Deployment and Versioning
Distribution and Access
SNOMED CT drug content is distributed using Release Format 2 (RF2), the standard format for both the International Edition and all SNOMED CT extensions.
More information
The same format is used for both international content and national drug extensions, ensuring that they can be combined into a single, coherent terminology dataset.
Drug extensions are typically distributed as RF2 packages that are explicitly aligned to a specific version of the SNOMED CT International Edition. This dependency must be respected when deploying and updating terminology content.
Approach
Implementations of the SNOMED CT Drug Model require loading a combined dataset consisting of:
the SNOMED CT International Edition, which provides the clinical backbone (MP, MPF, CD)
the National Drug Extension, which provides real-world product and packaging concepts (RCD, CDP, RPCD)
associated terminology derivatives, e.g. mappings to local drug dictionaries (this may be part of a reference set in the Drug Extension)
Drug extensions are not standalone. They depend on the International Edition and extend it by adding product identity and packaging information. As a result, implementations must treat the terminology as a layered model, where extension concepts build directly on international concepts.
To ensure correct resolution of dependencies, content should be loaded in the following order:
International Edition
National Drug Extension
Derivatives
Key considerations include:
maintaining clear module separation between international and extension content
ensuring version alignment between the extension and its dependent International Edition
validating that all referenced concepts (e.g. substances, dose forms) are present
building indexes to support efficient querying across the drug hierarchy
Version Management
SNOMED CT drug content evolves continuously due to:
introduction of new medicinal products
regulatory changes
updates to substances, dose forms, and strengths
refinement of the concept model
RF2 uses an append-only versioning mechanism, where each change to a component results in a new version identified by its effective time and active status. Previous versions are retained, enabling full historical traceability.
Implementations should support:
regular ingestion of updated releases
identification of new, modified, and inactivated concepts
reclassification and reindexing after updates
Concepts may be inactivated due to product withdrawal, modeling changes, or replacement by more accurate representations. Systems must detect inactive concepts and, where available, use historical association reference sets to identify replacement concepts. This is particularly important for maintaining continuity in longitudinal data.
Historical traceability allows systems to:
interpret past medication records correctly
support audit and regulatory requirements
perform time-based analytics
Reference sets are also versioned and must be updated alongside core terminology content. This is especially important for:
language refsets, which determine preferred terms for display
mapping refsets, which link SNOMED CT concepts to local drug dictionaries