OHDSI Home | Forums | Wiki | Github

How to handle vocabulary updates for existing data?

When there is a new Athena release, How to handle existing mappings of data ?

What if few concepts were mapped to something which is not present or updated to in the latest vocabularies ?

One can’t afford to do full ETL every time CDM or Vocabularies or Source data changes.
What procedure generally you guys follow?

@rohans:

It’s a good question, and we need to write it down in the documentation better. But bottom line:

To your question about mapping to something that is not valid any longer: You always keep the vocabulary you used for the latest ETL refresh of the data. If you download a new vocabulary you have to re-ETL.

And generally, this is good enough. Unless you are waiting for a vocab release because it contains fixes you need, the difference between two subsequent releases is marginal. All concepts are stable, unless they are deprecated or upgraded, which happens to a small number of them each release. So, technically you can be stale after a new vocab comes out, but in reality this problem is not very significant. But if you wait for a longer time (a year or more) then you will run the chance that a study or analyses created outside your own dataset will no longer work with your data.

Thanks for the response @Christian_Reich

t