We are using the following query in the VOCAB to make the hop from MeSH to RxNorm:
SELECT A.CONCEPT_CODE source_concept_code, A.CONCEPT_NAME source_concept_name, B.CONCEPT_CODE concept_code, B.CONCEPT_NAME concept_name, B.CONCEPT_ID concept_id, B.CONCEPT_CLASS_ID concept_class_id
FROM CONCEPT A, CONCEPT B, CONCEPT_RELATIONSHIP CR
WHERE A.VOCABULARY_ID = 'MeSH' AND
CR.CONCEPT_ID_2 = A.CONCEPT_ID AND
B.CONCEPT_ID = CR.CONCEPT_ID_1 AND
B.VOCABULARY_ID = 'RxNorm' AND
CR.RELATIONSHIP_ID = 'Mapped from'
If you use this map in Vocab v20160311, rituximab gets mapped to MeSH tag of C092560 not the D000069283 being used in PubMed. Thus we would not be able to take this PubMed article in LAERTES with the drug rituximab.
I’d like to understand this better what is going on here. Particularly if you understand the differences between C’ and D’ codes.
You have to ask the NLM why they change things. We are just putting things in the way they are.
The C codes are generally Supplemental Concepts, and the D codes Main Headings (Descriptors). Supplementals are usually more detailed, the Mains more general. You can read about it here.
So, they promoted Rituximab from Supplemental to Main. They just can do these things. Next release will have it the way you need it.
When updating the Vocab, will you be including all drug Descriptors and Supplementary Concept Records with maps to RxNorm. I’m using rituximab as an example here and it is probably not the only one in this situation.
And as I always ask I need to know when you’ll be creating this map. The timing will help me make some decisions on what to do in LAERTES for this item.
Actually, we are doing lazy incorporation: Only codes that come with a map will be incorporated. If Rituximab is mapped to RxNorm it’s in. If not - that would be a problem. Because MeSH is gigantic, and it would be very hard to play secondary NLM. They have a $ multi-million budget.