Study Of Relation Between Epidemic Features Of Influenza In Tianjin And Meteorological Factors

Objective:Based on the analyses of the data reported by influenza supervision hospitals and etiological test results in the labs in Tianjin within four years from 2006 to 2009, this article analyzes the epidemic features of influenza in Tianjin by combination of descriptive epidemiological methods, trend analyses and time series studies. In the same time, the article gets the meteorological data within the same period to discuss the relations between the epidemic features and the meteorological factors. The article also establishes forecast models by time series analytical methods and forecasts and gives early warnings for the influenza epidemiology so as to provide data and evidences for the prevention and control of influenza.Methods:Collect the ILI cases from the first week in 2006 to the 52nd week in 2009, get the ILI samples for lab tests and conduct descriptive analyses of all the epidemiological data and etiological test results by the supervision of the designated hospitals. Based on the aforementioned influenza materials, establish ARIMA model of the monthly morbidity by SPSS statistic software, testify the adaptability of the model and examine its forecast accuracy by the method of off-group back substitution. Combine meteorological data with the influenza cases, analyze the relativity, establish forecast models and testify the results.Results:the endemic influenza attack all round the year in Tianjin with the peak of ILI percentage from the 48th week to the 8th week next year. In 2009, the peak occurred in advance in the 44th week. The ILI percentage of groups in different age periods is ranked from the highest to the lowest one as follows:0～years old,5～years old,25～years old,15～years old and 60～years old. The low age group (0～years old and 5～years old) accounts for 77.31% as the incidence group. The advantage of influenza virus types is evident, i.e. alternative advent of type A and type B or alternative advent of H1 and H3 in type A. Separation of influenza virus is the highest in winters, less sensitive in springs and the lowest in summers and even become zero in some weeks. It means that influenza virus is sensitive to meteorological factors. Seen from the analyses, the separation of virus is the highest when the average temperature is -5 to 10℃, average relative humidity is 60 to 80% and the average air pressure is 1020 to 1040hPa, which is also the meteorological conditions for the peak of ILI percentage. Establishment of ARIMA(0,1,1)×(0,1,1)12 can forecast the influenza ILI percentage more accurately. The expression of the model is:(1-B)(1-B12)Zt=(1-0.711B)(1-0.953B12)atConclusions:The peak of influenza morbidity in Tianjin appears in winters. This region should make solid work in influenza vaccine vaccination and the recommended best time for influenza vaccine vaccination is autumns (from the 39th week to the 42nd week) and the major vaccination groups are senior people and children. Enhance the identification and diagnosis of respiratory diseases to avoid misdiagnoses. Strengthen the separation extent of the influenza virus from the 39th week to the 13th week in the next year to identify strain variation at early time and be alert to the advent of peak of ILI epidemic when advantageous strains appear. Make further attempts in optimizing the alerting methods and developing data sources to provide practical evidences for establishing highly effectively early forecasting model of influenza.