User Tools

Site Tools


documentation:vocabulary:changes_from_version_4

Changes from Vocabulary Version 4.0

The Standardized Vocabularies are undergoing a constant revision and improvement. Therefore, only significant structural changes are listed in the following.

Changes to Conventions

  • Concept Designation as Standard, Classification and Source
    • Source Concepts: In contrast to V4.0 of the Standardized Vocabularies, there is no longer a distinction between source codes and Concepts. All source codes now have their own Concept. However, in order to distinguish between codes that are used to represent a clinical entity in the data tables, Concepts are now designated Standard Concepts or Source Concepts.
    • For some vocabularies, source codes did have records in the Concept table, such as VA Product, RxNorm Concept Classes “Drug Form”, SNOMED Concepts other than Condition and Procedure. These Concept had a concept_level = 0. The concept_id of these non-standard Concepts has been preserved between the versions.
    • Standard Concepts: These Concepts were available in previous versions of the Standardized Vocabularies. They were designated through a concept_level value greater than 0. The concept_id of Standard Concepts has been preserved between the versions.
    • Classification Concepts: Like the Standard Concepts, the Classification Concepts existed in V4.0 of the Standardized Vocabulary. Their concept_level was typically 3 for drug classifications, 1-5 for MedDRA (Condition Classification) and 1-2 for SNOMED Procedure Classifications. Their concept_id has been preserved as well.
  • Mapping between source and Standard Concepts.
    • Each source code has now a representation as a Source Concept, and the mapping is achieved through the CONCEPT_RELATIONSHIP table. A new relationship_id = “Maps to” (reverse_relationship_id = “Mapped from”) links the Source to the Standard Concepts.
    • The SOURCE_TO_CONCEPT_MAP table is no longer used, and therefore no longer populated in the Standardized Vocabularies. However, it still can be used to map local source codes for specific ETL procedures using non-public vocabularies not supported here.

Changes to Tables

TableNature of change
DOMAINNewly created reference table for all Domains. Please refer to the discussion of the Domains.
CONCEPT_CLASSNewly created reference table for all Concept Classes of Concepts. Please refer to the discussion of the individual Vocabularies for detais.
SOURCE_TO_CONCEPT_MAPNo longer used for mapping of source codes to Standard Vocabularies. Instead, source codes have now a representation in the CONCEPT table, and these Source Concepts are mapped to Standard Concepts using records in the CONCEPT_RELATIONSHIP table “Maps to” and “Mapped from”. Please refer to the discussion of the mapping and the individual Vocabularies for detais.
COHORT_DEFINITIONNewly created reference table for the definition of Cohorts in the COHORT table. This table is not supported with standardized content, but for internal use at each CDM instance.
ATTRIBUTE_DEFINITIONNewly created reference table for the definition of Cohort Attributes in the COHORT_ATTRIBUTE table. This table is not supported with standardized content, but for internal use at each CDM instance.

Changes to Fields

TableFieldNature of change
CONCEPT and VOCABULARYvocabulary_idThis is no longer a numerical but a character identifier of length 20. The character string refers to the Vocabulary in a human readable format for convenience of the analyst during querying. E.g., vocabulary_id 1 became vocabulary “SNOMED”, vocabulary_id 2 became “ICD9CM”. Please refer to the discussion of the Vocabularies.
CONCEPTconcept_class_idThis used to be a character field called concept_class containing Concept Class as a string. Now, it contains a 20 character identifier referring to the CONCEPT_CLASS table, similar to how Vocabularies are referred to in the vocabulary_id field. Please refer to the discussion of the individual Vocabularies for details.
CONCEPTdomain_idThis is a newly added field for a 20 character identifier referring to the DOMAIN table, similar to how Vocabularies are referred to in the vocabulary_id field. Please refer to the discussion of the Domains.
CONCEPTstandard_conceptThis is a newly added field designating whether the Concept is a Standard Concept (“S”), a Classification Concept (“C”) or a Source Concept (NULL). Please refer to the discussion of how the Concept designation is used in practice.
CONCEPTconcept_levelThis field used to designate the hierarchical level within a Vocabulary or Domain, and is no longer available. Hierarchical levels only work for strictly stratified hierarchies such as RxNorm or MedDRA, and are no longer supported. Concept Level 0 used to designate non-Standard Concepts, which now is captured in the standard_concept field and explained here.
VOCABULARYvocabulary_referenceThis is a newly added field referring to external documentation of the Vocabulary.
VOCABULARYvocabulary_versionThis is a newly added field capturing the version of the Vocabulary as indicated in the source.
VOCABULARYvocabulary_concept_idThis is a newly added field referring to the Concept representing the Vocabulary.
CONCEPT_RELATIONSHIP and RELATIONSHIPrelationship_idThis is no longer a numerical but a character identifier of length 20. The character string refers to the Relationship in a human readable format for convenience of the analyst during querying. E.g., relationship_id 10 became “Subsumes”, relationship_id 353 became “Maps to”. Please refer to the discussion of the Vocabularies for details of the Concept Relationships.
RELATIONSHIPrelationship_concept_idThis is a newly added field referring to the Concept representing the Relationship.
CONCEPT_SYNONYMlanguage_concept_idThis is a newly added field referring to the Concept of the language used. Currently, only English is supported in the Standardized Vocabularies.
DRUG_STRENGTHamount_unit_concept_idThis field used to be called amount_unit containing the unit as a character string. Now, it is a foreign key to the Unit Concept in the CONCEPT table.
DRUG_STRENGTHnumerator_valueThis field used to be called enumerator_value.
DRUG_STRENGTHnumerator_unit_concept_idThis field used to be called enumerator_unit containing the unit as a character string. Now, it is a foreign key to the Unit Concept in the CONCEPT table.
DRUG_STRENGTHdenominator_unit_concept_idThis field used to be called denominator_unit containing the unit as a character string. Now, it is a foreign key to the Unit Concept in the CONCEPT table.
documentation/vocabulary/changes_from_version_4.txt · Last modified: 2015/01/20 03:13 by cgreich