Analysis Of Influenza Surveillance Results In Shengzhou And ARIMA Model Prediction

Background:Influenza is an acute respiratory infectious disease caused by influenza viruses,characterized by fever,cough,and generalized aches and pain.Influenza is one of the class C infectious disease stipulated in the law of the People’s Republic of China on infectious disease.Influenza is the first infectious disease monitored globally.As part of Shaoxing city,a national influenza surveillance site,Shengzhou city has been included in the influenza surveillance system as influenza surveillance cooperative city since 2009.Currently,Shengzhou city has 1 influenza surveillance network laboratory(PCR Laboratory of Shengzhou center for disease control and prevention)and 1 sentinel hospital for monitoring the symptoms of influenza-like(the people’s Hospital of Shengzhou City).Shengzhou is a relative developed county-level city which locates in the Middle East part of Zhejiang Province.The population density was relatively high,the floating population was also large,and it had experienced several epidemics of influenza and bird flu in the past.Therefore,long-term influenza surveillance is necessary to prevent and control the spread of influenza in Shengzhou city.Influenza surveillance could also provide effective and scientific basis for the establishment of targeted influenza prevention and control measures,protect the health of the people,and promote the sustainable and healthy development of social economy.Objective:1.To conduct a comprehensive analysis of influenza surveillance results from 2010 to 2016 in Shengzhou city from the characteristics of etiology and epidemiology.To master the epidemic situation,the distribution of outbreaks,and the change of influenza dominant strains of the influenza-like illness of 7 years of Shengzhou City.To provide scientific basis for the scientific development of influenza prevention and control measures.2.The monthly distribution data of the percentage of visits of influenza-like symptoms of influenza sentinel surveillance hospital in Shengzhou during 2010-2016 were recorded in the seasonal ARIMA model to model and fitted to predict the incidence of influenza in the short term,and to make an effective exploration of the early warning and forecasting techniques of influenza and other infectious diseases.Methods:1.Through the subsystem of "China Disease Prevention and control informationsystem","China flu Surveillance information system",descriptive analysis of the monitoring data,the outbreak data and the etiological data of influenza-like illness collected from 2010 to 2016 years in Shengzhou in the subsystem was carried out.2.Used the ARIMA model of SPSS 19.0 software prediction function,on the basis of modeling and fitting the monthly data of the percentage of visits of influenza-like cases in Shengzhou during 2010-2016 influenza,to explore the optimal model for the forecast of the data of January to June 2017,and compare with the actual value to evaluate the effect of the forecast model.Results:1.From 2010 to 2016,a total of 1,735,648 outpatient and emergency cases were reported at the Shengzhou influenza sentinel surveillance hospital,among which the numbers of influenza like illness(ILI)reported was 29,649,and the percentage of visits of influenza like cases(ILI%)was 1.71%.In addition to autumn,which is the active season of flu,the peak of influenza occurred in winter,spring and summer.Winter and spring are usually the main active season,followed by summer.2.From 2010 to 2016,at the influenza sentinel surveillance hospital in Shengzhou,the top three of the age group of the influenza-like cases were 25～,0～and 5～years old age group,15～and 60～years old age groups constitute relatively small portion of influenza-like cases.Overall,the under 15 years old age group was a high-risk group.3.From 2010 to 2016,a total of 4,553 throat swab specimens of flu-like cases were detected in the influenza network laboratory of Shengzhou Center for Disease Control and Prevention,the number of positive specimens of viral nucleic acid was 758,positive rate was 16.65%(758/4553).The lowest positive rate of influenza nucleic acid was 5.61%in 2013,and the highest positive rate of influenza nucleic acid in was 24.46%in 2016.4.From 2010 to 2016,the monthly variation of viral nucleic acid types of influenza like specimens is different in every year in Shengzhou.The predominant strains of influenza in Shengzhou took turns to be epidemic among H3,H1N1 and B.5.From 2010 to 2016,a total of 28 outbreaks of influenza-like cases were disposed in Shengzhou.For the location of the ourbreak,there are 27 outbreaks in schools,and one in a village.For the time distribution of outbreaks,the outbreacks mainly occurred in 2016(14 outbreaks),and for monthly distribution,mainly occurred in November and December in each year,with a total of 15.6.According to the timing distribution map of the percentage of visits of influenza-like cases during 2010-2016 in Shengzhou City,the active seasons of flu sere winter and spring,followed by summer,the timing was seasonal.Through the logarithmic transformation of the original time series data,and carrying out 1 times and 1 times the seasonal difference to reach the stationary of time series,and through debugging and fitting,the optimal model of multiple seasonal ARIMA(1,1,1)(1,1,1)12 was obtained,and its standard numerical value of BIC was-0.377,this BIC is the smallest value of all debug mode,the statistic of residual sequence Ljung-Box of the model was 8.005,P value is 0.889,which indicates that the residual sequenc was random errore,the sequence conformed to the requirements of white noise.Through the comparison between the actual value and predictive value of the percentage of visits of influenza like illness from January to June in 2017 in Shengzhou City,the average error is 31.99%,because the basic data is small,the error is in a reasonable range,from change trend of the predicted value and actual value both can be consistent.Conclusions:1.From 2010 to 2016,the active seasons of influenza in Shengzhou were winter and spring,followed by summer,and autumn is relatively inactive.The incidence of influenza-like illness was concentrated in people under the age of 15.The average positive rate of influenza virus nucleic acid of 7 annual samples was 16.65%,the predominant strains of influenza in Shengzhou took turns to be epidemic among the new H1N1,H3 and B,schools were the main site of outbreaks.2.By modeling,fitting and predicting of the product seasonal ARIMA model,the model can predict the value and trend of the percentage of influenza-like cases(ELI%)well in Shengzhou city,and provide reference for the prediction and warning of other infectious diseases.