We carried out an ecological study in which the spatial distribution of 979 new tuberculosis cases between 2000 and 2005 were analyzed. The risk was estimated using the local empirical Bayesian method. The local indicators of spatial association were calculated to evaluate autocorrelations of incidence between adjoining districts. We found that the tuberculosis cases were heterogeneously distributed between districts, and it was possible to identify regions in which there was a high risk of becoming ill.