Analyzing factors associated with cancer occurrence: A geographical systems approach

Geographical Information Systems (GIS) is defined as an automated set of functions that provides professionals with advanced capabilities for the storage, retrieval, display and analysis of geographically located data. This paper aims to show a GIS based approach to analysis of registry-based cancer incidence data. Case data were obtained from the Pennsylvania Cancer Registry for 1993-1997 period and corresponding population data were obtained from the National Center for Health Statistics. The incidence rate was calculated as the number of cases diagnosed per 100,000 persons living in any given area. The rates were age-adjusted (direct method) against the 1970 U.S. standard population. GIS methods such as clustering, autocorrelation were used to analyze prostate cancer incidence spatially. A lower incidence area in the northeast was observed. The significant variables that are associated with the incidence in this period were percentage of households with less than 2 bedrooms (p=0.022) and distance to toxic sites (p=0.007). These two factors were thought to represent different dimensions of urbanicity. GIS analyses for ecological analyses of health related outcomes is feasible and useful in generating hypothesis and identifying areas for intervention. [Turk J Cancer 2004;34(2):67-70]