是重症监护室 (ICU) 的常见并发症,导致不良结果.
- 预测液体过载是很困难的,而ICU药物使用可能会影响其发展.
- 机器学习 (ML) 对预测复杂临床结果的传统回归有潜在的优势.
研究的目的:
- 为了比较传统回归技术和ML模型的预测性能,用于识别ICU患者的液体过载.
- 确定临床上有意义的流体过载预测因素,包括患者和药物相关因素.
主要方法:
- 成人ICU患者 (停留≥72小时) 的回顾性观察队列研究,具有可用的液体平衡数据.
- 开发和比较传统的物流回归和各种监督ML模型 (例如XGBoost) 来预测流体过载 (≥10%的体重增加).
- 使用接收器操作特征下的面积 (AUROC),正预测值 (PPV) 和负预测值 (NPV) 评估模型性能.
主要成果:
- 在391名患者中,有12.5%的患者出现过多的液体负荷.
- XGBoost ML模型获得了最高的性能 (AUROC 0.78,PPV 0.27,NPV 0.94).
- XGBoost的性能与最优的传统物流回归模型 (AUROC 0.70,PPV 0.20,NPV 0.94) 相当.
- 功能重要性突出了疾病严重程度得分和药物数据作为关键预测指标.
结论:
- 在这个队列中,ML和传统模型都在预测ICU流体过载方面表现出类似的有效性.
- 疾病的基线严重程度和ICU药物治疗方案的复杂性是液体过载的重要预测因素.
- 机器学习方法,特别是XGBoost,对流体过载预测有前途,其特征的重要性提供了临床见解.
更多相关视频
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
8.3K
10:38Observational Study Protocol for Repeated Clinical Examination and Critical Care Ultrasonography Within the Simple Intensive Care Studies
Published on: January 16, 2019
20.2K
相关概念视频
01:22Heart Failure VI: Adjunct Therapies
15
Additional therapies for treating patients with heart failure (HF) may include procedural interventions, supplemental oxygen, the management of sleep disorders, and nutritional therapy.Procedural InterventionsImplantable Cardioverter-Defibrillator: For patients at risk of life-threatening arrhythmias due to severe left ventricular dysfunction, an Implantable Cardioverter-Defibrillator (ICD) can detect and terminate these arrhythmias, preventing sudden cardiac death and improving survival rates.
15
01:30Typical Model Studies
363
Fluid mechanics model studies often utilize scaled-down systems to predict fluid behavior in full-scale environments, such as river flows, dam spillways, and structures interacting with open surfaces. Maintaining Froude number similarity in river models is crucial, as it replicates surface flow features like wave patterns and velocities.
363