预测支原体肺炎病儿并发大叶性肺炎风险的列线图模型的建立

Establishment of a risk prediction line chart model for lobar pneumonia in children with mycoplasma pneumonia

目的探讨支原体肺炎病儿并发大叶肺炎的影响因素,并依此建立列线图预测模型。方法回顾性分析宝鸡市妇幼保健院2019年3月至2022年3月收治的142例支原体肺炎病儿的临床资料,并将其作为模型组。比较模型组中并发组(89例)和未并发组(53例)两组病人的一般资料,采用多因素logistic回归分析法分析模型组支原体肺炎病儿并发大叶肺炎的影响因素,并采用R 3.4.3软件包绘制列线图模型,另收集2016年1月至2019年1月该院收治的224例支原体肺炎病儿的临床资料,将其作为验证组,绘制校准曲线图并采用Bootstrap法检验预测列线图模型的校准度。结果多因素logistic回归分析结果显示,模型组年龄、发热时间、合并胸腔积液、发病至应用大环内酯类药物时间、C-反应蛋白(CRP)和乳酸脱氢酶(LDH)表达水平升高均为影响支原体肺炎病儿并发大叶性肺炎的危险因素(OR=1.84、2.65、4.10、2.84、3.50、2.37,P<0.05)。将以上6个影响因素作为预测指标,构建模型组支原体肺炎病儿并发大叶肺炎的风险预测列线图模型,经Bootstrap法进行验证,模型组的其一致性指数(C-index)为0.832,验证组的C-index为0.829,且校准曲线的校准度较好。结论年龄、发热时间、合并胸腔积液、发病至应用大环内酯类药物时间、CRP和LDH表达水平升高均是支原体肺炎病儿并发大叶肺炎的危险因素,基于以上危险因素构建的列线图模型具有良好的校准度,有助于临床预测并发大叶性肺炎的概率,为临床治疗提供更有力指导。

Objective To explore the influencing factors of children with mycoplasma pneumonia complicated with lobar pneumonia,and to establish a line chart prediction model.Methods The clinical data of 142 children with mycoplasma pneumonia admitted to Baoji Maternal and Child Health Hospital from March 2019 to March 2022 were retrospectively analyzed as the model group.The general data of the concurrent group(89 cases)and the non-concurrent group(53 cases)in the model group were compared.Multivariate Logistic regression analysis was used to analyze the influencing factors of mycoplasma pneumonia complicated with lobar pneumonia.The R3.4.3 software package was used to draw the line graph model.The clinical data of 224 children with mycoplasma pneumonia admitted to the hospital from January 2016 to January 2019 were selected as validation group.The calibration curve was drawn and the Bootstrap method was used to test the calibration degree of the predicted plot model.Results Multivariate logistic regression analysis showed that age,fever time,pleural effusion,time from onset to application of macrolides,elevated levels of C-reactive protein(CRP)and lactate dehydrogenase(LDH)in the model group were risk factors for lobar pneumonia in children with mycoplasma pneumonia(OR=1.84,2.65,4.10,2.84,3.50,2.37,P<0.05).The above six influencing factors were used as the prediction indexes to construct the risk prediction line graph model of children with mycoplasma pneumonia complicated with lobar pneumonia in the model group.The Bootstrap method was used to verify the consistency index(C-index)of the model group was 0.832,and the C-index of the verification group was 0.829,and the calibration degree of calibration curve was better.Conclusion Age,fever time,pleural effusion,time from onset to application of macrolides and elevated levels of CRP and LDH are risk factors for children with mycoplasma pneumonia complicated with lobar pneumonia.The line graph model based on the above risk factors has good calibration,which is helpful to predict the probability of lobar pneumonia and provide more powerful guidance for clinical treatment.