Effects of meteorological factors on occurrence of HFMD using time series analysis

Huifen FENG,Qiumin ZHAO,Guangcai DUAN,Guang ZHU,Shuling LI

DOI: https://doi.org/10.13705/j.issn.1671-6825.2015.02.007

2015-01-01

Abstract:Aim: To examine the potential effects of meteorological factors on the hand-foot-mouth disease ( HFMD) epidemic, so as to explore the method to prevent and control HFMD .Methods:The daily incidence and weather surveil-lance data were provided by Erqi Center for Disease Control and Prevention and Zhengzhou Meteorological Administration respectively.The correlation between the meteorological parameters and the weekly number of HFMD in Erqi District were analyzed with Spearman′s rank correlation tests with and without time-lag.Time series modeling was carried out using mult-ivariate SARIMA models when there was significant predictor meteorological variable .Results:HFMD was prevalent from March to July and peaks in April and May .Spearman′s rank correlation test indicated that average atmospheric temperature at lag 2 weeks (rS =0.248, P<0.05), maximum atmospheric temperature at lag 2 weeks (rS =0.170, P<0.05) and minimum atmospheric temperature at lag 2 weeks ( rS =-0.223, P<0.05) were correlated with the weekly number of HFMD.Average atmospheric temperature at lag 2 weeks was identified as a significant predictor for the weekly number of HFMD.SARIMA(1,1,0) (0,1,0) 52 associated with average atmospheric temperature at lag 2 weeks were developed and validated for description and predication of the weekly number of HFMD .The goodness of fit [ stationary square R( R2 )=0.797] and the predictive power [root mean square error(RMSE)=11.573] of the model were enhanced by the inclusion of climatic variable to forecast the number of HFMD in the year 2014 compared with the univariate model ( R2 =0.833, RMSE=10.611).Conclusion:Temperatures affect the occurrence of HFMD , and might be used as meteorological pre-dictors for the epidemic of HFMD in this region .