Spatial analysis of the distribution of Lyme disease in Wisconsin.

Abstract

Surveillance measures for human cases of Lyme disease in Wisconsin were compared and associated with tick distribution and vegetation coverage. During 1991-1994, 1,759 confirmed human cases of Lyme disease reported to the Wisconsin Division of Health were assigned a county of residence, but only 329 (19%) could be assigned with certainty a county of exposure. Distributions of cases by county of exposure and residence were often consistent from year to year. Tick distribution in 46 of 72 Wisconsin counties was mapped based on collections by researchers, statewide surveys of infested deer, and submissions from the public. Satellite data were used to calculate a normalized difference vegetation index (NDVI) for each county. A geographic information system (GIS) was used to map distributions of human Lyme disease cases, ticks, and degree of vegetation cover. Human case distribution by county of exposure was significantly correlated with tick distribution; both were positively correlated with high NDVI values in spring and fall, when wooded vegetation could be distinguished from agricultural crops in the satellite image. Statistical analysis of spatial patterns using a measure of spatial autocorrelation indicated that counties with most human cases and ticks were clustered in parts of western Wisconsin. A map delineating the counties with highest risk for Lyme disease transmission was generated based on numbers of exposed human cases and tick concentrations.