In epidemiology, it is essential to understand the spatial characteristics of diseased population for developing more sophisticated models or sampling programs. The distribution pattern of disease reveals how a disease spreads and what the disease risk factors are. The spatial analysis has been developed for visualizing and analyzing geographically referenced health and environmental data in recent years. Such an analysis method has motivated researchers to identify the potential disease risk factor and to design the efficient disease surveillance as well as control programs.
In this study, we investigated the leading causes of death in the township level to observe the spatial-temporal clustering patterns in Taiwan. We used age-standardized mortality rates by gender from1998 to 2002 and from 2003 to 2007 in 358 townships. First, the exploratory analysis was used to detect whether there were spatial clustering townships with higher or lower mortality rates at local and global levels for each specific cause of death. Second, considering the effect of spatial autocorrelation, the spatial regression model was used to explore the relationships between medical resources, population density and the accident mortality rates.
