Assessment for spatial driving forces of HFMD prevalence in Beijing, China

Abstract

Hand-foot-mouth disease (HFMD) outbreak greatly threatened Beijing city, the capital city of China, in 2008. The control prevention of HFMD has become an urgent mission for Beijing Center for Disease Control and Prevention and a focus problem for the citizens. Medical, social and environmental situations account for much of HFMD morbidity. The spatial driving forces of HFMD occurrence vary across geographical regions, whereas the factors that play a significant role in HFMD prevalence may be concealed by global statistics analysis. This study aims at the identification of the association between the spatial driving forces and HFMD morbidity across the study area and the epidemiological explanation of the results. HFMD spatial driving forces are represented by 6 factors which was obtained by Pearson Correlation analysis and Stepwise Regression method. Compared to Classical Linear Regression Model (CLRM), Geographically weighted regression (GWR) techniques were implemented to predict HFMD morbidity and examine the nonstationary of HFMD spatial driving forces. Informative maps of estimated HFMD morbidity and statistically significant spatial driving forces were generated and rigorously evaluated in quantitative terms. Prediction accuracy by GWR was higher than that by CLRM. The residual led to by CLRM suggested a significant degree of spatial dependence, while that by GWR indicated no significant spatial dependence. In the three regions plotted by Beijing city Ring Roads, HFMD morbidity was found to have significantly positive or negative association with the 6 kinds of spatial driving forces. GWR model can effectively represent the spatial heterogeneity of HFMD driving forces, significantly improve the prediction accuracy and greatly decrease the spatial dependence. The results improve current explanation of HFMD spread in the study area and provide valuable information for adequate disease intervention measures.