Handling multiple transformation requests

Implementers
1

Dear all, I am not sure if this scenario applies to any other institution out there, or if this scenario even applies at all.

At our university, we have been receiving requests from PIs to transform their current data into the OMOP CDM. So pretty much, think about multiple people coming to your office, giving you a bunch of data and asking it to transform it to OMOP CDM so they can easily share it, explore it with the available tools, etc…

Do I need to create a new database each time I have a request like this and create all tables and import all concepts?

or

Is it correct to import everything keeping a log of the persons, visits and providers and then put that into a cohort in the cohort/cohort attributes tables? If so… could someone please expand on how to accomplish this in the best way possible?

or

Somehow different.

Thanks in advance. Have a great week.

2

Most institutions convert their entire dataset into CDM first and then use “cohorts” to provide data to individual researchers within a given organization.

So if all requesting researchers use some master data to get their cohorts data, it is best to convert the master data.

3

I figure that is what usually happens, but here we don’t have a centralize data warehouse and each researcher has his own cohort of patients.

Thanks for the reply.