可解释的基于机器学习的28天死亡预测模型,用于患有急性损伤的老年患者
Yueru Jiao1,2, Zhen Wu2, Yabin Zhang2
1Chinese PLA Medical School, Chinese PLA General Hospital, Beijing, 100853, China.
BMC nephrology
|February 5, 2026
概括
机器学习模型可以预测急性损伤 (AKI) 的老年患者的死亡风险. XGBoost模型实现了高精度,识别了早期干预的关键因素,如机械通风和肌水平.
科学领域:
- 医疗信息学 医疗信息学
- 老年医学 老年医学
- 腎臟病學 (nephrology) 是一種醫學專業.
背景情况:
- 患有急性损伤 (AKI) 的老年患者有高死亡风险.
- 早期预测死亡率对于及时干预和资源分配至关重要.
- 机器学习 (ML) 为开发准确的预测模型提供了潜力.
研究的目的:
- 开发和验证ML模型来预测老年AKI患者的28天死亡率.
- 确定这一群体中死亡率的关键预测因素.
- 通过预测性医疗分析来增强临床决策.
主要方法:
- 来自PLAGH (2008-2018) 的1290名老年AKI患者的回顾性分析.
- 五种ML算法 (L2-logistic,LASSO,XGBoost,RF,MLP) 与过量采样技术的比较.
- 使用接收器操作特征曲线 (AUC) 下面的面积进行性能评估;通过夏普利添加式解释 (SHAP) 进行可解释性.
主要成果:
- 随机过量抽样的XGBoost模型在验证队列中实现了0.8659的AUC.
- 在eICU-CRD数据集中的外部验证得出AUC为0.6317.
- 28天死亡率的最高预测因素包括机械通风,血清肌素峰值,AKI阶段,尿蛋白和血清白蛋白.
结论:
- ML模型,特别是XGBoost,可以有效地预测老年AKI患者的死亡率.
- 预测模型增强了临床决策,并可以改善患者的治疗结果.
- 通过SHAP分析确定的关键因素为有针对性的干预提供了洞察力.
相关概念视频
01:22Acute Kidney Injury I: Introduction
768
Introduction:Acute Kidney Injury (AKI) describes a swift decrease in kidney function occurring over hours to days, characterized by the kidneys' failure to remove waste products from the bloodstream. This leads to dangerous complications like metabolic acidosis, fluid overload, and electrolyte imbalances, such as hyperkalemia, which can cause life-threatening arrhythmias. AKI is common in both hospital and outpatient settings, often triggered by dehydration, sepsis, or exposure to nephrotoxic...
768
01:29Acute Kidney Injury II: Pathophysiology
1.1K
Acute kidney injury (AKI) causes are categorized into three primary categories based on the location of the injury: prerenal, intrarenal (or intrinsic), and postrenal causes. This classification guides clinical management and illustrates how different pathways can impair kidney function.Etiology and Pathophysiology of Acute Kidney Injury1. Prerenal causesEtiology: Prerenal Acute Kidney Injury, the most common type, occurs when reduced blood flow to the kidneys decreases filtration capacity...
1.1K
01:22Acute Kidney Injury VI: Nursing Management
481
Acute Kidney Injury (AKI) results in an inability to maintain fluid, electrolyte, and acid-base balance. Effective nursing management is critical in improving patient outcomes and includes comprehensive patient assessment and targeted interventions.Comprehensive Patient AssessmentA detailed history collection is essential, focusing on any recent infections, nephrotoxic medication use, or chronic conditions such as hypertension and diabetes that may contribute to AKI. During the physical...
481
01:20Acute Kidney Injury V: Interprofessional Care
364
Acute Kidney Injury (AKI) requires a collaborative healthcare approach to restore renal function and prevent complications. Essential management strategies involve monitoring fluid and electrolyte balance, adjusting medications, initiating dialysis when necessary, and providing nutritional support.Fluid and Electrolyte ManagementFluid Monitoring: Regularly monitoring body weight, central venous pressure, and urine output helps detect fluid imbalances early. Patient intake and output are...
364
01:29Acute Kidney Injury III: Clinical Manifestations
961
Acute Kidney Injury (AKI) progresses through distinct clinical phases: the oliguric, diuretic, and recovery phases, each marked by unique manifestations and challenges.Oliguric Phase:The oliguric phase is the initial stage of AKI, typically lasting 10 to 14 days. This phase is marked by a significant reduction in urine output, usually less than 400 mL per day, indicating decreased kidney function. Fluid retention is a prominent feature, leading to symptoms such as edema, hypertension, and...
961
01:30Acute Kidney Injury IV: Diagnostic Studies and Prevention
343
Accurate diagnosis and effective prevention are critical in managing Acute Kidney Injury (AKI), which is linked to high mortality rates ranging from 10% to 80%. Timely recognition of at-risk patients and careful monitoring can significantly reduce the likelihood of kidney damage.Diagnostic Assessments:The diagnostic process starts with a comprehensive medical history to identify prerenal, intrarenal, and postrenal causes.Prerenal causes, such as dehydration, hypotension, or blood loss, should...
343