Instrumental Variables
This seminar has 3 goals. The first is to provide a conceptual overview of the selection bias problem in observational comparative effectiveness studies. The second is to introduce practice pattern-based instrumental variables methods for eliminating selection bias as well as methods for testing these models. The last is to present an example of forthcoming research that uses these techniques to compare alternative medications for type 2 diabetes. Click here to view the presentation.
Propensity Scoring
This seminar has 2 goals. The first is to provide a conceptual overview of propensity score matching as a methodological approach for reducing the selection bias problem in observational studies. The second is to introduce the implementation of a propensity score matching approach using R, a free software environment for statistical computing and graphics. Examples of ongoing research in primary care and hospital settings are used for demonstration purposes. Click here to view the presentation.