Visualizing health data with micromaps.

Maps have long been used to display the geographic patterns of disease. Identification of cancer "hot spots", clusters of high rates, and subsequent speculation as to their cause led to important epidemiologic findings such as a link between smokeless tobacco use and oral cancer. Recent advances in computer systems have made interactive display of spatial data possible on everyone's desktop. In this paper, we describe how spatial patterns of disease and local characteristics can be examined visually and simultaneously using graphics called micromaps. These are graphics that link statistical information to an organized set of small maps in order to explore and communicate patterns in the outcome variable, related covariates, geographic locations and the associations among them. Three types of micromaps are described and illustrated: linked, conditioned and comparative.