I’ll add a few more bars here since I know this Frank guy.
Frank is working with a team who is trying to think through the use cases their CDM needs to support. It’s new and intoxicating to think of all the amazing things they can do research-wise. One thing they want to get a feel for are the common types of studies or analyses that other people run with OMOP CDMs.
So yes, characterization, estimation and prediction would all be example constructs. But more specifically, if you were lining up a list of “things your CDM must do to be a viable CDM”… what studies/analyses would you prioritize? This harks back to some of the discussion @Patrick_Ryan @jennareps and others have discussed around “what’s the minimum amount of data needed to run a study”.