Undergraduate Honors Thesis Presentation: Investigating Spatial, Environmental, and Socioeconomic Risk Factors of Breast Cancer Incidence in Illinois

Geographic disparities in cancer incidence often reflect underlying environmental exposures and socioeconomic inequalities. This study employs a Bayesian conditional autoregressive (CAR) framework to examine spatial variation in observed breast cancer counts across Illinois ZIP codes. Markov chain Monte Carlo (MCMC) sampling provided posterior estimates of both structured (spatial) and unstructured random fluctuations. The roles of geographic areas, proximity to Toxic Release Inventory (TRI) facilities, high-risk ZIP codes for pediatric lead poisoning, and Social Vulnerability Index (SVI) are investigated. Results consistently highlighted significant roles for environmental and socioeconomic covariates. Overall, the CAR-based approach accommodated for spatial dependence and offered refined estimates of relative risk, guiding public health authorities toward targeted interventions in potentially vulnerable neighborhoods. This research underscores the importance of spatially explicit models in unveiling complex interactions between place, pollutants, socioeconomic status, and cancer outcomes, while demonstrating the practical utility of hierarchical Bayesian methods for epidemiological analysis.

Advisor: Debashis Mondal