OBJECTIVES: To account for the non-stationarity of relationships in space, aspatial regression can be supplemented with geographically-weighted regression (GWR), whereby the regression model is fitted within local windows and each observation is weighted according to its proximity to the center of the window. This study aims to conduct regression analysis in a spatial context to assess the local impacts of putative factors on late-stage diagnosis of prostate cancer in Florida during the period 2001-2007. METHODS: A logistic regression was performed aspatially and at the nodes of a 5 km spacing grid overlaid over Florida and using all the cancer cases within a radius of 125 km of each node. Each observation was weighted as a function of its proximity to the center of the window (bisquare adaptive weight function). Covariates included age, race, marital status, smoking, type of health insurance and diagnosis facilities, presence of comorbidities (healthy (no comorbidity), average (1-2 comorbidities), above-average), census-tract median income and presence of farmhouse, year of diagnosis, county-level provider-to-case ratios. RESULTS: Variables increasing the likelihood of late-stage diagnosis included having 1 to 2 comorbidities (odds=1.697) and more than 2 comorbidities (odds=3.963), smoking (odds=1.283), being African American (odds=1.199) and living in census tracts with farmhouses (odds=1.124). Having private insurance (odds=0.533), having public insurance (odds=0.470), being married (odds=0.787) or diagnosed in a for-profit facility (odds=0.886), as well as living in census tracts with high income (odds=0.994) reduces the likelihood. CONCLUSIONS: There are significant spatial associations between late-stage prostate cancer incidence and observed individual, socioeconomic, behavioral, environmental and demographic factors in Florida. This emphasizes the need for local strategies and cancer control interventions to reduce the percentage of late-stage diagnosis and ultimately eliminate health disparities.