GEOGRAPHICAL CLUSTERING OF MORTALITY FROM SYSTEMIC-SCLEROSIS IN THE SOUTHEASTERN UNITED-STATES, 1981-90

Objective. To determine whether elevated rates of mortality from syste mic sclerosis (SSc) in the Southeastern United States result from loca l, multicounty clusters of the disease. Methods. Detection of spatial clusters of SSc mortality by applying the method of Kulldorff and Naga rwalla to death certificate data from 955 counties in 12 southeastern states. Results. From 1981 to 1990, significant excess mortality from SSc in the Southeastern US occurred among white males [standardized mo rtality ratio (SMR) = 1.2; p = 0.0004] and black males (SMR = 1.2; p = 0.04), but not among white females (SMR = 0.98; p = 0.55) or black fe males (SMR = 1.1; p = 0.06). When the cluster detection algorithm was applied to data for white males, 3 significant clusters were identifie d. The primary cluster (p = 0.001) was centered around Coffee, Tenness ee. Two smaller clusters overlapped the primary cluster -one centered at Calhoun, Alabama, (p = 0.008) and another centered at Chattooga, Ge orgia, (p = 0.04). Analysis of data for block males resulted in a sing le significant cluster (p = 0.02) centered at Northampton, North Carol ina. When data for white or black females were analyzed, no clusters r eached statistical significance. In combination, excess SSc mortality in the detected clusters accounted for 79.0 and 66.2%, respectively, o f the excess deaths among white and black males across the whole South east. Conclusion. Elevation of SSc mortality rates in the Southeastern US results from local clusters of concentrated mortality. These clust ers may be artifacts of regional variation in death certificate qualit y. If not, distinctive environmental factors in these areas may provid e new insights into the etiology of SSc.