DEVELOPING INNOVATIVE STATISTICAL MODELS TO PREDICT CANCER INCIDENT CASES AND OUTCOMES AT THE SMALL-AREA LEVEL

Cancer incidence and outcomes vary across populations and geographies over time. Understanding the extent and causes of disparities is essential to cancer control. This thesis developed statistical models to predict cancer incident cases and outcomes at the small-area level. Studies described geographic and temporal patterns, examined factors explaining the variation, and performed projections in geographic areas. My thesis provides novel clinical contributions and methodological framework in terms of identifying factors that explain cancer disparities and quantify the existing and future burden of cancer at the small-area level, essential for cancer prevention and control. These models can be implemented in other cancers and diseases.