Integrated detection and analysis on the clusters of schistosomiasis based on geographic information system

Abstract

To detect and analyze the clusters of schistosomiasis on marshland and lake areas based on geographic information system (GIS) in 2008 and to provide suggestions for the development of integrated methodology on the detection of clusters on related diseases. Moran's I of global spatial autocorrelation, Getis-Ord Gi of local spatial autocorrelation and SaTScan were used to detect the schistosomiasis clusters based on GIS and comparison of the results for different methods were performed. Results from the global Moran's I tests for all the marshland and lake areas related to the schistosomiasis were statistically significant (P < 0.05) and indicated spatial heterogeneity;the z-score outcomes as calculated by Getis-Ord Gi indicated high cluster that 50 clusters were categorized at the 0.05 significance level and the z-score of these 45 clusters were more than 0. Results of SaTScan statistics appeared the same as local spatial autocorrelation and almost showing the existence of 5 cluster areas. The geographical distribution of clusters where schistosomiasis was prevalent showed that it was closely corresponded to the middle and lower Yangtse river and in particular, many clusters were located near the boundary of Hubei and Hunan province.