Considerations for pharmacoepidemiology study design

Series: causal inference

Right now, an unorganized list of concepts I don’t want to forget about. Pharmacoepidemiology, general Rapid changes in the natural history of a disease or changes in treatment decisions will mean strong effects in relation to calendar time. Example: The management of COVID-19 improved throughout the pandemic independent of medication use. If you wanted to assess the comparative effectiveness of a drug, you would need to control for calendar time.

Causal Inference - a language & symbols cheat sheet

Series: causal inference

The “notes” category indicates that much of this post is my summary and interpretation from sources I’m using to learn about a topic of interest. As such, most of the following are not my original thoughts. Learning hard stuff is hard, and it continues to be hard long after you would like it to have become easier. It’s that ‘I know this but I also don’t KNOW this" feeling.

Notes on Mediation Analysis, Part II

Series: mediation analysis

I’m spending the last semester of my masters working on an independent study project on causal mediation analysis. It seemed a perfect opportunity to try this Learning in Public thing for the first time. I’m currenlty adding to this post on an almost daily basis. This is Part II in a two-part series. Part I covers general mediation concepts and terminology. This is where I dive into some of the more complex issues in mediation analysis.

Notes on Mediation Analysis, Part I

Series: mediation analysis

I’m spending the last semester of my masters working on an independent study project on causal mediation analysis. It seemed a perfect opportunity to try this Learning in Public thing. I’m currenlty adding to this post on an almost daily basis. Part I covers general mediation concepts and terminology. Part II covers more complex issues in mediation analysis. Corrections or requests welcome in the comments section! Note: most of these notes are not specific to analysis in any language, but if I use code samples they will be in R.