ABSTRACT
The data available to address important questions about harms and benefits to human health are often rather limited. Ethical concerns, logistical constraints, and resource limitations often mean that only observational studies can be conducted to address a given question about a possible determinant of health. And the data from observational studies are indeed prone to be limited in fundamental ways. Thus, while statistical methodologies play a central role across the health sciences, they become acutely critical in the observational study realm. As will be showcased here, Bayesian statistical methods can be applied to infer as much as possible about the health question at hand, while acknowledging the limitations of the available data.
