Study Definitions

[sebbacon]

The aspiration of a study definition is that it should be a single place to describe the data an analyst needs to run their analysis.

It should:

• Not require them to think about the differences between backends
• Be easy to review - compact and expressive, diffable
• Have a happy path where it can be generated via a web UI
• Be composable - if someone QAs a new variable, it should be easy to package this up and make it available via codelists (Codelists and variable logic repo #3)