个体化预测小儿反复上呼吸道感染诱发风险的列线图模型构建与评估  

Construction and evaluation of a nomogram model for individualized prediction of the risk of recurrent upper respiratory tract infection in children

作　　者：吴卫华 WU Weihua(Department of Pediatrics,Wuyishan Municipal Hospital,Fujian Province,Wuyishan354300,China)

机构地区：[1]福建省武夷山市立医院儿科,福建武夷山354300

出　　处：《中国当代医药》2023年第28期8-12,17,共6页China Modern Medicine

摘　　要：目的分析急性呼吸道感染(AURTI)患儿反复上呼吸道感染(rURTIs)的影响因素,并构建列线图预测模型。方法选取2021年6月至2022年6月武夷山市立医院收治的426例AURTI患儿作为研究对象,按照7∶3的比例将426例AURTI患儿以简单随机抽样法分为建模队列(n=298)与验证队列(n=128)。采用logistic回归分析影响AURTI患儿发生rURTIs的危险因素,应用R软件绘制列线图预测模型,并对模型进行验证。结果426例AURTI的患儿中,rURTIs者为188例(44.13%),其中建模队列298例AURTI患儿有134例发生rURTIs(44.97%),验证队列128例AURTI患儿有54例发生rURTIs(42.19%)。单因素分析显示,rURTIs组和非rURTIs组哮喘史、过敏史、抗生素使用次数≥3次/年、母乳喂养时间<6个月、被动吸烟、户外活动<2 h/d、早产、偏食、入托的患儿占比比较,差异有统计学意义(P<0.05)。多因素logistic分析结果显示,哮喘史(β=1.439,OR=4.215,95%CI:1.346~13.201)、过敏史(β=0.896,OR=2.451,95%CI:1.254~4.791)、抗生素使用次数≥3次/年(β=1.369,OR=3.930,95%CI:2.145~7.202)、被动吸烟(β=1.496,OR=4.463,95%CI:1.984~10.040)、偏食(β=1.172,OR=3.229,95%CI:1.695~6.152)、入托(β=1.042,OR=2.834,95%CI:1.452~5.533)是影响AURTI患儿发生rURTIs的独立危险因素,而户外活动≥2 h/d(β=-1.371,OR=0.254,95%CI:0.133~0.486)、母乳喂养时间≥6个月(β=-1.462,OR=0.232,95%CI:0.116~0.462)是保护因素(P<0.05)。利用以上8个风险预测指标构建列线图模型,其内部验证的曲线下面积为0.859,H-L检验的χ^(2)=5.429,P=0.711;外部验证的曲线下面积为0.898,H-L检验的χ^(2)=5.602,P=0.708。结论基于rURTIs的8项诱发因素,构建的列线图模型有较好预测能力,可指导临床医护人员早期识别高危儿童,并采取综合措施进行干预。Objective To analyze the influencing factors of recurrent upper respiratory tract infection(rURTIs)in children with acute respiratory tract infection(AURTI),and construct a line graph prediction model.Methods A total of 426 children with AURTI in Wuyishan Municipal Hospital from June 2021 to June 2022 were selected.According to the ratio of 7∶3,426 children with AURTI were divided into modeling cohort(n=298)and validation cohort(n=128)by simple random sampling method.Logistic regression was used to analyze the risk factors for rURTIs in children with AURTI,and R software was used to draw a nomogram prediction model and verify the model.Results Among the 426 children with AURTI,188 cases(44.13%)had rURTIs.Among the 298 children with AURTI in the modeling cohort,134 cases(44.97%)had rURTIs,whlie 54 cases(42.19%)of the 128 children with AURTI in the verification cohor had rURTIs.Univariate analysis showed that there were statistically significant differences between the rURTIs group and the non-Rurtis group in the proportion of children with asthma history,allergy history,antibiotic use times≥3 times/year,breastfeeding time<6 months,passive smoking,outdoor activities<2 h/d,premature birth,partial eclipse,and admission to child care(P<0.05).The results of multivariate logistic analysis showed that asthma history(β=1.439,OR=4.215,95%CI:1.346-13.201),allergy history(β=0.896,OR=2.451,95%CI:1.254-4.791),antibiotic use times≥3 times/year(β=1.369,OR=3.930,95%CI:2.145-7.202),passive smoking(β=1.496,OR=4.463,95%CI:1.984-10.040),partial eclipse(β=1.172,OR=3.229,95%CI:1.695-6.152),admission to child care(β=1.042,OR=2.834,95%CI:1.452-5.533)were independent risk factors for rURTIs in children with AURTI,and outdoor activities≥2 h/d(β=-1.371,OR=0.254,95%CI:0.133-0.486),breastfeeding time≥6 months(β=-1.462,OR=0.232,95%CI:0.116-0.462)were protective factors(P<0.05).The above 8 risk prediction indexes were used to construct a line graph model.The area under the curve of the internal verification was 0.859,the H-L t

关 键 词：反复上呼吸道感染 诱发因素 列线图 验证 

分 类 号：R725.6[医药卫生—儿科]