Determinants of the Incidence of Hand, Foot and Mouth Disease in China Using Geographically Weighted Regression Models

Background: Over the past two decades, major epidemics of hand, foot, and mouth disease (HFMD) have occurred throughout most of the West-Pacific Region countries, causing thousands of deaths among children. However, few studies have examined potential determinants of the incidence of HFMD. Methods: Reported HFMD cases from 2912 counties in China were obtained for May 2008. The monthly HFMD cumulative incidence was calculated for children aged 9 years and younger. Child population density (CPD) and six climate factors (average-temperature [AT], average-minimum-temperature [AT(min)], average-maximum-temperature [AT(max)], average-temperature-difference [AT(diff)], average-relative-humidity [ARH], and monthly precipitation [MP]) were selected as potential explanatory variables for the study. Geographically weighted regression (GWR) models were used to explore the associations between the selected factors and HFMD incidence at county level. Results: There were 176,111 HFMD cases reported in the studied counties. The adjusted monthly cumulative incidence by county ranged from 0.26 cases per 100,000 children to 2549.00 per 100,000 children. For local univariate GWR models, the percentage of counties with statistical significance (p < 0.05) between HFMD incidence and each of the seven factors were: CPD 84.3%, AT(max) 54.9%, AT 57.8%, AT(min) 61.2%, ARH 54.4%, MP 50.3%, and AT(diff) 51.6%. The R-2 for the seven factors' univariate GWR models are CPD 0.56, AT(max) 0.53, AT 0.52, MP 0.51, AT(min) 0.52, ARH 0.51, and AT(diff) 0.51, respectively. CPD, MP, AT, ARH and AT(diff) were further included in the multivariate GWR model, with R-2 0.62, and all counties show statistically significant relationship. Conclusion: Child population density and climate factors are potential determinants of the HFMD incidence in most areas in China. The strength and direction of association between these factors and the incidence of HFDM is spatially heterogeneous at the local geographic level, and child population density has a greater influence on the incidence of HFMD than the climate factors.