利用结构化电子病历数据与机器学习构建术后感染检测模型并验证
本研究基于5家大型医院2013至2019年的30,639例手术,利用结构化电子病历数据开发并验证简约、可解释的术后感染监测模型。模型仅含4至8个变量,在保留测试集中AUC>0.91、特异度>81%、敏感度>87%、阴性预测值>99%、阳性预测值10%至15%。结论认为可用电子病历数据和简单线性回归模型对所有手术进行高精度感染监测。
要点
- 术后感染占所有术后并发症一半以上,自动监测可覆盖全部手术。
- 基于电子病历的简约模型在保留数据中AUC>0.91,阴性预测值>99%。
- 模型仅需4至8个变量,便于临床部署与解释。
核心结论
- 利用结构化电子病历数据和简单线性回归模型可对所有手术实施高精度的术后感染监测与报告。
研究设计与样本
- 回顾性研究,纳入2013至2019年5家大型医院的30,639例独特手术,将结构化电子病历数据与美国外科医师学会国家手术质量改进计划(NSQIP)的术后结局数据链接。
- 采用knockoff filter(一种控制I型错误的惩罚回归技术)进行变量选择,并在按时间划分的保留数据集中验证模型。
主要发现
- 7%的患者发生至少一种术后感染。
- 所选模型包含4至8个变量,在保留测试集中AUC>0.91、特异度>81%、敏感度>87%、阴性预测值>99%、阳性预测值10%至15%。
关键数值与推荐
- 感染性并发症包括手术部位感染、尿路感染、脓毒症和肺炎,均限于术后30天内。
- 预测变量来自电子病历中的诊断、操作和药物。
临床怎么用
- 可用于替代部分人工病历审查,实现对所有手术的术后感染自动化监测与报告。
- 模型变量少且可解释,便于在电子病历系统中集成和推广。
局限性
- 原文未报告模型在外部人群中的验证结果。
- 阳性预测值仅为10%至15%,可能需结合其他数据确认感染事件。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Postoperative infections constitute more than half of all postoperative complications. Surveillance of these complications is primarily done through manual chart review, which is time consuming, expensive, and typically only covers 10% to 15% of all operations. Automated surveillance would permit the timely evaluation of and reporting of all operations. METHODS: The goal of this study was to develop and validate parsimonious, interpretable models for conducting surveillance of postoperative infections using structured electronic health records data. This was a retrospective study using 30,639 unique operations from 5 major hospitals between 2013 and 2019. Structured electronic health records data were linked to postoperative outcomes data from the American College of Surgeons National Surgical Quality Improvement Program. Predictors from the electronic health records included diagnoses, procedures, and medications. Infectious complications included surgical site infection, urinary tract infection, sepsis, and pneumonia within 30 days of surgery. The knockoff filter, a penalized regression technique that controls type I error, was applied for variable selection. Models were validated in a chronological held-out dataset. RESULTS: Seven percent of patients experienced at least one type of postoperative infection. Models selected contained between 4 and 8 variables and achieved >0.91 area under the receiver operating characteristic curve, >81% specificity, >87% sensitivity, >99% negative predictive value, and 10% to 15% positive predictive value in a held-out test dataset. CONCLUSION: Surveillance and reporting of postoperative infection rates can be implemented for all operations with high accuracy using electronic health records data and simple linear regression models.
引用
- Colborn KL, Zhuang Y, Dyas AR, et al. Development and validation of models for detection of postoperative infections using structured electronic health records data and machine learning. Surgery. 2022. PMID: 36470694. doi:10.1016/j.surg.2022.10.026
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。