Semiparametric spatial frailty modeling for survival data based on copulas.

In this study, we describe a frailty model for spatially correlated survival data based on the class of Archimedean copulas in the spatial regression, assuming that counties are allowed to have varying sizes of spatial structure. In the spatial frailty model, a Bayesian nonparametric prior for a baseline survival function is provided to improve model flexibility. The proposed approach was evaluated by analyzing survival data for acute myeloid leukemia in the North West of England. [ABSTRACT FROM AUTHOR]