Using Spatially Adaptive Filters to Map Late Stage Colorectal Cancer Incidence in Iowa

Abstract

Disease rates computed for small areas such as zip codes, census tractsor census block groups are known to be unstable because of the smallpopulations at risk. All people in Iowa diagnosed with colorectal cancerbetween 1993 and 1997 were classified by cancer stage at the time oftheir first diagnosis. The ratios of the number of late-stage cancers tocancers at all stages were computed for spatial aggregations of circlescentered on individual grid points of a regular grid. Late-stagecolorectal cancer incidence rates were computed at each grid point byvarying the size of the spatial filter until it met a minimum thresholdon the total number of colorectal cancer incidences. Thesedifferent-sized areas are known as spatially adaptive filters. Thevariances analyzed at grid points showed that the maps produced usingspatially adaptive filters gave higher statistical stability in computedrates and greater geographic detail when compared to maps produced usingconventional fixed-size filters.