【作者】 陈丽丽；

【导师】 刘爱忠； 胡世雄；

【作者基本信息】 中南大学 ， 公共卫生（专业学位）， 2023， 硕士

【摘要】 目的:了解湖南省近10年的流感流行规律,及时发现有意义的流感病毒变异,探究流感与气象因素的相关性,有效预测流感的流行,为流感的防控提供理论依据。方法:本研究基于湖南省流感监测数据及流感病原学监测数据,首先采用描述性流行病学的方法对湖南省2012-2021年流感的流行特征进行分析,包括流感流行概况、三间分布情况等;然后用Spearman秩相关分析流感样病例与气象因素的关系并运用分布滞后非线性模型(Distributed lag non-linear model,DLNM)探究湖南省气象因素对流感的影响;利用移动流行区间法(Moving epidemic method,MEM)预警流感活动的流行强度;最后,构建差分自回归移动平均模型(Autoregressive integrated moving average,ARIMA)预测流感流行情况。结果:(1)2012-2021年湖南省共报告门急诊就诊病例50518324人次,其中流感样病例(Influenza Like Illness,ILI)为2164934人次,占门急诊病例就诊总数的4.29%。2012-2021年湖南省报告的ILI整体呈上升趋势,2019年报告的ILI最多,为313047人次。湖南省流感有明显季节性,每年冬春季和夏季是发病高峰期,每年7月-8月、12月-次年1月是流感高发季节。2012-2021年湖南省流感在各个年龄段均有发病,以0～14岁人群为主,尤以0～4岁的病例居多,占总报告例数的87.52%。研究期间,岳阳市和永州市报告的流感样病例就诊百分比(Influenza-Like Illness Percentage,ILI%)居于全省前两位,总体ILI%水平均达到7%以上。(2)2012-2021年湖南省检测的ILI标本中共20236份为阳性,其中季H3型最多,共6570份,其次是BV型5408份和新甲H1型5024份,季H1型较少,共4份。流感病毒检测总阳性率为11.73%。2019年ILI标本流感病毒检测阳性率最高,为20.55%,2020年ILI标本流感病毒检测阳性率最低,为4.18%。研究期间,除25～59岁年龄组人群流感病毒优势株为新甲H1型外,其余各年龄组流感病毒优势株相同,均为季H3型。在湖南省各地区中,研究期间ILI标本中检测阳性率高于15%的地区有怀化市、娄底市;ILI标本中检测阳性率在10%～15%之间的的地区有株洲市、岳阳市、长沙市、常德市、郴州市是、邵阳市、湘潭市;张家界市ILI标本检测阳性率最低,为5.70%。(3)以湖南省各气象因素的中位数作为参考值,发现各气象因素均与ILI%呈非线性暴露反应关系,低平均风速(<1.74m/s)、低平均气温(<18.44℃)、高平均气压(>1000hpa)、高平均降水量(>3.13mm)、高平均相对湿度(>79%)、低平均日照时数(<3.41h)、低平均温差(<7.37℃)或高平均温差(>11℃)能够增加流感样病例发病风险。(4)MEM模型对湖南省流感预警结果显示,2020年第41周至2021年第20周,从第13周开始进入流行期,为低流行水平。2021年第21周至第40周,从第21周开始进入流行期,第21至23周为中等流行水平,第24周开始进入低流行水平。模型灵敏度及特异度均较高。(5)ARIMA模型可用于预测湖南省ILI%(MAE=0.39,RMSE=0.58,MAPE=8.86%)。结论:(1)湖南省2012-2021年流感样病例就诊比总体有明显起伏,在2013年、2018年和2020年各有1个发病高峰,其余年份均存在冬春季和夏季两个流行高峰。研究期间,长沙市和邵阳市报告的ILI居于全省前两位。高发人群集中在0～15岁儿童。(2)湖南省流感流行的主要病原型别为季H3型、BV型、新甲H1型及BY型,各病原型别呈交替流行或共同流行。5～14岁人群流感病毒检测阳性率最高。(3)各气象因素均与流感样病例就诊比呈非线性滞后分布关系,低平均风速(<1.74m/s)、低平均气温(<18.44℃)、高平均气压(>1000hpa)、高平均降水量(>3.13mm)、高平均相对湿度(>79%)、低平均日照时数(<3.41h)、低平均温差(<7.37℃)或高平均温差(>11℃)能够增加流感样病例发病风险。(4)本研究利用MEM建立的预警模型灵敏度和特异度都较高,可用于在湖南省建立流感预警系统,识别流感流行和评估流行强度。用于湖南省ILI%预测的ARIMA模型有良好的预测效果。图29幅,表21个,参考文献75篇更多还原

