Course Description
Teaches advanced epidemiology students to synthesize biologic content, epidemiologic designs, and statistical analysis into a coherent research plan. The didactic portion of the course will cover causality in epidemiology, scientific method, including hypothesis testing and interval estimation, sample size estimation, sampling strategies, and advanced regression models.
Athena Title
Causal Inference in Epid
Prerequisite
(EPID 8010 or EPID 8020 or EPID 8040) and (BIOS 8010)
Semester Course Offered
Offered fall and spring
Grading System
A - F (Traditional)
Course Objectives
Students will be able to: 1. Build an argument for cause and effect between an epidemiologic exposure and outcome. 2. Write a statistical analysis plan, including specific aims, causal diagrams, data descriptions, and details of regression models. 3. Design and perform model building using causal regression methods. 4. Review and critique analytic approaches used in the published literature on observational epidemiologic studies.
Topical Outline
The course will integrate four overarching themes: cause and effect, identifiability conditions for causal inference, bias and random error, and regression model building. 1. Cause and effect a. Philosophical foundations and history of causality in science b. Component causes c. Counterfactual theory d. Causal modeling versus association modeling (i.e., marginal structural models) e. Randomized experiments 2. Identifiability conditions for causal inference a. Exchangeability b. Positivity c. Consistency d. Target trial 3. Bias and random error a. Causal diagrams b. Confounding c. Selection bias d. Measurement bias e. Random variability 4. Regression Model Building a. Choice of regression model for analysis b. Inverse probability weighting and marginal structural models c. G-estimation and structural nested models d. Propensity score e. Instrumental variable estimation f. Doubly robust estimators g. Longitudinal studies h. Variable selection and machine learning
Syllabus
Public CV