Spatiotemporal analysis of West Nile virus incidences in the Twin Cities Metropolitan Area

Abstract: Various approaches are used to identify West Nile virus (WNV) exposure areas by using unusual sightings of infected dead birds, mosquito pools or human cases both prospectively and retrospectively. A significant and largely unmet need is to incorporate the temporal characterization of virus spread and locational information of the three components of transmission cycle, including birds (reservoir), mosquitoes (vector), and humans (host) on a localized scale. Exposure areas containing all three components of the WNV cycle in close proximity have higher potential to amplify an outbreak as compared to exposure areas delineated by a single component only. In this paper, we introduce a novel approach, termed 'Nearest Neighbor Distance Time' or NNDT to delineate and retrospectively monitor WNV transmission cycles on various scales in the Twin Cities of Minnesota. The method is based on the interaction between nonrandom spatial distribution of infected dead birds, mosquito pools, and human cases and the time required for the virus to be transmitted from one component to another in the cycle.  The NNDT model was implemented in a geographic information system using the data from the year 2002 to 2006. It successfully identified transmission cycles with an area ranging from 5 to 150 sq miles. The years 2002 and 2003 had three such WNV cycles, followed by one, two, and four in 2004, 2005, and 2006 respectively. The NNDT can be useful to locate chronically exposed areas and also allow researchers to generate hypotheses related to the transmission of WNV.