基于电子病历的重症老年患者急性肾损伤连续风险预测研究

doi:10.3969/j.issn.1006-7795.2022.04.014

首都医科大学学报 ›› 2022, Vol. 43 ›› Issue (4): 600-609.doi: 10.3969/j.issn.1006-7795.2022.04.014

• 医学信息学:应用和发展 • 上一篇下一篇

基于电子病历的重症老年患者急性肾损伤连续风险预测研究

邬金鸣¹, 孙海霞², 王嘉阳², 钱庆^3*

1.中国医学科学院/北京协和医学院医学信息研究所医学数据共享研究室,北京 100020;
2.中国医学科学院/北京协和医学院医学信息研究所医学智能计算研究室,北京 100020;
3.中国医学科学院/北京协和医学院医学信息研究所,北京 100020

收稿日期:2022-03-18 出版日期:2022-08-21 发布日期:2022-10-28
基金资助:
医学知识管理与智能化知识服务关键技术研究(2021-I2M-1-056)。

Continuous risk prediction of acute kidney injury in elderly critically-ill patients based on electronic medical records

Wu Jinming¹, Sun Haixia², Wang Jiayang², Qian Qing^3*

1. Medical Data Sharing Division, Institute of Medical Information/Medical Library, Chinese Academy of Medical Science & Peking Union Medical College, Beijing 100020, China;
2. Medical Intelligent Computing Division, Institute of Medical Information/Medical Library, Chinese Academy of Medical Science & Peking Union Medical College, Beijing 100020, China;
3. Institute of Medical Information/Medical Library, Chinese Academy of Medical Science & Peking Union Medical College, Beijing 100020, China

Received:2022-03-18 Online:2022-08-21 Published:2022-10-28
Contact: *E-mail:qian.qing@imicams.ac.cn
Supported by:
This study was supported by Key Techniques of Medical Knowledge Management and Intelligent Service(2021-I2M-1-056).

摘要/Abstract

摘要： 目的探索重症老年患者(≥60岁)急性肾损伤早期连续风险预测的可行性,促进机器学习在临床决策支持中的应用。具体实现以6 h为单位连续预测重症老年患者在未来48 h的急性肾损伤发病风险,并探索可实现何种程度的早期预测,以及比较当前数据和累积数据的预测效果。方法基于重症监护医学信息数据库(Medical Information Mart for Intensive Care,MIMIC)-Ⅲ,应用逻辑回归、支持向量机、随机森林和轻量梯度提升机(light gradient boosting machine,LightGBM)建模预测。基于曲线下面积(area under curve,AUC)、精确度和召回率进行结果评估。结果共11 261条重症老年患者记录纳入研究。基于当前6 h数据预测时,LightGBM的AUC达0.845~0.925,随机森林、支持向量机和逻辑回归的最高AUC均低于0.73。基于入重症监护病房最初6 h数据,LightGBM效果最好,AUC达0.845。LightGBM应用当前数据比累积数据获得更高的AUC、精确度和召回率,随机森林、支持向量机和逻辑回归反之。结论利用LightGBM对重症老年患者进行急性肾损伤早期连续预测切实可行,仅基于重症监护病房前6 h数据的预测结果就可以达到24 h积累数据的预测效果。此外,不同模型对数据的接收能力和适用性不同,LightGBM在当前数据中表现优于累积数据,其他3种模型在累积数据中表现优于当前数据。

关键词: 机器学习, 疾病预测, 急性肾损伤, 电子病历, 重症监护病房

Abstract: Objective To explore the feasibility of early continuous risk prediction of acute renal injury in severe elderly patients (≥60 years old) and promote the application of machine learning in clinical decision support. Methods The data were collected from the Medical Information Mart for Intensive Care (MIMIC)-Ⅲ database. Logistic regression (LR), support vector machine (SVM), random forest (RF), and light gradient boosting machine (LightGBM) were applied to predict the risk of acute kidney injury (AKI). The prediction results were evaluated based on area under curve (AUC), accuracy, and recall. Results A total of 11 261 intensive care unit (ICU) records were included. When the data of every six hours was used for continuous prediction, AUCs yielded with LightGBM were 0.845－0.925, and those with RF, SVM, and LR were all less than 0.73. As for using the data of the first 6 hours, LightGBM reached AUC 0.845. Compared current data with the cumulative data of ICU, LightGBM yielded higher AUC, accuracy, and recall, whilst it was opposite in RF, SVM, and LR. Conclusion LightGBM completed AKI continuous prediction task with acceptable performance. It is practical to use the data of the first 6 hours on ICU admission for AKI early prediction, which achieve prediction effect of 24－hour accumulated data. In addition, different models have different data applicability. LightGBM performed better based on current data while the other three models favored cumulative data.

