Abstract:
|
Worldwide, breast cancer is the most common malignancy in women. With an early diagnosis, survival rates can increase for women with breast cancer. Modern statistical methods allow us to analyze these district-level incidences of breast cancer in Portugal. In population surveillance research, age-period-cohort (APC) models are widely used to examine underlying trends of disease incidence and mortality and these models have been recently extended to allow spatially-varying effects. Here, we apply spatial APC models to estimate relative risk and temporal trends of breast cancer incidence in 20 districts in Portugal, which include 18 contiguous districts and two island districts. Our model allows us to report on country-wide trends, but also to investigate geographic disparities between districts. Estimation takes place in a fully-Bayesian paradigm by applying the Adaptive No-U-Turn Hamiltonian Monte Carlo sampler via Stan, which frequently outperforms other Monte Carlo samplers in terms of efficiency.
|