摘要
目的:探究儿童重症支原体肺炎(mycoplasma pneumonia,MPP)早期筛查情况并构建预后预测列线图(nomogam)模型.方法:选取成都市第七人民医院、攀枝花市中心医院、温江区妇幼保健院2022年12月―2024年1月收治的400例重症MPP患儿,将患者按照7∶3比例(随机数字表法)随机分为建模组(280例)与验证组(120例).根据患儿预后情况将建模组分为预后良好组和预后不良组.儿童重症MPP预后不良的影响因素采用多因素Logistic回归分析;采用R软件构建Nomogam模型.绘制ROC曲线评估Nomogam模型的区分度;通过校准曲线来评价模型的一致性;使用临床决策曲线(DCA)对该模型的临床应用价值进行评估.结果:280例患儿有79例发生预后不良,发生率为28.21%.两组年龄、降钙素原(PCT)、白细胞介素-6(IL-6)、中性粒细胞与淋巴细胞比值(NLR)、乳酸脱氢酶(LDH)、C反应蛋白(CRP)和免疫球蛋白M(IgM)比较有差异.多因素Logistic回归分析年龄、PCT、IL-6、NLR、LDH、CRP和IgM是儿童重症MPP预后不良的危险因素.建模组和验证组AUC为0.974和0.967,校准曲线斜率与1接近,H-L检验为χ2=7.021/6.984,P=0.742/0.698,一致性良好.DCA曲线可知,高风险阈值概率在0.03～0.91时,该模型预测儿童重症MPP预后不良的临床使用价值较高.结论:年龄、PCT、IL-6、NLR、LDH、CRP和IgM是儿童重症MPP预后不良的影响因素,以此构建的列线图模型区分度与一致性良好,可早期筛查儿童重症MPP预后不良发生的风险.
Abstract
Objective To construct a Nomogam model for predicting prognosis based on multicenter data from early screen-ing of severe mycoplasma pneumoniae pneumonia(MPP)in children.Methods A total of 400 children with severe MPP ad-mitted to Chengdu Seventh People's Hospital,Panzhihua Central Hospital,and Wenjiang Maternal and Child Health Hospital from December 2022 to January 2024 were gathered.Patients were randomly grouped into a modeling group(280 cases)and a validation group(120 cases)in a 7∶3 ratio(random number table method).According to the prognosis of the patients,the mod-eling group was separated into a good prognosis group and a poor prognosis group.The influencing factors for poor prognosis of severe MPP in children were analyzed using multiple logistic regression analysis.R software was applied to build the Nomogam model.ROC curve was plotted to evaluate the discrimination of the Nomogam model.Model consistency was applied to draw calibration curves for evaluation.Decision Curve Analysis(DCA)was applied to evaluate the clinical application value of the model.Results Out of 280 children,79 had poor prognosis,with an incidence rate of 28.21%.There were differences in age,PCT,IL-6,NLR,LDH,CRP,and IgM between the two groups.Multivariate logistic regression analysis showed that age,PCT,IL-6,NLR,LDH,CRP,and IgM were risk factors for poor prognosis of severe MPP in children.The AUC of the modeling group and validation group was 0.974 and 0.967,and the slope of the calibration curve was close to 1,the H-L test showed χ2=7.021 and 6.984,P=0.742 and 0.698,indicating good consistency.DCA curve showed that when the probability of high-risk thresh-old was between 0.03 and 0.91,the model had high clinical value in predicting poor prognosis of severe MPP in children.Conclusion Age,PCT,IL-6,NLR,LDH,CRP,and IgM are factors influencing poor prognosis in children with severe MPP,and the column-line diagram model constructed in this way has good discrimination and consistency,and can be used to screen the risk of poor prognosis in children with severe MPP at an early stage.
NETL
NSTL
万方数据