Key words: machine learning, disease prediction, acute kidney injury, electronic medical records, intensive care unit

中图分类号:

邬金鸣, 孙海霞, 王嘉阳, 钱庆. 基于电子病历的重症老年患者急性肾损伤连续风险预测研究[J]. 首都医科大学学报, 2022, 43(4): 600-609.

Wu Jinming, Sun Haixia, Wang Jiayang, Qian Qing. Continuous risk prediction of acute kidney injury in elderly critically-ill patients based on electronic medical records[J]. Journal of Capital Medical University, 2022, 43(4): 600-609.

参考文献

[1] Bellomo R, Kellum J A, Ronco C. Acute kidney injury[J]. Lancet, 2012, 380(9843):756－766.
[2] Siew E D, Davenport A. The growth of acute kidney injury: a rising tide or just closer attention to detail?[J]. Kidney Int, 2015, 87(1):46－61.
[3] Hoste E A J, Clermont G, Kersten A, et al. RIFLE criteria for acute kidney injury are associated with hospital mortality in critically ill patients: a cohort analysis[J]. Crit Care, 2006, 10(3):R73.
[4] 王俊琳, 严玉澄. 老年人急性肾损伤的特点及研究进展[J]. 中国中西医结合肾病杂志, 2012, 13(3):272－275.
[5] 陆任华, 方燕, 高嘉元, 等. 住院患者急性肾损伤发病情况及危险因素分析[J]. 中国危重病急救医学, 2011, 23(7):413－417.
[6] Kellum J A, Lameire N, KDIGO AKI Guideline Work Group. Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1)[J]. Crit Care, 2013, 17(1):204.
[7] Ahmed A, Vairavan S, Akhoundi A, et al. Development and validation of electronic surveillance tool for acute kidney injury: a retrospective analysis[J]. J Crit Care, 2015, 30(5):988－993.
[8] Pickkers P, Ostermann M, Joannidis M, et al. The intensive care medicine agenda on acute kidney injury[J]. Intensive Care Med, 2017, 43(9):1198－1209.
[9] Kristovic D, Horvatic I, Husedzinovic I, et al. Cardiac surgery-associated acute kidney injury: risk factors analysis and comparison of prediction models[J]. Interact Cardiovasc Thorac Surg, 2015, 21(3):366－373.
[10] Lee H C, Yoon S B, Yang S M, et al. Prediction of acute kidney injury after liver transplantation: machine learning approaches vs. logistic regression model[J]. J Clin Med, 2018, 7(11):428.
[11] Kate R J, Perez R M, Mazumdar D, et al. Prediction and detection models for acute kidney injury in hospitalized older adults[J]. BMC Med Inform Decis Mak, 2016, 16:39.
[12] Flechet M, Güiza F, Schetz M, et al. AKIpredictor, an online prognostic calculator for acute kidney injury in adult critically ill patients: development, validation and comparison to serum neutrophil gelatinase-associated lipocalin[J]. Intensive Care Med, 2017, 43(6):764－773.
[13] Mohamadlou H, Lynn-Palevsky A, Barton C, et al. Prediction of acute kidney injury with a machine learning algorithm using electronic health record data[J]. Can J Kidney Health Dis, 2018, 5: 2054358118776326.
[14] Sanchez-Pinto L N, Khemani R G. Development of a prediction model of early acute kidney injury in critically ill children using electronic health record data[J]. Pediatr Crit Care Med, 2016, 17(6):508－515.
[15] Zimmerman L P, Reyfman P A, Smith A D R, et al. Early prediction of acute kidney injury following ICU admission using a multivariate panel of physiological measurements[J]. BMC Med Inform Decis Mak, 2019, 19(1):16.
[16] Sun M X, Baron J, Dighe A, et al. Early prediction of acute kidney injury in critical care setting using clinical notes and structured multivariate physiological measurements[J]. Stud Health Technol Inform, 2019, 264:368－372.
[17] Li Y K, Yao L, Mao C S, et al. Early prediction of acute kidney injury in critical care setting using clinical notes[J]. Proceedings (IEEE Int Conf Bioinformatics Biomed), 2018, 2018:683－686.
