移动流行区间法在北京市房山区流感流行分级预警中的应用  

Application of moving epidemic methods in the early warning of influenza epidemic classification in Fangshan District of Beijing

作　　者：魏柯雯 马仲慧 杨鹏[2] 王宇 董帅兵 李丽丽[1] WEI Kewen;MA Zhonghui;YANG Peng;WANG Yu;DONG Shuaibing;LI Lili(Fangshan Center for Disease Control and Prevention,Beijing 102488,China;Beijing Center for Disease Control and Prevention,Bejing 100013,China)

机构地区：[1]北京市房山区疾病预防控制中心,北京102488 [2]北京市疾病预防控制中心,北京100013

出　　处：《预防医学情报杂志》2024年第9期1043-1048,共6页Journal of Preventive Medicine Information

基　　金：首都卫生发展科研专项(项目编号:首发2022-1G-3014)。

摘　　要：目的探讨移动流行区间法(moving epidemic methods,MEM)在流感流行预警中的应用,估计2020—2021年北京市房山区流感流行强度,并对其应用效果进行评价。方法收集2014—2020年北京市房山区二级以上医疗机构每周的流感样病例(influenza-like illness,ILI)数和流感样病例百分比(percentage of Influenza-like illness,ILI%)分别建立MEM模型,建立最优模型估计2020—2021年流感流行阈值和分级强度阈值,并用交叉验证法对MEM的预警效果进行验证。结果通过ILI和ILI%拟合模型效果均较好。基于ILI数据建模显示,2020—2021年流行阈值,中、高和极高强度阈值分别为603例、1110例、1705例和1925例。灵敏度为0.84、特异度为0.97、马修斯相关系数为0.80和约登指数为0.80;基于ILI%数据建模显示,2020—2021年流行阈值,中、高和极高强度阈值分别为1.29%、2.36%、3.35%和3.70%。灵敏度为0.84、特异度为0.96、马修斯相关系数为0.79和约登指数为0.80。结论本研究利用ILI和ILI%2种数据建立的MEM模型在北京市房山区具有较高的灵敏度和特异度,可以利用此模型建立流感预警系统,为流感防控提供参考依据。Objective To explore the application of moving epidemic methods(MEM)in the early warning of influenza epidemic,to estimate the influenza epidemic intensity in Fangshan District of Beijing from 2020 to 2021,and to evaluate the application effects.Methods The weekly numbers of influenza-like cases(ILI)and the percentage of influenza-like cases(ILI%)by hospitals of grade two or above in Fangshan District of Beijing from 2014 to 2020 were collected to establish MEM models respectively.An optimal model was established to estimate the epidemic threshold and classification intensity threshold of influenza from 2020 to 2021.The early warning effect of MEM was verified by cross-validation method.Results The models fitted by ILI and ILI%data had good results.Modeling based on ILI data showed that,the epidemic threshold,medium,high and extremely high intensity thresholds from 2020 to 2021 were 603,1110,1705 and 1925 cases,respectively.The sensitivity was 0.84,the specificity was 0.97,the Matthews correlation coefficient was 0.80,and the Jorden′s index was 0.80.Modeling based on ILI%data showed that,the epidemic threshold,medium,high and extremely high intensity thresholds from 2020 to 2021 were1.29%,2.36%,3.35%and 3.70%,respectively.The sensitivity was 0.84,the specificity was 0.96,the Matthews correlation coefficient was 0.79,and the Jorden′s index was 0.80.Conclusions The MEM models based on ILI and ILI%data in this study had high sensitivity and specificity in Fangshan District of Beijing,which can be used to establish an influenza early warning system and provide reference for influenza prevention and control.

关 键 词：流感 移动流行区间法 预警 

分 类 号：R511.7[医药卫生—内科学] R181.3[医药卫生—临床医学]