OHDSI Home | Forums | Wiki | Github

Concept relationship issue - NDC to standard concept


(Shinjinaka) #1

Hi,
I used the concept_relationship to map NDC to standard concept.
When I compared the source description and mapped NDC description, the mapping seems fine.
However, the NDC description and “Maps to” concept_name seems not make sense.
Please check the attached file.

OHDSI_concept_relationship_need_fix.xlsx (12.0 KB)


(Melanie Philofsky) #2

You should post this to the Vocabulary Users category. You’ll get a quicker reply from knowledgeable persons.


(Alexander Davydov) #3

Hi @shinjinaka,
This is probably an NDC code reuse issue.
The problem is we still cannot identify 2 reliable sources with a different meaning.
Your descriptions match the current ones in the NDC source, so we will need to update the mappings.
Can you please provide MIN and MAX event dates associated with these codes/descriptions?
Is this a full list?


(Shinjinaka) #4

This data is coming from outpatient prescription database. Physician enter prescription, and system picks up a NDC associated with the medication description/code. Having the same description that OMOP has gives us confident to use the NDC for mapping.

The problem is “Maps to” RxNorm description doesn’t match with “Mapped from” NDC description.

Do you need MIN and MAX event dates for validation purpose? Although I can provide the MIN and MAX event dates associated with these codes/descriptions, this data is not dispensing records. Each description tight with one NDC. Thus, I am not sure you should use the data to determine valid starts date and end dates.

Please let me know if you still need the dates data.


(Alexander Davydov) #5

Exactly, the current mappings for these NDCs with current descriptions are wrong, except those that mapped to ‘Multivitamin preparation Oral Tablet’.

Yes, please. We will not use them as a reference at this point. Just to find any patterns.


(Shinjinaka) #6

Min and Max were added.
OHDSI_concept_relationship_need_fix.xlsx (15.0 KB)

Can we have a concept_id with “Multivitamin with Iron preparation Oral Tablet” or “Multivitamin with Mineral preparation Oral Tablet”? Although some of the “multivitamin” product contains Iron (i.e. Centrum), I feel Iron won’t be under the multivitamin categories and should be separated. Iron should be under “Mineral” categories.


(Alexander Davydov) #7

Thank you.

Usually, new ‘RxNorm Extension’ concepts are created to reflect the source concepts of different drug vocabularies. ‘Multivitamin preparation’ is something different since we may use it as a target concept to map various vitamins where the composition is not clearly defined. In theory, more detailed vitamin concepts might be created. But the general feeling is that the vitamins, whether containing minerals or not, are out of the scope of scientific use cases. Please let us know if you think different.


(Shinjinaka) #8

I got a response from a researcher as below.

“Both can be essential for treatment and they are not the same thing. As you have each individual item, obvioulsy this is not going to make you lose data either way. If your framework is very granular, I would separate, if very gross then maybe not necessary. So I would be prudent and leave them separate.if you have no compelling reason not to split.”

Can we create concept like “Multivitamin with Iron preparation Oral Tablet” or “Multivitamin with Mineral preparation Oral Tablet”?


(Christian Reich) #9

What is actually the question? Not losing data is not a use case. What is the researcher trying to determine? The effects of multivitamin with iron vs multivitamin with mineral? Which effects?


(Shinjinaka) #10

Sorry that I should have include the actual question.

My question:
"I have a question about how do you think about multivitamin data. As a pharmacist, I consider multivitamin as vitamins (A,B,C,D,E,K), and doesn’t include any minerals.
Currently, in OMOP model, “multivitamin preparation” is including both vitamins and minerals.
I prefer separating vitamins and minerals because minerals tends to interact with certain medications (ie. fluoroquinolones antibiotics).
However, I don’t know how important to distinguish this intern of research. Do you feel like it will be nice to distinguish vitamins and minerals, or it doesn’t matter at this point. "

Response from researcher:
“Both can be essential for treatment and they are not the same thing. As you have each individual item, obviously this is not going to make you lose data either way. If your framework is very granular, I would separate, if very gross then maybe not necessary. So I would be prudent and leave them separate.if you have no compelling reason not to split.”

I mentioned “multivitamin with iron” earlier, because that’s something that came to my mind first, but there are other minerals such as zinc, calcium, copper, etc. I have second thought about it because there are so many combination of these. Thus, as start, can we create a concept for "Multivitamin with Mineral preparation Oral Tablet” and map multivitamin with mineral preparation ( i.e. THEREMS TABLET
44977082 NDC) to that?


t