While establishing record delta from source to OMOP, it looks like some patients were identified as duplicates in source data, therefore merged in source data, and the duplicate ID in source record was thereafter deleted. Detecting the deletions is easy in source, but how have folks been representing this in their CDM? Do they just remove/merge a person_id? If so, how do you represent merges in a way that travels to other tables? If it’s a matter of doing an update across all tables, that’s fine too, but I’m a bit uncomfortable doing an update on an OMOP record, and then not having record of the original in OMOP.
OMOP is going to be a part of our larger data lake/fabric, so being able to trace and update records downstream from OMOP is important. This may not always be possible in real time, so it’s important to have the process downstream from OMOP be able to search for deleted and changed records within OMOP tables. Is this a good use case for metadata table usage? If so, what has the metadata update process looked like for you with regard to deleting/updating person records?