[18] Chiofolo C, Chbat N, Ghosh E, et al. Automated continuous acute kidney injury prediction and surveillance: a random forest model[J]. Mayo Clin Proc, 2019, 94(5):783－792.
[19] Kiers H D, van den Boogaard M, Schoenmakers M C J, et al. Comparison and clinical suitability of eight prediction models for cardiac surgery-related acute kidney injury[J]. Nephrol Dial Transplant, 2013, 28(2):345－351.
[20] Lee H C, Yoon H K, Nam K, et al. Derivation and validation of machine learning approaches to predict acute kidney injury after cardiac surgery[J]. J Clin Med, 2018, 7(10):322.
[21] Callejas R, Panadero A, Vives M, et al. Preoperative predictive model for acute kidney injury after elective cardiac surgery: a prospective multicenter cohort study[J]. Minerva Anestesiol, 2019, 85(1):34－44.
[22] Huen S C, Parikh C R. Predicting acute kidney injury after cardiac surgery: a systematic review[J]. Ann Thorac Surg, 2012, 93(1):337－347.
[23] Mizota T, Yamamoto Y, Hamada M, et al. Intraoperative oliguria predicts acute kidney injury after major abdominal surgery[J]. Br J Anaesth, 2017, 119(6):1127－1134.
[24] Wilson T, Quan S, Cheema K, et al. Risk prediction models for acute kidney injury following major noncardiac surgery: systematic review[J]. Nephrol Dial Transplant, 2016, 31(2):231－240.
[25] Caragata R, Wyssusek K H, Kruger P. Acute kidney injury following liver transplantation: a systematic review of published predictive models[J]. Anaesth Intensive Care, 2016, 44(2):251－261.
[26] Lin K Y, Zheng W P, Bei W J, et al. A novel risk score model for prediction of contrast-induced nephropathy after emergent percutaneous coronary intervention[J]. Int J Cardiol, 2017, 230:402－412.
[27] Yin W J, Yi Y H, Guan X F, et al. Preprocedural prediction model for Contrast-Induced nephropathy patients[J]. J Am Heart Assoc, 2017, 6(2):e004498.
[28] Silver S A, Shah P M, Chertow G M, et al. Risk prediction models for contrast induced nephropathy: systematic review[J]. BMJ, 2015, 351:26316642.
[29] Zhou Y H, Fan W F, Dong J J, et al. Establishment of a model to predict the prognosis of pregnancy-related acute kidney injury[J]. Minerva Urol Nefrol, 2018, 70(4):437－443.
[30] Johnson A E, Pollard T J, Shen L, et al. MIMIC-Ⅲ, a freely accessible critical care database[J]. Sci Data, 2016, 3:160035.
[31] MIT Critical Data. Secondary analysis of electronic health records[M]. Cham: Springer, 2016.
[32] Tomaev N, Glorot X, Rae J W, et al. A clinically applicable approach to continuous prediction of future acute kidney injury[J]. Nature, 2019, 572(7767):116－119.
[33] Kate R J, Perez R M, Mazumdar D, et al. Prediction and detection models for acute kidney injury in hospitalized older adults[J]. BMC Med Inform Decis Mak, 2016, 16:39.
[34] Ataei N, Ameli S, Yousefifard M, et al. Urinary neutrophil gelatinase-associated lipocalin (NGAL) and cystatin C in early detection of pediatric acute kidney injury; a diagnostic accuracy study[J]. Emerg (Tehran), 2018, 6(1):e2.
[35] 唐咨林. 基于电子病历时态数据的住院病人急性肾损伤风险预测[D]. 广州: 暨南大学, 2018.
[36] Ke G L, Meng Q, Finley T, et al. LightGBM: a highly efficient gradient boosting decision tree[C]//von Luxburg U, Guyon I, Bengio S,et al.Proceedings of the 31st International Conference on Neural Information Processing Systems. NY USA: Curran Associates Inc, 2017: 3149-3157.
[37] 张渊, 冯聪, 李开源, 等. ICU患者急性肾损伤发生风险的LightGBM预测模型[J]. 解放军医学院学报, 2019, 40(4):316－320.
[38] 蔺轲, 谢俊卿, 胡永华, 等. 支持向量机在ICU急性肾损伤患者住院死亡风险预测中的应用[J]. 北京大学学报:医学版, 2018, 50(2):239－244.
[39] 曹正凤. 随机森林算法优化研究[D]. 北京: 首都经济贸易大学, 2014.

