Advisors: Adrian Dobra and Jon Wakefield
Understanding the relationships between disease incidence and risk factors such as demographic characteristics, life style factors, and environmental contaminants is a central goal in public health and epidemiology. Often outcomes and risk factors are measured at specific locations or at particular times. We present flexible Bayesian models for spatial and temporal data to address important public health questions in two examples. In the first example, we consider low birthweight and preterm birth along with three risk factors in North Carolina. We introduce a new model that uses binary trees together with graphical log-linear models to account for spatial heterogeneity in the strengths of associations between the outcomes and risk factors (different parameter values for the same set of interactions) and spatial heterogeneity in the interaction structure itself (different sets of interactions by location). In the second example, we use Bayesian age-period-cohort models to fit and forecast rates for breast cancer and lung cancer incidence using data from the Washington State Cancer Registry. We also derive an aggregate mean model for estimating the individual level effects of tobacco use and cancer screening using data from the Behavioral Risk Factor Surveillance System.