Observational Health Data Sciences and Informatics

Conditions are records of a Person suggesting the presence of a disease or medical condition stated as a diagnosis, a sign or a symptom, which is either observed by a Provider or reported by the patient. Note that the Condition Domain makes no distinction between a mere observation of a symptom, a fully worked out diagnosis or anything inbetween.

Structure of Condition Domain

Standard Concepts are taken from the SNOMED vocabulary. All valid (invalid_reason is null) SNOMED Concepts of the Condition Domain are Standard Concepts and can therefore be used in the condition_concept_id in the CONDITION_OCCURRENCE and CONDITION_ERA tables.

SNOMED is a hierarchical vocabulary. Therefore SNOMED concepts can also be used as Class Concepts: descendants of any SNOMED Concept in the CONCEPT_ANCESTOR table can expected to be a correct semantic match in a query.

MeDRA is used as an additional classification system. The connection between MedDRA and SNOMED is accomplished at the level of MedDRA Preferred Term Concept to the equivalent SNOMED concept. As a result, MedDRA Concepts can also be used to query data, where SNOMED descendants of any MedDRA Concept in the CONCEPT_ANCESTOR table can expected to be a correct semantic match. However, MedDRA Low Level Term Concepts are not mapped to SNOMED. Also, due to this selective mapping, SNOMED Concepts that are more general than their MedDRA Preferred Term equivalent cannot be retrieved by a CONCEPT_ANCESTOR query.

The Source Vocabularies are mapped to SNOMED Standard Concepts. The mapping is either provided by the vocabulary source (Read, CIEL, MeSH) or developed by the Vocabulary Team.

All internal relationships provided by SNOMED are loaded, but renamed to follow the relationship_id formatting rules and allow for bi-directional mappings, see SNOMED