Spatial heterogeneity of the associations of economic and health care factors with infant mortality in China using geographically weighted regression and spatial clustering.

Abstract

Economic factors and health care resources are important influential factors of infant mortality. We aimed to examine prefecture-level spatial heterogeneity and clustering of the associations of economic and health care factors with infant mortality rates (IMR) in China. IMR data in 348 prefectures were calculated and adjusted, and economic and health care data were collected in each prefecture in China, 2010. Stepwise regression was used to select important variables, and geographically weighted regression (GWR) was applied to examine the spatial variations of the relationships between economic and health care factors and IMR. The k-means clustering was developed to elucidate the spatial clustering patterns of the GWR coefficients. The results showed that three important variables were selected in the multivariable regression model, including per capita income of rural residents, Engel's coefficient of rural residents, and proportion of government health expenditure. The GWR with these three variables revealed spatial heterogeneity of the associations between IMR and economic and health care factors; western China generally had higher GWR R-squares and stronger associations between IMR and all the three variables than the middle-eastern part of China. Based on the GWR coefficients, three distinct spatial clusters were identified. This study contributes new findings on the spatial heterogeneity of the associations between economic and health care factors and infant mortality rate in China, which calls for region-specific policies to reduce infant mortality in China.