OHDSI Home | Forums | Wiki | Github

One-to-many mapping for drugs_exposure

Hi guys,

I’m doing a code mapping from ATC to RxNorm and RxNorm Extension based on the mapping from Athena. I found some of the ATC concepts map to multiple RxNorm concepts. Here is an example:

21600006 sodium monofluorophosphate; oral
maps to
19068715 monofluorophosphate
42899013 fluorine

In this case, should I create two drug_exposure records for one source record? I think it’s ok for the “Condition” domain according to the previous post ICD10CM to SNOMED. But not sure if the same applies to “Drug” domain as well.

Thank you very much and please let me know if I get this wrong.

hi @Dawei_Zhou ,

That is the way it works. It is true for all domains since to populate columns such as drug_concept_id or condition_concept_id you always use the concept_relationship table and the Maps to relationship to get the Standard concepts we should populate

1 Like

Thank you, Jose. I get it now.


following this thread, two questions:

  1. when using the ATC we are working on the domain of “class”. In the documentation (OMOP CDM v6.0), when we consider the description of DRUG_CONCEPT_ID have the following at the end: “Note: If only the drug class is known, the DRUG_CONCEPT_ID field should contain 0”
    Q1) Where should we consider the ATC mapping? We though it would be on DRUG_CONCEPT_ID

  2. When using ATC the possibility of analysing the dose seems to be lost in OMOP.
    Q2) Is this a correct assumption? How could we consider the dose when our source codes are not mapped to RxNorm but on their original format have the dose?

Thank you,