中小学校水痘暴发疫情风险预测模型构建  

Establish of the risk predictive model for varicella outbreaks in primary and middle schools

作　　者：郑永韬 叶春梅 倪作为 张佳妮 来芬华 高艳敏 杨东波 王岩梅 ZHENG Yongtao;YE Chunmei;NI Zuowei;ZHANG Jiani;LAI Fenhua;GAO Yanmin;YANG Dongbo;WANG Yanmei(Xiaoshan District Center for Disease Control and Prevention,Hangzhou(311203),Zhejiang Province,China)

机构地区：[1]浙江省杭州市萧山区疾病预防控制中心,311203 [2]杭州市临平区疾病预防控制中心 [3]浙江大学医学院附属第一医院 [4]上海市杨浦区控江社区卫生服务中心 [5]杭州市萧山区瓜沥镇社区卫生服务中心 [6]上海中医药大学附属市中医医院

出　　处：《中国学校卫生》2024年第6期873-877,共5页Chinese Journal of School Health

基　　金：上海中医药大学科技发展项目(23YGZX06);杭州市卫生科技计划项目(B20230317);萧山区农业和发展重大科技计划项目(20231070)。

摘　　要：目的 探究中小学校水痘暴发疫情的流行特征并构建风险预测模型,为学校水痘疫情防控提供科学指导。方法采用巢式病例对照研究,将2023年1—12月上海市(杨浦区、静安区)及杭州市(萧山区、临平区)4个区县的中小学校作为研究对象,观察学校水痘疫情发生情况。分析学校水痘暴发疫情发生的影响因素,构建学校水痘疫情发生风险预测模型,并采用H-L拟合优度检验、受试者工作特征曲线(ROC)、Calibration校准曲线、决策曲线(DCA)等对预测模型进行评价。结果 水痘暴发疫情中小学校有98所,同期无水痘暴发疫情的195所学校作为对照,厕所温水洗手台、厕所洗手台洗手液、班级平均学生数、学生每日在校时长、专职校医、校长水痘疫苗犹豫、水痘疫苗1剂次接种率、水痘疫苗2剂次接种率均是学校水痘暴发疫情发生的影响因素,组间差异均有统计学意义(χ^(2)/Z值分别为10.01,20.49,17.43,9.74,32.17,6.60,2.20,3.39,P值均<0.05)。采用上述8个变量构建风险预测模型,H-L拟合优度检验χ^(2)值为5.863(P>0.05);ROC曲线下面积(AUC)为0.846(95%CI=0.799~0.893);Calibration校准曲线显示模型的预测值和实际值一致性较好,DCA曲线显示在较大范围内模型的预测效果较好。结论 学校水痘暴发疫情风险预测模型具有较好的准确性和效果。可利用好学校水痘暴发疫情风险预测模型并采取相应措施,降低水痘疫情在学校的传播风险。Objective To investigate the epidemiological characteristics of varicella outbreaks in primary and middle schools,and to establish a risk predictive model,so as to provide scientific guidance for the prevention of varicella outbreaks in schools.Methods Based on a nested case-control study,primary and middle schools in 4 districts of Shanghai(Yangpu District and Jingan District)and Hangzhou(Xiaoshan District and Linping District)from January to December 2023 were selected to observe the status of varicella outbreaks.Associated factors of varicella outbreaks were investigated and used for establishing the predictive model,which was evaluated by the Hosmer-Lemeshow(H-L)goodness of fit test,receiver operating characteristic(ROC)curve,Calibration curve,decision curve analysis(DCA).Results A total of 98 varicella outbreaks were included,with 195 schools without varicella outbreaks during the same period as controls.Eight factors,including the availability of warm water in restroom,availability of hand soap in restroom,average class size,duration of student attendance at school per day,presence of a full-time school doctor,hesitancy of the school principal towards varicella vaccination,and rates of first and second doses of varicella vaccination,were identified as potential factors for school varicella outbreaks,with statistically significant differences(χ^(2)/Z=10.01,20.49,17.43,9.74,32.17,6.60,2.20,3.39,P<0.05).The 8 variables above were employed to construct a risk predictive model,and Hosmer-Lemeshow goodness of fit test yielded aχ^(2) value of 5.863(P>0.05);the area under the ROC curve(AUC)was 0.846(95%CI=0.799-0.893);Calibration curve analysis indicated good consistency between predicted and actual values of the model.DCA demonstrated favorable predictive performance of the model over a wide range.Conclusions The predictive model for school varicella outbreaks demonstrates satisfactory accuracy and efficacy.It suggested to make good use of this prediction model and take relevant measures to reduce the risk of va

关 键 词：水痘 疾病暴发流行 列线图 学生保健服务 

分 类 号：G4788[文化科学—教育学] R181.32[文化科学—教育技术学] R511.5[医药卫生—流行病学]