Yesterday I ran our gender source terms through Usagi. As expected, male & female matched and unknown did not. We have decided to use the OHDSI supported PCORNet concepts for unknown and other.
Usagi produced a csv that mapped the source concept to the standard concept and provided the standard concept_id. My question: why did Usagi give me all the attributes for a standard concept? Isn’t the concept_id enough data since the concept_id is the PK to the CONCEPT table which contains all the attributes? I’m asking because yesterday when my source concept of unknown didn’t match, I was a lemur and filled in all the attributes for the appropriate PCORNet concept. Thinking about it on my drive home I realized that all I need to add to the csv file is the target concept_id. Would I be missing something, breaking something, or worse if I only add the concept_id of a supported OHDSI concept to the csv for the CONCEPT_RELATIONSHIP table?
Hi @MPhilofsky!
The attributes are in the mapping file for silly technical reasons. There is no harm in not having them in the CSV file. Maybe if I have time I’ll start work on Usagi v2.0 (or maybe it should be called v1.0), which will use a very different technical architecture and will no longer require these fields.
Note that there’s also an option to export to the source_to_concept_map table format, which is a bit cleaner.