儿童重症肺炎危险因素及风险预测模型的构建  

Risk factors and establishment of nomogram prediction model of severe pneumonia in children

作　　者：白心怡 李娟[2] 李克晶 卢媛 BAI Xinyi;LI Juan;LI Kejing;LU Yuan(The First School of Clinical Medicine,Lanzhou University,Lanzhou 730000,Gansu;Central Laboratory,the First Hospital of Lanzhou University,Lanzhou 730022,Gansu,China)

机构地区：[1]兰州大学第一临床医学院,兰州730000 [2]兰州大学第一医院中心实验室,兰州730022

出　　处：《临床检验杂志》2023年第9期692-699,共8页Chinese Journal of Clinical Laboratory Science

基　　金：兰州市科技局项目(2020-XG-48)。

摘　　要：目的探讨影响儿童重症肺炎的危险因素并构建风险列线图模型,评估并验证该模型的可行性。方法收集2020年1月至2022年1月兰州大学第一医院诊断为支气管肺炎(188例)和重症肺炎(112例)住院儿童的临床资料共300例为建模组,分别设为儿童轻症肺炎组和儿童重症肺炎组。另收集2022年3月至4月、2023年3月至4月兰州大学第一医院诊断为支气管肺炎(56例)和重症肺炎(67例)住院儿童的临床资料共123例为验证组。建模组与验证组为同一纳入及排除标准,收集患儿临床资料,比较两组资料后用多因素Logistic回归分析筛选儿童重症肺炎的危险因素,采用R软件构建预测儿童重症肺炎的列线图模型,并评价列线图模型的校准度、区分度和获益率,计算模型的简化应用。结果ICU住院史、住院时长、C反应蛋白(CRP)、喘息、单核细胞百分比、心血管并发症为儿童重症肺炎独立危险因素(P<0.05)。利用独立危险因素构建的列线图预测模型采用Bootstrap法重复抽样1000次对列线图模型进行内部验证,校正曲线趋近于理想曲线,C-index=0.925,Hosmer-Lemeshow拟合优度检验显示χ^(2)=11.060,P=0.198,预测模型的ROC曲线下面积(AUC ROC)为0.925(95%CI:0.895~0.955),临床决策曲线(DCA)评估显示获益率良好,并用R软件再次构建模型的简化应用及验证,验证结果与实际一致性较高。结论利用ICU住院史、住院时长、CRP、喘息、单核细胞百分比、心血管并发症这6项独立危险因素构建的儿童重症肺炎的风险列线图模型具有较好的预测价值,可为临床儿童重症肺炎的防治提供参考。Objective To explore the risk factors affecting severe pneumonia in children,construct a nomogram model,and evaluate and verify the feasibility of the model.Methods From January 2020 to January 2022,the clinical data of 300 hospitalized children diagnosed with bronchopneumonia(188 cases)and severe pneumonia(112 cases)in the First Hospital of Lanzhou University were used as the modeling group,including the mild pneumonia group and severe pneumonia group.From March 2022 to April 2022 and March 2023 to April 2023,the clinical data of 123 hospitalized children diagnosed with bronchopneumonia(56 cases)and severe pneumonia(67 cases)in the First Hospital of Lanzhou University were used as the validation group.The modeling group was the same inclusion and exclusion criteria as the validation group.The clinical data of the patients were collected,and the information of the modeling group was compared with the validation group.The multi-factor logistic regression analysis was used to screen the risk factors affected severe pneumonia.The R software was used to build a nomogram model predicting severe pneumonia in children.The calibration and distinction of the nomogram model were evaluated,and the simplified application of the model was calculated.Results The ICU hospitalization history,length of stay,C-reactive protein(CRP),wheeze,percentage of monocyte and cardiovascular complications were independent risk factors affecting severe pneumonia in children(P<0.05).The nomogram prediction model constructed by independent risk factors was performed repeated sampling for 1000 times using the Bootstrap method,and the results showed that the calibration curve of the nomogram model was closed to the ideal curve,C-index=0.925.The Hosmer-Lemeshow good of fit test showedχ^(2)=11.060,P=0.198,the area under the ROC curve(AUC ROC)=0.925,and 95%CI:0.895-0.955.The evaluation of decision curve analysis(DCA)showed a good clinical benefit rate,and the simplified application and validation of the model were reconstructed using the R software,

关 键 词：重症肺炎 儿童 列线图模型 评分模型 

分 类 号：R725.6[医药卫生—儿科]