OHDSI Home | Forums | Wiki | Github

Maps to vs. Maps to Value; single concept split into two

Vocabulary Users
#1

We noticed concept 4172372 Remission phase [Measure Value] {Qualifier Value} Snomed has a “Maps to” relationship to concepts but is in the “Meas Value” domain. The conditions and disorders are a “Maps to”.

This affects multiple condition concepts that were updated and split into two.

Now Non-Standard:
alcohol use, unspecified, in remission
cocaine use, unspecified, in remission
cannabis use, unspecified, in remission
other stimulant use, unspecified, in remission
opioid use, unspecified, in remission

Standard:

987891000000105/46286594 Problem related to lifestyle (finding) Observation Snomed
4172372 Remission phase [Measure Value] {Qualifier Value} Snomed

46975003 / 4264766 Organic mental disorder caused by cocaine Condition Snomed
4172372 Remission phase [Measure Value] {Qualifier Value} Snomed

77355000/4300092 Cannabis-induced organic mental disorder Condition Snomed (Disorder)
4172372 Remission phase [Measure Value] {Qualifier Value} Snomed

50026000/4178114 Organic anxiety disorder caused by psychoactive substance Condition Snomed (Disorder)
4172372 Remission phase [Measure Value] {Qualifier Value} Snomed

14784000/4032799 Opioid-induced organic mental disorder Condition Snomed(Disorder)
4172372 Remission phase [Measure Value] {Qualifier Value} Snomed

Is this correct or is an update needed?
Thank you, Tina

#2

Hello, @Tina_French

Thank you for your question.
Yes, this requires fixing. Meas Value concepts should be moved back to the Observation domain.
Provided mappings also should be fixed as Qualifier Value concepts don’t carry the semantics of clinical events and we should not use them to map conditions.
Unless these fixes are implemented I would propose you use concepts from the hierarchy of Disorders in remission to map these non-standard concepts.

Regards,
Masha.

#3

Hi @m-khitrun,
Thank you for responding so promptly. I wanted to clarify, the mappings you see at the bottom of my message are the new standard concepts for these conditions after the last vocab update. When you say “they need to be fixed”- is this something OHDSI will do? Our ETL expert, Katherine Simon, created a chart that outlines the issue very well- I have attached it. Please let me know if I need to provide additional information. Thanks again, Tina
Copy of Maps to Value Vocabulary Issue.xlsx (47.7 KB)

#4

Hi, @Tina_French

Thank you for the chart provided.

Yes, our Vocabulary Team should deprecate these erroneous mappings and remap the mentioned ICD10CM concepts to the proper targets. We expect these fixes to be implemented during the nearest release.

Kind regards,
Masha

#5

Hi @m-khitrun, You mentioned that Qualifier Value concepts should not be used to map conditions. Is there OHDSI documentation explaining more about how/when to use Qualifier Value concepts?

Thank you,
Alison

#6

Hello @Alison_Park

Qualifier Value concepts do not carry the semantics of clinical events. Thus, if a source concept represents a clinical event (condition, procedure, etc.) it should not be mapped to a Qualifier Value.

Qualifier Values provide values for ‘Measurement/Value’ or ‘Observation/Value’ pairs and are used to fill the value_as_concept_id fields in the Measurement and Observation CDM tables, respectively.
Or, if you perform a mapping job, you may use them as targets to create ‘Maps to value’ relationships, not ‘Maps to’.

Also, you may find additional information about Qualifier Values in the SNOMED documentation.

Regards,
Masha.

#7

Thank you!

#8

@m-khitrun - Thank you so much for responding and clarifying. We appreciate all that you do to keep things moving forward! Tina

t