在电子健康档案数据中识别终末期肾病的算法验证
本研究利用诊断与操作代码开发并验证识别终末期肾病(ESKD)的算法,开发队列来自Geisinger健康系统559,615例患者,验证队列来自纽约大学朗格尼健康系统767,186例患者。算法在开发队列敏感度73.9%、特异度99.83%,验证队列敏感度71.8%、特异度99.95%,特异度高而敏感度中等。
要点
- 算法在Geisinger队列识别ESKD的敏感度73.9%、特异度99.83%、阳性预测值79.9%、阴性预测值99.76%。
- 算法在纽约大学朗格尼队列敏感度71.8%、特异度99.95%、阳性预测值91.6%、阴性预测值99.79%。
- 算法与USRDS记录的ESKD发病日期中位差异在Geisinger为-3天,在纽约大学朗格尼为0天。
核心结论
- 联合诊断与操作代码的算法识别ESKD具有高特异度和中等敏感度,可作为基于电子健康档案数据研究的工具。
研究设计与样本
- 诊断算法研究,开发队列纳入Geisinger健康系统1996年1月至2018年6月治疗的559,615例患者,验证队列纳入纽约大学朗格尼健康系统2018年1月至2020年12月治疗的767,186例患者。
主要发现
- Geisinger队列中算法识别4,766例(0.85%)ESKD,USRDS报告5,155例(0.92%)ESKD。
- Geisinger队列算法敏感度73.9%(95%CI 72.7-75.1%)、特异度99.83%(95%CI 99.82-99.84%)、阳性预测值79.9%(95%CI 78.9-81.0%)、阴性预测值99.76%(95%CI 99.75-99.77%)。
- 纽约大学朗格尼队列算法敏感度71.8%(95%CI 70.7-73.0%)、特异度99.95%(95%CI 99.95-99.96%)、阳性预测值91.6%(95%CI 90.8-92.4%)、阴性预测值99.79%(95%CI 99.78-99.80%)。
- 算法与USRDS的ESKD发病日期中位差异在Geisinger为-3天(IQR -21至83天),在纽约大学朗格尼为0天(IQR -12至69天)。
关键数值与推荐
- 算法与USRDS数据中的名义金标准标识进行比较,以敏感度、特异度、阳性预测值和阴性预测值评价算法性能。
- 在部分病例中比较算法与USRDS的ESKD发病日期。
- 结局定义为ESKD,包括维持性透析、既往接受肾移植或肾移植手术。
临床怎么用
- 原文未据此给出推荐。
局限性
- 仅使用结构化电子健康档案数据。
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
RATIONALE & OBJECTIVES: Accurate ascertainment of end-stage kidney disease (ESKD) in electronic health records (EHRs) data is important for much epidemiological research. This study developed and validated an algorithm using diagnosis and procedure codes to identify patients with ESKD (treated with maintenance dialysis or kidney transplantation) in EHR data. STUDY DESIGN: Study of diagnostic algorithms. SETTING & PARTICIPANTS: The development cohort included 559,615 patients treated at the Geisinger Health System (January 1996-June 2018). The validation cohort included 767,186 patients treated at New York University Langone Health System (January 2018 to December 2020). ALGORITHMS COMPARED: The algorithm used diagnosis and procedure codes compared with a nominal gold standard designation within the United States Renal Data System (USRDS) data. The performance of the algorithm was characterized by sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The dates of incident ESKD between the algorithm and USRDS were compared in a subset of cases. OUTCOME: ESKD (maintenance dialysis, prior recipient of a kidney transplant, or kidney transplantation surgery) cases. RESULTS: In Geisinger, we developed an ESKD algorithm that identified 4,766 (0.85%) ESKD cases; there were 5,155 (0.92%) ESKD cases reported by the USRDS. The sensitivity, specificity, PPV, and NPV of the algorithm were 73.9% (95% CI, 72.7-75.1%), 99.83% (99.82-99.84%), 79.9% (78.9-81.0%), and 99.76% (99.75-99.77%), respectively. When applying the algorithm to New York University Langone Health System data, the sensitivity, specificity, PPV, and NPV were 71.8% (95% CI, 70.7-73.0%), 99.95% (99.95-99.96%), 91.6% (90.8-92.4%), and 99.79 (99.78-99.80%), respectively. The median difference between dates of incident ESKD (algorithms minus USRDS) was-3 (IQR, -21 to 83) days for Geisinger and 0 (IQR, -12 to 69) days for New York University Langone Health. LIMITATIONS: Use of structured EHRs data only. CONCLUSIONS: Algorithms combining diagnosis and procedure codes show high specificity and modest sensitivity for identifying patients with ESKD, providing a research tool to inform future EHRs-based studies. PLAIN-LANGUAGE SUMMARY: Although electronic health records (EHRs) data holds great promise for advancing kidney research, little work has been done to accurately identify ESKD cases in these data. This study developed and validated an algorithm using diagnosis and procedure codes to identify ESKD in EHRs. Our findings showed that the algorithm performed consistently in 2 different health systems, demonstrating high specificity and negative predictive values but lower sensitivity and positive predictive value. This algorithm may inform future ESKD research using EHR data.
引用
- Gao C, Xu Y, Mehta S, et al. Validation of an Algorithm to Identify End-Stage Kidney Disease in Electronic Health Records Data. Am J Kidney Dis. 2025. PMID: 40381931. doi:10.1053/j.ajkd.2025.03.021
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。