OHDSI Home | Forums | Wiki | Github

A need for Standard drug categories

Vocabulary Users
#1

Hi everyone!
In a mapping process, the Vocabulary team regularly encounters source drugs that do not specify any ingredients, but which determine a certain pharmacological category. In such a case we cannot use the RxNorm or RxNorm Extension whilst ATC drug categories look suitable except the fact they are classifications, so we cannot use them.
However, entries that we lose when targeting the concept id to 0 could be useful for identifying certain patients, for example:

  • Unspecified insulin can be an indicator of patients with Diabetes Mellitus.
  • Unspecified low-molecular-weight heparin gives us people at risk for bleeding.
  • Unspecified interferon indicates patients on immunotherapy and with altered immunity.

Dear experts, is there hope that categories of drugs will become standard one day?

@Christian_Reich
@abedtash_hamed
@aostropolets
@Alexdavv
@Dymshyts

3 Likes
Drug concepts with lower granularity (medicine groups)
Missing Concepts Discovered relating to Allergies
Missing Concepts Discovered relating to Allergies
#2

They are Classifications per se. This is more about why don’t we admit Classifications to be used as event_concept_id in general. As long as we don’t, why the Drug domain should be an exception?

How many usecases are there? Aren’t we ok having the workarounds of Procedure/Observation concepts?

ID CODE NAME CLASS CONCEPT VALIDITY DOMAIN VOCAB
4215993 39543009 Administration of insulin Procedure Standard Valid Procedure SNOMED
40481876 443464003 Low molecular weight heparin therapy Procedure Standard Valid Procedure SNOMED
40492039 448884005 Injection of interferon Procedure Standard Valid Procedure SNOMED
#3

Agree with @Alexdavv. We haven’t really pushed the promotion of ATC to Standard for two reasons:

  • We are still working on the mapping to RxNorm (and waiting for some input from the WHO)
  • Not all drugs are in ATC, and not all ATC have drugs attached to them.

At some point we should just bite the bullet.

#4

We are facing this issue trying to map cancer data for COVID-19 patients. The CCC19 registry that is being captured via REDCap chart abstraction often only captures ATC or Hemonc.org classification concepts. Because the emphasis is on rapid data capture, RxNorm ingredients cannot always be specified. It would be great if this problem could be solved.

t