Hello @bertoverduin , thank you for your question. The language_concept_id values you listed appear to be SNOMED CT-related IDs, representing different languages or dialects. However, the SNOMED CT vocabulary may not have been included in your download from Athena if you didn’t explicitly select it or if it wasn’t part of the default download.
Since these language_concept_id values correspond to SNOMED concepts, the absence of the SNOMED CT vocabulary in your dataset is causing a mismatch between the CONCEPT_SYNONYM and CONCEPT tables. The CONCEPT_SYNONYM table includes these language_concept_id entries because they are part of the terminology used to manage synonyms in SNOMED, but without SNOMED itself, those IDs don’t have corresponding entries in the CONCEPT table.
While you specifically downloaded the LOINC and Gender vocabularies, some of the terms in those vocabularies may rely on SNOMED CT for concept definitions, including language-specific synonyms. As a result, the language_concept_id entries ended up in your download because they are used to manage synonyms in different languages or dialects. However, without the SNOMED CT vocabulary present, those IDs don’t have corresponding entries in the CONCEPT table.
You may try downloading and loading the SNOMED CT vocabulary from Athena. This will ensure that all language_concept_id values in the CONCEPT_SYNONYM table have corresponding entries in the CONCEPT table. Once you’ve downloaded SNOMED CT, reload all vocabularies into your v5.4 CDM to ensure that the missing SNOMED concept IDs are properly included. After loading the SNOMED CT vocabulary, you can try to apply the constraints without encountering the error, as the missing language_concept_id values will now be present in the CONCEPT table.
