Assymetric mapping between terminologies?

[Kim Tang]

Hello,

I would like to make use of the feature to use the "Mapping between terminologies" to map from ICD10 to SNOMED-CT and vice versa.

However I encountered a problem where a SNOMED-CT concept maps to an ICD code, but the ICD code maps to another concept:



Shouldn't it be a symmetric relation/mapping between concepts of different terminologies, e.g. if A maps to B, then B maps to A as well?

For clarity regarding the code, it is essentially the code from the documentation but used on the "mapper" class:

        PYM = snomed_world.get_ontology("http://PYM/").load()
        mapper.SNOMEDCT = PYM["SNOMEDCT_US"]
        mapper.ICD10 = PYM["ICD10"]

Am I not using the mapping feature correctly or do I have to change the call somehow?

Kind regards,
Kim Tang

[Jiba]

Hello,

The mapping should not be expected as being symmetric, due to the difference in terminologies. Here, the concept J98.4 "Other disorders of lung" does not exist in SNOMED CT (NB in that concept, "other" means "not associated with any other ICD10 code", so this concept exists only in ICD10).

Thus there is no direct mapping for J98.4. In such cases, Owlready tries to map the parent concept, leading to the more general concept of Disorder of the respiratory system in SNOMED CT.

Jiba

[Kim Tang]

Ah I see, I did not know that Owlready tries to match the parent concept in such a case. Then I understand why it is not acting symmetrically when mapping concepts from ICD to SNOMED-CT and vice versa.

Does Owlready only check in such cases if a mapping of the parents exist, or does it continue to iterate through the ancestors to find a mapping (e.g. if the parents also don't have a mapping)?

[Jiba]

> Does Owlready only check in such cases if a mapping of the parents exist, or does it continue to iterate through the ancestors to find a mapping (e.g. if the parents also don't have a mapping)?

Yes, it continues to iterate through the ancestors.

Jiba