OHDSI Home | Forums | Wiki | Github

Adding new ‘RxNorm extension clinical drug’ concepts: concept names without density information for injective drugs are allowed?

Vocabulary Users
#1

In Korea, we are currently working on adding RxNorm extension concepts for drugs that do not matches to existing RxNorm (or RxNorm extension) concepts.

In Korea, every drug is mapped to 9 digit ingredient code.

For example, Korean drug code ‘052700241’, ‘Lincomycin Inj. 300 mg’, is mapped to ingredient code ‘184230BIJ’.

Because this ingredient code include information about ingredient name, and quantity/unit, route and type, and not density (or total volume), both drug of ‘lincomycin 300 mg/ml Injection (total 1ml)’ and ‘lincomycin 150 mg/ml Injection (total 2ml)’ are mapped to the same ingredient code ‘148230BIJ’.

So, it is easier for us to make new RxNorm clinical drug name like ‘lincomycin hydrochloride 300 mg injective solution’ than to calculate the density.

In some injective drugs of powder form, the clinical drug name should be like the expression above. Then, in drugs of liquid form, is same expression possible? Or, is it better to calculate the density for those drugs?

#2

@Sooyeon_Cho:

Very interesting.

Any chance you can use the process defined here? It is very hard to keep all the logic right, and have a seamless integration with RxNorm. We have a whole QA process to make sure it works. We are very happy to help you convert your data into the input tables DRUG_CONCEPT_STAGE, INTERNAL_CONCEPT_STAGE, RELATIONSHIP_TO_CONCEPT and DS_STAGE.

Yes, this will be tricky. Because in RxNorm 1 mL of 300 mg and 2 mL of 150 mg are not hte same thing. However, we have incorporated 300 mg solution for injection as a Clinical Drug, even though it violates the rule that dosing information for liquids should a concentration (I think you call it “density”), rather than a total mass.

Do ‘300 mg/ml Injection (total 1ml)’ and ‘lincomycin 150 mg/ml Injection (total 2ml)’ have the same drug codes in your data?

Can you provide some example data?

#3

We’ve been reviewing those documents and sql queries, they were quite difficult and complicated, but well organized. Thank you.

In fact, we have 3 brand drugs of lincomycin 300mg from different company, but same form as ‘300mg/ml’. It was only to illustrate an example of a potential problem.

The cases of ‘different concentration having same ingredient code’ are rare.
Here is other right example of two drug.

. Drug / Drugcode / Igdrcode

  • Mucosten Inj 300mg/2mL / 646001141 / 101802BIJ
  • Mucosten Inj 300mg/3mL / 646001130 / 101802BIJ

. Igdrcode / Igrd info

  • 101802BIJ / acetylcysteine 300mg Injection

Though they have different ‘drug code’, we’ve thought mapping Korean ‘ingredient code’ to RxNorm is easier because about 70,000 drug codes have only 12,000 ingredient codes.
However, if we add Koren drug codes to OMOP voca, we have to fill ‘Drug strength’ table and total volume information, too.

So, we’ll follow the ‘concentration for liquids’ rule as it possible.

#4

@Sooyeon_Cho:

Those are no problem to represent.

These violate the RxNorm rule. But we can put them in RxNorm Extension, and it will work with the OMOP system.

That’s pretty normal. If you could map the Korean ingredients, we could probably do the rest and bring in your drug data properly. Just like the Canadian, UK, French and German drugs.

We can help you too with some parsing, if you need to get that information from strings.

Let me know.

t