Observational Health Data Sciences and Informatics

The drug exposure domain concepts capture records about the utilization of a Drug when ingested or otherwise introduced into the body. A Drug is a biochemical substance formulated in such a way that when administered to a Person it will exert a certain physiological or biochemical effect. The following products are not considered Drugs, but Devices:

• Radiopharmaceuticals
• Contrast material for imaging
• Nutritional products and supplements, including baby formula. In reality that results in the slightly arbitrary and in some cases difficult to ascertain situation that solutions of salts for parental use are Drugs (hydrating patients and maintaining ionic balance), while the addition of nutrients such as glucose or vitamins makes them devices (feeding patients).
• Products directly derived from blood (e.g. erythocytes or plasma)

Drug domain Concepts should be used in the drug_concept_id of the DRUG_EXPOSURE or DRUG_ERA (only at the Ingredient level) tables, or in the value_as_concept_id field of the OBSERVATION or MEASUREMENT table (e.g. for observations like “Plasma level”).

Structure of Drug Domain

Standard Concepts are taken from the RxNorm and DMD vocabularies. All Standard RxNorm and DMD Concepts can be used in the drug_concept_id in the DRUG_EXPOSURE and DRUG_ERA tables.

RxNorm is the foundation for the combined Domain space. All valid RxNorm Concepts (invalid_reason is null) are used as Standard Concepts. All DMD Concepts that have an equivalence relationship to a RxNorm Concept are non-Standard Concepts, i.e. they are the equivalent in the UK market. All DMD Concepts without such equivalence realationship are considered to be unique UK Concepts and get Standard Concept status. That way, a combined Drug Domain is constructed, in which entities common to both markets exist in their RxNorm identity, and the additional DMD Concepts only present in the UK get added.

Note: In future releases, more international Drug Concepts will be added that exist neither in the US (RxNorm) or UK (DMD). These will again become Standard Concepts. As a result, a mixed Drug Domain will exist with each Standard Concept guaranteed to be unique. However, this is a purely operational mechanism, the vocabulary_id of a Standard Concept should not be used to identify whether a drug exists in a market or not, as they become Standard on a “first get served first” approach.

RxNorm, DMD and all added vocabularies, are hierarchical. Therefore the Concepts can also be used as Class Concepts: descendants of any Standard Concept in the CONCEPT_ANCESTOR table can expected to be a correct semantic match in a query. For example, a query for the descendants of an ingredient should turn up all drug products (Clinical Drug or Branded Drug) containing that ingredient.

On top of this combined RxNorm/DMD Drug Domain are classification systems. These are derived from a variety of different vocabularies. Note: Classification Concepts are not considered unique, but may exist in different classifications. For example, NSAID Analgesics exist as ETC, VA Class and SNOMED classes.

The Source Concepts are loaded as is. CPT4 and HCPCS drug are so-called procedure drugs, which are typically administered by a healthcare provider. Not all CPT4 and HCPCS drugs can only be mapped to a distinct RxNorm Concept if such a Concept can unambigously identified. In many cases, such as chemotherapy administration, the detail about the precise chemotherapeutic agent is not coded.

All Source Vocabularies are mapped to the RxNorm Standard Concepts, and Multilex and DMD also to the DMD Standard Concepts. The mapping is either constructed from the vocabulary source and in some cases curated by the vocabulary team.

All these are provided by RxNorm.

These are manually curated by the vocabulary team

This is provided by the UMLS.

These are inferred from public source and curated by the vocabulary team.

These are inferred from public source and curated by the vocabulary team.

All relationships provided by RxNorm are loaded as is, but renamed to follow the relationship_id formatting rules and allow for bi-directional mappings, see RxNorm. All internal relationships of DMD are derived from the source and converted to generate “pseudo”-RxNorm relationship, omitting the Drug Form and Drug Component Concepts. See DMD for details. All hierarchical relationships of the drug classes are loaded for the source as is and used to form the CONCEPT_ANCESTOR table.