Diabetes prevention is a critical area of focus in chronic disease. Increasingly, epidemiologic studies also seek to understand potential causal relationships and pathways between non-individual level predictors and diabetes. This dissertation aims to contribute to the growing body of literature examining non-individual and modifiable factors associated with diabetes, in the hopes of informing efforts to create environments in which it easier for individuals to make the changes needed to live a healthier life while managing their diabetes. First, a legal mapping study is conducted to describe key features and variations in how states implement copayment caps on insulin. Several important features emerged, which can be summarized by how each state handled the extent of the scope of caps, and specificity in the language used to describe these caps. A minority of states included provisions for price caps to be applied to an emergency supply of insulin (n=4), an out-of-pocket cap for insulin provided via an insulin affordability program (3 states), or a cap on out-of-pocket costs for devices or supplies needed to manage diabetes (6 states). Specificity of language was particularly noticeable regarding handling of a deductible, whether all types of insulin would be affordable on health coverage plans, and whether the cap applied to one prescription or the aggregate total of all prescriptions. In the subsequent two analyses, I examine spatial non-stationarity in the relationship between diabetes prevalence and variables related to the food environment, access to physical activity, and access to primary care in Georgia. This analysis was conducted at both the census tract and the county level. Both analyses find evidence of spatial autocorrelation in the distribution of adult diabetes prevalence across Georgia. Clusters of high and low prevalence were identified. Walkability remains a significant predictor of diabetes in Georgia, even when controlling for a wide set of demographic variables. At the census tract level, there is evidence to suggest this association between walkability and diabetes prevalence changes in magnitude across the state. At the county level, a global regression model may sufficiently capture the association. Key predictors for greenspace, access to care, and the food environment were not significant predictors of diabetes prevalence in Georgia, when controlling for demographic characteristics. Much of the variation in diabetes prevalence across Georgia can be statistically explained by these demographic characteristics alone.