Joint Modeling of Disease Pairs

Aim: Exploring the spatial patterns in joint distribution of incidences of two diseases. Material and method: A Poisson-Binomial regression model was used in analysing hospitalisation counts data in Hungary, 2008. Model parameters were estimated by MCMC implemented in WinBUGS. Results: Spatial patterns of laryngeal and hypopharyngeal cancer differ significantly from that of their ratio. Conclusion: The Poisson-Binomial model proposed here might help clarify us the different spatial dependencies of the sum and the ratio of incidences of two diseases.