作者投稿
专家审稿
编辑办公
邮件订阅
RSS
Establishing AKI Warning System in Peking Union Medical College Hospital from a Machine Learning Approach: A Single-center Research Protocol
-
摘要:
研究背景及目的 院内急性肾损伤(acute kidney injury, AKI)是住院患者住院时间延长和预后不良的独立危险因素。利用住院电子病历系统(electronic medical record, EMR)早期预警模型对AKI进行识别并及时干预，对降低AKI发生率、减轻AKI严重程度并改善患者预后具有重要意义。目前基于EMR的AKI相关研究主要针对单学科病房住院患者，采用传统统计学方法进行回顾性分析，尚缺乏基于人工智能技术的大规模多学科病房含时效信息的AKI风险预警模型并以此进行前瞻性干预的研究。本研究计划基于全病程全病历系统收集住院患者的完整临床信息，通过大样本数据及机器学习算法，旨在建立多学科病房的AKI预测模型。 方法 本研究计划分为回顾性研究和前瞻性研究两部分。回顾性研究中，纳入2016年1月1日至2020年12月31日北京协和医院所有成年住院患者。通过全病程全病历系统，收集其一般资料、临床诊断、生命体征、实验室检查结果和住院病历等相关信息，采用Logistic回归、朴素贝叶斯、随机森林、支持向量机、梯度提升决策树、循环神经网络的机器学习算法，构建可预测AKI发生风险的预警模型，并对模型的准确性进行验证。前瞻性研究纳入北京协和医院连续12个月的所有成年住院患者。其中AKI预警系统启动前6个月的所有成年住院患者为对照组，AKI预警系统启动后6个月的所有成年住院患者为干预组。干预组中，将AKI预警系统嵌入EMR，对所有住院24 h以上的患者每6小时进行1次实时未来48 h内AKI风险评估，并对高危患者进行早期干预。对照组无AKI风险评估及报警提示，无相应干预措施。比较两组患者AKI与AKI 3级发生率、AKI缓解率、终末期肾病进展率、住院期间死亡率及住院时间、住院费用等指标差异。 预期结果 回顾性研究中，共纳入约127 000例住院患者，其中院内AKI患者14 605例。构建的多学科病房AKI预测模型可提前24~48 h预测住院患者发生AKI的风险，其中提前24 h预测AKI的受试者工作特征曲线下面积＞0.80。前瞻性研究中，纳入34 748例住院患者，其中干预组和对照组均为17 374例。干预组肾脏替代治疗的时间、住院时间较对照组缩短(P＜0.05)，肾脏替代治疗的比例、AKI与AKI 3级发生率、终末期肾病进展率、住院期间死亡率、住院费用均低于对照组(P＜0.05)，AKI缓解率高于对照组(P＜0.05)。 预期结论 基于EMR构建的多学科病房AKI预测模型可提前24~48 h预测住院患者发生AKI的风险，降低AKI发生率及其严重程度，改善患者预后。 Abstract:Background and Objective In-hospital acute kidney injury (AKI) has a significant negative impact on patients' outcome and the length of hospital stay. It is significantly important to use the early warning of electronic medical records (EMR) to identify and intervene AKI in a timely manner so as to reduce the severity of AKI and to improve the prognosis of patients. At present, AKI-related research based on the EMR system mainly uses traditional statistical methods for retrospective analysis, mainly for inpatients in single-disciplinary wards, and there is still a lack of early warning models of AKI risk based on artificial intelligence technology in large-scale multi-disciplinary wards with time-sensitive information and further prospective research. This study aims to develop a multiple-ward AKI prediction model tailored for general hospitals in China based on machine-learning algorithms and big data acquired by the EMR system. Methods This single-center study consists of both a retrospective observational study and a prospective study. All hospitalized adult patients admitted in Peking Union Medical College Hospital (PUMCH) between 2016 and 2020 were included in the retrospective study. Logistic regression, naive Bayes, random forest, support vector machine, gradient boosting and recurrent neural network will be used for modeling based on demographics, clinical feature, vital signs, imaging, lab results and hospitalized medical records, which aims to predict AKI 24-48 h in advance and will be internally validated. The prospective study intends to include all adult inpatients in PUMCH for 12 consecutive months. Among them, all adult hospitalized patients within 6 months before the AKI early warning system is launched will be of the control group, and all adult hospitalized patients within 6 months after the AKI early warning system is launched will be of the intervention group. In the intervention group, the AKI early warning system will be embedded in the EMR, and all patients hospitalized for more than 24 hours will be assessed for AKI risk in the next 48 hours in real time every 6 hours. Early intervention will be carried out for high-risk patients. The control group does not have above-mentioned high-risk and alarm prompts of AKI, and no corresponding intervention measures. The incidence of AKI