In this paper we propose a hierarchical Bayesian method to estimate the relative risk for female breast cancer at the municipality level in the province of Trento. To model the relative risk we use the so called convolution model for count data, which takes into account random and spatially correlated random effects (un- correlated and correlated heterogeneity). The method is adopted to obtain reliable estimates of the relative risk in those areas where the low number of observations makes the rough relative risk estimates unstable (where rough means a relative risk estimated using only observations and target population data at a given area level). This Bayesian method can be applied to a wide range of problems. Here we apply this method to epidemiological data aiming at an on-line computational solution.
Bayesian Hierarchical Model for Small Area Disease Mapping: a Breast Cancer Study
MARCHETTI, STEFANO;
2010-01-01
Abstract
In this paper we propose a hierarchical Bayesian method to estimate the relative risk for female breast cancer at the municipality level in the province of Trento. To model the relative risk we use the so called convolution model for count data, which takes into account random and spatially correlated random effects (un- correlated and correlated heterogeneity). The method is adopted to obtain reliable estimates of the relative risk in those areas where the low number of observations makes the rough relative risk estimates unstable (where rough means a relative risk estimated using only observations and target population data at a given area level). This Bayesian method can be applied to a wide range of problems. Here we apply this method to epidemiological data aiming at an on-line computational solution.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
