×

Hierarchical models for detecting geographical effects in cancer incidence and survival. (English) Zbl 1135.62086

Summary: This study is motivated by the importance of assessing small-area variation for the development and implementation of medial and educational interventions to reduce disparities in breast cancer survival. The data were collected state-wide and post-stratified to the county level by the Iowa SEER program. We propose a Bayesian hierarchical model for Weibull distributions by incorporating conditional autoregressive priors for the transformed rate parameters to analyze spatial-temporal effects on breast cancer survival. The N. E. Gibbs, W. G. Poole and P. K. Stockmeyer algorithm [ACM Trans. Math. Softw. 2, 322–330 (1976; Zbl 0345.65014)] on the sparse adjacency matrix enhances the efficiency of the Gibbs sampler in the simulation study. Results of breast cancer survivals for aged 65 or older women in Iowa are presented. Comments and further discussions are given.

MSC:

62P10 Applications of statistics to biology and medical sciences; meta analysis
62F15 Bayesian inference
62N01 Censored data models
62J12 Generalized linear models (logistic models)
65C60 Computational problems in statistics (MSC2010)

Citations:

Zbl 0345.65014
PDFBibTeX XMLCite