and AKI grade 3, AKI remission rate, end-stage renal disease progression rate, mortality during hospitalization, length of stay, hospitalization expenses and other indicators will be compared between the two groups. Expected Results An estimated number of 127 000 in-hospital patients will be included in the retrospective study, among which 14 605 patients suffer from AKI. The prediction model is expected to predict AKI 24-48 h in advance and the aim for area under receiver operating characteristics curve should be > 0.80. In the prospective study, 34 748 inpatients will be enrolled, including 17 374 in both the intervention group and the control group. The duration time of renal replacement therapy and length of hospital stay in the intervention group should be shorter than those in the control group (P < 0.05); the proportion of renal replacement therapy, the incidence of AKI and AKI 3, the rate of progression of end-stage renal disease, the mortality rate during hospitalization, and the hospitalization cost should be lower than those in the control group (P < 0.05), and the AKI remission rate should be higher than that in the control group (P < 0.05). Expected conclusion EMR-based multi-ward AKI prediction model will predict AKI risk 24-48 h in advance, which will lower AKI incidence and severity, and improve clinical outcomes. -
Key words:
- acute kidney injury /
- machine learning /
- multi-ward /
- prediction model
作者贡献：郑华负责实验方案具体设计及文章撰写；张萌负责病案信息提取、清理、统计；赵泽、任菲负责机器学习相关方案设计；林剑峰负责统计方法设计及部分文章撰写；国秀芝负责检验科具体数据提取、清理及统计；夏鹏负责相关临床特征变量选取及临床干预方案设计；邱玲负责检验科相关方案设计、工作协调与合作；周炯负责病案科相关方案设计、工作协调与合作；陈丽萌成立该项目，负责整体框架设计并协调合作。利益冲突：无 -
图 2 预测模型的时间结构及预测的AKI发生风险与SCr变化
A.以1例经门诊就诊并入院治疗的患者为例，该患者电子病历所有数据信息被均分为4个6 h为1个单位的时间段，图中以正方形表示。所有无具体时间的数据信息均归为该天的第5个时间段。现对患者入院后第1天(D1)第3个时段的数据进行分析。除该时间段数据外，同时将其近48 h及更早期的既往数据一起纳入模型。患者在住院第3天(D3)的第3个时间段发生了AKI事件。B.基于D1第3个时间段的数据，预测模型认为未来(以48 h内为例)发生AKI的概率达报警阈值，与第3天(D3)第3时间段内真实AKI事件对应，预测结果准确。C. 患者住院期间SCr变化趋势，同时利用机器学习进行SCr预测，结合SCr预测值，辅助AKI判断
AKI、SCr：同图 1 -
[1] Okusa MD, Davenport A. Reading between the (guide)lines-the KDIGO practice guideline on acute kidney injury in the individual patient[J]. Kidney Int, 2014, 85: 39-48. doi: 10.1038/ki.2013.378 [2] Tang X, Chen D, Yu S, et al. Acute kidney injury burden in different clinical units: Data from nationwide survey in China[J]. PLoS One, 2017, 12: e0171202. doi: 10.1371/journal.pone.0171202 [3] Chawla LS, Amdur RL, Shaw AD, et al. Association bet-ween AKI and long-term renal and cardiovascular outcomes in United States veterans[J]. Clin J Am Soc Nephrol, 2014, 9: 448-456. doi: 10.2215/CJN.02440213 [4] Ikizler TA, Parikh CR, Himmelfarb J, et al. A prospective cohort study of acute kidney injury and kidney outcomes, cardiovascular events, and death[J]. Kidney Int, 2021, 99: 456-465. doi: 10.1016/j.kint.2020.06.032 [5] Yang L, Xing G, Wang L, et al. Acute kidney injury in China: a cross-sectional survey[J]. Lancet, 2015, 386: 1465-1471. doi: 10.1016/S0140-6736(15)00344-X [6] Cheng Y, Luo R, Wang K, et al. Kidney disease is associated with in-hospital death of patients with COVID-19[J]. Kidney Int, 2020, 97: 829-838. doi: 10.1016/j.kint.2020.03.005 [7] Kellum JA, Prowle JR. Paradigms of acute kidney injury in the intensive care setting[J]. Nat Rev Nephrol, 2018, 14: 217-230. doi: 10.1038/nrneph.2017.184 [8] Kolhe NV, Reilly T, Leung J, et al. A simple care bundle for use in acute kidney injury: a propensity score-matched cohort study[J]. Nephrol Dial Transplant, 2016, 31: 1846-1854. doi: 10.1093/ndt/gfw087 [9] Kolhe NV, Staples D, Reilly T, et al. Impact of Compliance with a Care Bundle on Acute Kidney Injury