EPID 785R

This course focuses on the foundational relationships between study design and regression for doctoral students in Epidemiology. The course starts with a focus on study design principles, using the target trial emulation framework. The course then focuses in depths on applied regression modeling, including:

• Regression as a toolkit for descriptive, predictive, and causal inference
• Generalized linear models, variance estimation, conditional versus marginal adjustment
• Flexible and penalized regression
• Regression for outcome dependent sampling designs
• Models for time-to-event outcomes, including nonparametric estimators, parametric, and semiparametric approaches

This course adopts the philosophy that regression machinery is fundamentally descriptive in nature, at the level of the statistical model. The scientific use of a regression model, whether descriptive, predictive, or causal, is determined by the question being asked and the assumptions brought to the analysis. This course will demonstrate how and why.