Spatio-temporal patterns of dengue fever cases in Kaoshiung City, Taiwan, 2003–2008

Dengue fever is a serious vector-borne disease with worldwide incidence and is now considered to be a pandemic. A large body of literature has been developed to explore how the virus spreads and the role of its mosquito vector in this process. A data set of individual dengue fever cases collected in Kaoshiung City, Taiwan, has enabled us to examine dengue's spatial and temporal patterns between 2003 and 2008. Using geocoded individually reported cases, we detected and studied the changes in the spatial clusterness over time. We applied geospatial analysis to further study the spatio-temporal patterns of dengue fever cases in our data set, including hot spot/cold spot analysis and geographically weighted regression models. Using this data of individual cases, we confirmed, to some degree, the importance of the roles that population density, transportation arteries, and water bodies play in the spread of dengue fever. While these factors tend not to change drastically over time, the volatile changes in spatio-temporal patterns detected by our analysis suggest that additional environmental or socioeconomic factors need to be explored.

► DF cases were found to be partially correlated to population density and proximity to roads/rivers.
► Year-to-year variation occurred in locations and numbers of DF cases.
► Spatial diffusion of DF was a combination of expansion and relocation diffusion.
► Measures for containing spread of DF should not be localized efforts.
► More spatially and temporally details are needed for better understanding of how DF cases spread.