Outcomes: A Prospective Observational Study[J]. PLoS One, 2015, 10: e0132279. doi: 10.1371/journal.pone.0132279 [10] Chandrasekar T, Sharma A, Tennent L, et al. A whole system approach to improving mortality associated with acute kidney injury[J]. QJM, 2017, 110: 657-666. doi: 10.1093/qjmed/hcx101 [11] Hodgson LE, Roderick PJ, Venn RM, et al. The ICE-AKI study: Impact analysis of a Clinical prediction rule and Electronic AKI alert in general medical patients[J]. PLoS One, 2018, 13: e0200584. doi: 10.1371/journal.pone.0200584 [12] Cheng P, Waitman LR, Hu Y, et al. Predicting Inpatient Acute Kidney Injury over Different Time Horizons: How Early and Accurate?[J]. AMIA Ann Symp Proc, 2017, 2017: 565-574. http://europepmc.org/abstract/MED/29854121 [13] Koyner JL, Adhikari R, Edelson DP, et al. Development of a Multicenter Ward-Based AKI Prediction Model[J]. Clin J Am Soc Nephrol, 2016, 11: 1935-1943. doi: 10.2215/CJN.00280116 [14] Tomasev N, Glorot X, Rae JW, et al. A clinically applic-able approach to continuous prediction of future acute kidney injury[J]. Nature, 2019, 572: 116-119. doi: 10.1038/s41586-019-1390-1 [15] Barton AL, Williams SBM, Dickinson SJ. Acute Kidney Injury in Primary Care: A Review of Patient Follow-Up, Mortality, and Hospital Admissions following the Introduction of an AKI Alert System[J]. Nephron, 2020, 144: 498-505. doi: 10.1159/000509855 [16] Wilson FP, Martin M, Yamamoto Y, et al. Electronic health record alerts for acute kidney injury: multicenter, randomized clinical trial[J]. BMJ, 2021, 372: m4786. http://www.bmj.com/content/372/bmj.m4786.abstract [17] Zhou LZ, Yang XB, Guan Y, et al. Development and Validation of a Risk Score for Prediction of Acute Kidney Injury in Patients With Acute Decompensated Heart Failure: A Prospective Cohort Study in China[J]. J Am Heart Assoc, 2016, 5: e004035. [18] Xu N, Zhang Q, Wu G, et al. Derivation and Validation of a Risk Prediction Model for Vancomycin-Associated Acute Kidney Injury in Chinese Population[J]. Ther Clin Risk Manag, 2020, 16: 539-550. doi: 10.2147/TCRM.S253587 [19] Li Y, Chen X, Wang Y, et al. Application of group LASSO regression based Bayesian networks in risk factors exploration and disease prediction for acute kidney injury in hospitalized patients with hematologic malignancies[J]. BMC Nephrol, 2020, 21: 162. doi: 10.1186/s12882-020-01786-w [20] 王洪玲, 田洁, 韩涛. 失代偿性肝硬化伴发急性肾损伤的危险因素分析[J]. 中华肝脏病杂志, 2014, 22: 420-424. doi: 10.3760/cma.j.issn.1007-3418.2014.06.005 Wang HL, Tian J, Han T. Analysis of risk factors for acute kidney injury in patients with decompensated cirrhosis[J]. Zhonghua Ganzangbing Zazhi, 2014, 22: 420-424. doi: 10.3760/cma.j.issn.1007-3418.2014.06.005 [21] 冯芳, 陈宇, 陈伟, 等. 基于危险因素分层的急性肾损伤早期预警模型联合血液灌流在脓毒症患者中的应用: 一项前瞻性观察性先导性研究[J]. 中华危重病急救医学, 2020, 32: 814-818. doi: 10.3760/cma.j.cn121430-20200326-00239 Feng F, Chen Y, Chen W, et al. Application of a risk stratification-based model for prediction of acute kidney injury combined with hemoperfusion in patients with sepsis: a prospective, observational, pilot study[J]. Zhonghua Wei-zhongbing Jijiu Yixue, 2020, 32: 814-818. doi: 10.3760/cma.j.cn121430-20200326-00239 [22] Parreco J, Soe-Lin H, Parks JJ, et al. Comparing Machine Learning Algorithms for Predicting Acute Kidney Injury[J]. Am Surg, 2019, 85: 725-729. doi: 10.1177/000313481908500731 [23] Song X, Yu ASL, Kellum JA, et al. Cross-site transportability of an explainable artificial intelligence model for acute kidney injury prediction[J]. Nat Commun, 2020, 11: 5668. doi: 10.1038/s41467-020-19551-w -
下载:
点击查看大图
图(3)
计量
- 文章访问数: 1115
- HTML全文浏览量: 176
- PDF下载量: 192
- 被引次数: 0
