Spatial analysis of West Nile virus: rapid risk assessment of an introduced vector-borne zoonosis

Abstract

The distribution of human risk for West Nile virus was determined by spatial analysis of the initial case distribution for the New York City area in 1999 using remote sensing and geographic information system technologies. Cluster analysis revealed the presence of a statistically significant grouping of cases, which also indicates the area of probable virus introduction. Within the cluster, habitat suitability for potentially infective adult mosquitoes was measured by the amount of vegetation cover using satellite imagery. Logistic regression analysis revealed satellite-derived vegetation abundance to be significantly and positively associated with the presence of human cases. The logistic model was used to estimate the spatial distribution of human risk for West Nile virus throughout New York City. Accuracy of the resulting risk map was cross-validated using virus-positive mosquito sample sites. These new epidemiological methods aid in rapid entry point identification and spatial prediction of human risk of infection for introduced vector-borne pathogens.