【Abstract】 Objective: This study aims to understand the influenza epidemic patterns in Hunan province over the past 10 years,timely identify meaningful influenza virus variation,explore the correlation between influenza and meteorological factors,effectively predict the epidemic of influenza.which help to to provide a theoretical basis for the prevention and control of influenza.Methods: This study is based on influenza surveillance data and influenza pathogenic surveillance data in Hunan province.Firstly,the descriptive epidemiological method was used to analyze the characteristics of influenza epidemic in Hunan province from 2012 to 2021,including the epidemiological profile and three distribution of influenza.Then,using the Spearman’s correlation coefficient for ranked data to analyze the relationship between influenza-like illness percentage and meteorological factors.The DLNM was used to explore the impact of meteorological factors on Influenza-like illness percentage in Hunan province.The MEM was used to warn the intensity of influenza activity.Finally,the ARIMA model was constructed to predict influenza epidemics.Results:(1)From 2012 to 2021,23 national influenza surveillance hospitals in Hunan province reported a total of 50,518,324 outpatient visits,including 2,164,934 cases of ILI,accounting for 4.29% of the total outpatient visits.The number of ILI reported in Hunan province from 2012 to 2021 showed an overall upward trend,with the highest number of ILI reported in 2019 at 313,047.Influenza in Hunan province shows clear seasonality,with peak incidence in winter,spring and summer,and peak season from July to August and December to January each year.Influenza occurred in all age groups during this period,with the majority of cases in the 0 to 14 age group,particularly in the 0 to 4 age group,which accounted for 87.52% of the total reported cases.During the study period,Changsha City and Shaoyang City reported the highest number of ILI in the province,with 676,192 and 234,863 cases respectively.Yueyang City and Yongzhou City reported the highest ILI% in the province,with an overall ILI% level of over 7%.(2)From 2012 to 2021,a total of 20,236 ILI specimens sent to and tested at national influenza surveillance sentinel hospitals in Hunan Province were positive,of which 6,570 were of the seasonal H3 type,followed by 5,408 of the BV type and 5,024 of the new A H1 type,and 4of the seasonal H1 type were less.The total positive rate for influenza was11.73%.In 2019,the positivity rate for influenza virus detection in ILI specimens was the highest at 20.55%,while in 2020,it was the lowest at4.18%.During the study period,except for the 25 to 59 age group,where the dominant strain was new A H1 type,the dominant strain in all other age groups was H3N2.In Hunan province,areas with a positivity rate for influenza virus detection in ILI specimens above 15% included Huaihua City and Loudi City,while areas with a positivity rate between 10% and15% included Zhuzhou City,Yueyang City,Changsha City,Changde City,Chenzhou City,Shaoyang City,and Xiangtan City.Zhangjiajie City had the lowest positivity rate for influenza virus detection in ILI specimens at5.70%.(3)Using the median of each meteorological factor in Hunan Province as a reference value,all meteorological factors were found to have a non-linear exposure response relationship with ILI%,with low mean wind speed(<1.74m/s),low mean temperature(<18.44℃)high mean air pressure(>1000hpa),high mean precipitation(>3.13mm),high mean relative humidity(>79%),low mean sunshine hours(< 3.41h),low mean temperature difference(<7.37℃)or high mean temperature difference(>11℃)can increase the risk of ILI%.(4)The results of the MEM for early warning of influenza in Hunan Province showed that from week 41 of 2020 to week 20 of 2021,the epidemic period started from week 13 with a low epidemic level,from week 21 to week 40 of 2021,the epidemic period started from week 21 with a medium epidemic level from week 21 to 23 and a low epidemic level from week 24.The sensitivity and specificity of the model were both high.(5)The ARIMA model could be used to predict ILI% in Hunan Province(MAE=0.39,RMSE=0.58,MAPE=8.86%).Conclusions:(1)The number of ILI in Hunan province from 2012 to2021 showed significant fluctuations,with peaks in 2013,2018,and 2020,and seasonal peaks in winter,spring and summer in other years.Changsha City and Shaoyang City reported the highest number of ILI in the province,and the most affected population was children aged 0 to 15 years.(2)The main pathogens of influenza epidemic in Hunan Province were Seasonal H3,BV,New A H1,and BY,with alternating or co-prevalence among the pathogens.The highest positivity rate for influenza virus detection was found in the 5 to14 year age group.(3)All meteorological factors showed a non-linear lagged distribution with the influenza-like case attendance ratio,with low mean wind speed(<1.74m/s),low mean temperature(<18.44℃),high mean air pressure(>1000hpa),high mean precipitation(>3.13mm),high mean relative humidity(>79%),low mean sunshine hours(<3.41h),low mean temperature difference(<7.37℃)or a high mean temperature difference(>11℃)can increase the risk of ILI%.(4)The MEM model established in this study has high sensitivity and specificity and can be used to establish an influenza warning system in Hunan province to identify influenza outbreaks and assess the intensity of the outbreak.The ARIMA model for predicting ILI% in Hunan Province had good performance.更多还原