OHDSI Home | Forums | Wiki | Github

Anatomic Sites mapped to Observations in latest vocabulary

For Specimen, I’ve got 4 Anatomic Sites, which in our earlier vocabulary (Jan 2022) were:

concept_id concept_name domain_id vocabulary_id concept_class_id standard_concept concept_code
4076390 Stoma Spec Anatomic Site SNOMED Body Structure S 245857005
4092909 Urinary reservoir Spec Anatomic Site SNOMED Body Structure S 281162000
4134665 S pouch Spec Anatomic Site SNOMED Body Structure S 262119009
4148332 End ileostomy stoma Spec Anatomic Site SNOMED Body Structure S 311412005

These have now been mapped as Observations in our latest version (May 2023)

concept_id concept_name domain_id vocabulary_id concept_class_id standard_concept concept_code
4076390 Stoma Observation SNOMED Morph Abnormality S 245857005
4134665 S pouch Observation SNOMED Morph Abnormality S 262119009
4092909 Urinary reservoir Observation SNOMED Morph Abnormality S 281162000
4148332 End ileostomy stoma Observation SNOMED Morph Abnormality S 311412005

I can’t tell if the previous version was corrected, or if a new error has been introduced into the vocabulary.

It is neither a bug nor a feature.

These changes were caused by Snomed. They changed concept_class_id of these concepts from Body Structure to Morph Abnormality and we incorporated these changes during our release in August 2022.
I guess, the logic behind the changes was the following: post-surgical anatomy should be considered not an anatomically normal body structure, hence morphological abnormality.

Our vocabularies are quite consistent with all the related concepts

SELECT ca.ancestor_concept_id, c.*
FROM concept_ancestor ca 
JOIN concept c 
ON c.concept_id = ca.descendant_concept_id
AND ca.ancestor_concept_id = 4084082; --Post-surgical anatomy

Okay. Good.

Now the question for the ETLer is whether these concepts still represent valid anatomic sites for the Specimen record. The Accepted Concepts from the current CDM require Spec Anatomic Site and Body Structure.

t