Predication of hand,foot and mouth disease incidence in Hunan province using SARIMA model

Objective To establish a seasonal autoregressive integrated moving average(SARIMA)model to predicate incidence trend of hand,foot and mouth disease(HFMD)in Hunan province for effective control of the disease.Methods Data on monthly incidence of HFMD from May 2008 to December 2013 were collected from "China Information System for Disease Control and Prevention" for the construction of the SARIMA model and then the established model was verified based on HFMD incidence data from January to July of 2014;finally monthly HFMD incidences from August 2014 to July 2015 were predicted with the model established.Modules of Define Dates and ForeCasting in SPSS 18.0 were used in the analyses.Results SARIMA(1,0,0)(1,1,0)₁₂ was established for monthly HFMD incidence in Hunan province;the autoregressive model parameter of AR1 was 0.765(t=8.789,P<0.001)and the residuals of the model were white noise(Ljung-Box Q=15.420,P=0.494).The relative error between actual and predicted values ranged from 6.90% to 46.31% and the average of the relative error was 20.37%.Based on the predication of the model,the incidence peak of HFMD in 2014 would be around in May and November,and the peak incidences in 2014 would be higher than those in same months of 2013 and a peak incidence in the first half year of 2015 would occur in May and could be lower than that in the same month of 2014.Conclusion The established SARIMA(1,0,0)(1,1,0)₁₂model is of good fitting effect and could be applied in short-term predication of HFMD incidence in Hunan province.