在异构的成年选择性手术患者群体中,为术后呼吸衰竭提供最少绝对收缩和选择操作员衍生的预测模型
Jacqueline C Stocking1, Sandra L Taylor1, Sili Fan1
1Division of Pulmonary, Critical Care and Sleep Medicine (J. C. S., C. E. S., J. Y. A., N. J. K., and T. E. A.), Department of Internal Medicine, the Department of Public Health Sciences (S. L. T. and S. F.), the Outcomes Research Group (G. H. U.), Department of Surgery, University of California Davis, Sacramento, the Department of Anesthesiology and Perioperative Medicine (T. W. and M. C.), University of California Los Angeles, the Department of Medicine (M. K. O.), University of California Los Angeles, the VA Greater Los Angeles Healthcare System (M. K. O.), Los Angeles, the Department of Statistics (C. D.), University of California Davis, Davis, the Department of Anesthesia and Perioperative Care (J. M. A. and M. A. G.), University of California, San Francisco, San Francisco, the Department of Medicine (A. N. A.), University of California Irvine, Irvine, the Department of Surgery (R. A. M. and L. G.), University of California San Diego, San Diego, the College of Medicine (C. B. and J. M.), University of Arizona Health Sciences, the Department of Biomedical Engineering (V. S.), College of Engineering, the Center for Health Outcomes and PharmacoEconomic Research (I. A.), University of Arizona, Tucson, AZ, and The University of Florida-Scripps Research Institute (J. G. N. G.), Jupiter, FL.
开发了一个预测模型,以识别患有术后呼吸衰竭 (PRF) 高风险的患者. 该模型使用电子健康记录数据来改进患者护理规划和资源分配.
科学领域:
- 医疗信息学 医疗信息学
- 临床预测模型临床预测模型
- 医疗保健服务研究 医疗服务研究
背景情况:
- 手术后呼吸衰竭 (PRF) 显著增加医疗保健成本,并对患者的结果产生负面影响.
- 准确的PRF预测对于优化术后护理,资源管理和以患者为中心的决策至关重要.
研究的目的:
- 开发和验证一个预测模型,用于识别患有PRF高风险的患者.
- 评估使用电子健康记录 (EHR) 数据用于PRF风险预测的可行性.
主要方法:
- 进行了对23999名接受选择性手术的成年患者的EHR数据的回顾性分析.
- 使用最少绝对收缩和选择操作员 (LASSO) 方法精制的后勤回归模型被开发用于预测PRF (定义为机械通风>48小时).
- 模型性能使用AUC-ROC,AUC-PR,灵敏度,特异性,PPV,NPV和Brier分数进行评估.
主要成果:
- 最终的18个变量模型确定了关键预测因素,包括手术类型,支付人,ASA类,BMI,麻醉持续时间,手术内流体平衡,通风参数和血管压缩剂使用.
- 该模型实现了0.835的乐观度校正的AUC-ROC和0.156.15的AUC-PR.
- 在研究队列中,PRF的发病率为0.94% (225名患者).
结论:
- 一个PRF预测模型可以有效地使用易于获得的EHR和手术内数据开发.
- 该模型表现出良好的预测性能,并且可用于多中心验证和临床实施.
- 该工具可以支持术后护理入院和治疗策略的临床决策.
相关概念视频
Acute Respiratory Failure-V
Ensure that patients are monitored continuously for their response to therapy, including changes in...
Acute Respiratory Failure-II
The underlying physiological abnormalities that contribute to hypoxemic respiratory failure include:


