Applying Spatial Copula Additive Regression to Breast Cancer Screening Data

Breast cancer is associated with several risk factors. Although genetics is an important breast cancer risk factor, environmental and sociodemographic characteristics, that may differ across populations, are also factors to be taken into account when studying the disease. These factors, apart from having a role as direct agents in the risk of the disease, can also influence other variables that act as risk factors. The age at menarche and the reproductive lifespan are considered by the literature as breast cancer risk factors so that, there are several studies whose aim is to analyze the trend of age at menarche and menopause along generations. Also, it is believed that these two moments in a woman’s life can be affected by environmental, social status, and lifestyles of women. Using the information of 278,282 registries of women which entered in the breast cancer screening program in Central Portugal, we developed a bivariate copula model to quantify the effect a woman’s year of birth in the association between age at menarche and a woman’s reproductive lifespan, in addition to explore any possible effect of the geographic location in these variables and their association. For this analysis we employ Copula Generalized Additive Models for Location, Scale and Shape (CGAMLSS) models and the inference was carried out using the R package SemiParBIVProbit