Time series analysis of Japanese encephalitis and weather in Linyi City, China

Objectives To examine the relationship between meteorological factors and epidemiological pattern of Japanese encephalitis in Linyi City during 1956-2004. Methods Autoregressive integrated moving average (ARIMA) models were used to evaluate the relationship between monthly JE incidence and weather factors. Logarithmic transformation was applied to the JE incidence series to assure the normality and homogeneity of variance of the residuals. The effect of mass vaccination on JE incidence was also evaluated using a transfer function in the time series analysis. Results The analysis suggested that monthly average temperature [beta = 0.0574, 95% confidence interval (CI) = (0.0172, 0.0976)] and relative humidity [beta = 0.0082, 95% CI = (0.0004, 0.0158)] were positively associated with the logarithmic incidence of Japanese encephalitis after adjusting for mass vaccination in this area. Conclusions Weather variables might be treated as possible predictors of Japanese encephalitis incidence for regions with similar geographic, weather, and socio-economic conditions to Linyi, China.