基于电子病历的重症老年患者急性肾损伤连续风险预测研究

Continuous risk prediction of acute kidney injury in elderly critically-ill patients based on electronic medical records

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 14

编辑推荐

Metrics

本文评价

[1]	陈晓媚, 马圆媛, 聂静, 朱风新. 急性肾损伤向慢性肾脏病转化中肾小管上皮细胞损伤修复机制的研究进展[J]. 首都医科大学学报, 2022, 43(5): 720-727.
[2]	杨丰春, 郑思, 李姣. 可解释机器学习方法在疾病预测中的应用:脓毒血症患者死亡风险研究[J]. 首都医科大学学报, 2022, 43(4): 610-617.
[3]	刘鑫子, 李自艳, 郑思思, 朱虹, 尹冬青, 贾竑晓. 基于机器学习的单相抑郁与双相抑郁鉴别的脉图参数分类模型[J]. 首都医科大学学报, 2022, 43(3): 407-414.
[4]	齐志丽, 林瑾, 董磊, 刘景峰, 段美丽. 重症监护病房血流感染危险因素的Meta分析[J]. 首都医科大学学报, 2021, 42(6): 923-930.
[5]	冷菲, 李巍. 基于XGBoost对肺鳞癌和肺腺癌的分类预测[J]. 首都医科大学学报, 2019, 40(6): 889-893.
[6]	岳竹, 李昂, 刘培, 白静, 庄海舟, 段美丽. 肝移植术后急性肾损伤的发生率及危险因素[J]. 首都医科大学学报, 2018, 39(1): 14-20.
[7]	于淑婧, 刘培, 林瑾, 刘景峰, 冀晓俊, 刘壮, 董磊, 段美丽. 脓毒症相关急性肾损伤患者肾功能转归的临床研究[J]. 首都医科大学学报, 2018, 39(1): 21-27.
[8]	齐志丽, 刘培, 庄海舟, 段美丽. 重症监护病房院内获得性感染的临床分析[J]. 首都医科大学学报, 2018, 39(1): 41-44.
[9]	方伯梁, 钱素云. PICU脓毒症及严重脓毒症患儿肠内营养摄入能量及蛋白质充分性及其与预后的关系[J]. 首都医科大学学报, 2016, 37(2): 125-129.
[10]	荣光, 胡佳卉, 孟庆刚, 谢晴宇. 重症急性肾损伤早期与晚期应用肾脏替代法疗效的Meta分析[J]. 首都医科大学学报, 2016, 37(2): 188-195.
[11]	赵立红, 尹宁宁, 李倩, 陈晗, 史中华, 徐明, 石广志, 周建新. 右美托咪定对开颅术后延迟拔管患者应激性激素和血流动力学的影响[J]. 首都医科大学学报, 2015, 36(6): 964-968.
[12]	栾海霞, 袁慧, 贺建勋, 马旭, 姚兴荣, 曾小莉, 桂媛, 刘玉磊. 尿NGAL不同时点浓度变化在心脏术后急性肾损伤的早期诊断价值[J]. 首都医科大学学报, 2015, 36(4): 592-596.
[13]	蒋波, 姜利, 席修明. 不同生物学标志物对感染性急性肾损伤肾脏替代治疗的预测作用[J]. 首都医科大学学报, 2013, 34(2): 187-190.
[14]	孟广蕊;李春盛. 肾损伤分子-1对于脓毒症相关肾损伤的早期预测价值[J]. 首都医科大学学报, 2012, 33(3): 394-398.