Bayesian bivariate spatial shared component model: mapping breast and cervical cancer mortality in Southern Brazil

Abstract

Spatial analysis techniques are used in the data analysis of ecological studies, which consider geographical areas as observation units. In this article, we propose a Bayesian bivariate spatial shared component model to mapping the breast and cervical cancer mortality in Southern Brazil, based on the models introduced by Knorr-Held and Best (2001) and Held et al. (2005). Markov Chain Monte Carlo (MCMC) methods were used to spatially smooth the standardized mortality ratios (SMR) for both diseases. Local Indicator of Spatial Association (LISA) was used to verify the existence of spatial clusters in specific geographical areas. This study was carried out using secondary data obtained from publicly available health information systems.