RxNorm

Every clinical drug entity in RxNorm is assigned a unique, unambiguous Concept Unique Identifier (RXCUI). This identifier is a string of numbers that has no intrinsic meaning, ensuring it remains stable even if the drug's name or brand changes. The RXCUI serves as the central pivot for mapping between different vocabularies, allowing a system to understand that 'Acetaminophen 500 MG Oral Tablet' from one source is the same concept as 'Tylenol 500 MG Tab' from another.

RxNorm structures drug knowledge using a rich set of Term Types (TTY) to represent different levels of granularity. These include:

• SCD (Semantic Clinical Drug): The most precise level, specifying active ingredient, strength, and dose form (e.g., 'Acetaminophen 500 MG Oral Tablet').
• SBD (Semantic Branded Drug): A branded version of an SCD (e.g., 'Tylenol 500 MG Oral Tablet').
• IN (Ingredient): The active chemical substance (e.g., 'Acetaminophen').
• BN (Brand Name): The proprietary trade name (e.g., 'Tylenol'). This hierarchical model allows systems to query at the exact level of specificity required for a clinical workflow.

A core function of RxNorm is its ability to act as a universal translator between major drug vocabularies. It provides a curated crosswalk linking its own RXCUIs to codes from source vocabularies such as GS (Gold Standard Drug Database), MMSL (Multum MediSource Lexicon), NDFRT (National Drug File - Reference Terminology), and VANDF (Veterans Health Administration National Drug File). This linking is essential for health information exchange, allowing a pharmacy system using one vocabulary to communicate seamlessly with a prescriber's EHR using another.

RxNorm is not a flat list but a semantic network where concepts are connected through explicit, named relationships. Key relationships include:

• has_ingredient: Links a clinical drug to its active component.
• has_tradename: Links a branded drug to its brand name.
• has_dose_form: Links a drug to its physical form (e.g., Tablet, Injection).
• constitutes: Links an ingredient to a multi-ingredient drug. These relationships enable algorithmic reasoning, such as identifying all branded products containing a specific active ingredient or finding generic alternatives for a prescribed brand.

To keep pace with the dynamic pharmaceutical market, RxNorm follows a rigorous monthly release cycle. Each release is a complete, self-contained snapshot of the entire dataset, identified by a specific version number. This regular cadence ensures that new drugs, generics, and market withdrawals are incorporated into the standard with minimal latency. The predictable versioning is critical for regulated clinical AI systems that must be auditable and reproducible, as a model's output can be traced back to the exact RxNorm release used.

The metathesaurus that binds RxNorm to over 100 other vocabularies. The UMLS assigns a Concept Unique Identifier (CUI) to link equivalent terms across disparate source vocabularies, such as linking RxNorm's normalized drug name to its corresponding SNOMED CT clinical drug concept and its MeSH pharmacological action. This semantic network is the backbone of cross-vocabulary clinical NLP.

The most comprehensive clinical terminology, containing hierarchical drug concepts that map to RxNorm. While RxNorm provides the normalized name and ingredient-level relationships, SNOMED CT provides the rich clinical context—linking a drug to its therapeutic role, disease indications, and contraindications. This mapping is critical for clinical decision support systems that need to reason about a patient's full clinical picture.

The FDA's identifier for packaged drug products, representing the physical packaging layer of a medication. RxNorm normalizes the many-to-many relationship between NDCs and clinical drugs. A single generic drug may have dozens of NDCs from different manufacturers, and RxNorm links them all to a single Semantic Clinical Drug (SCD) concept, enabling accurate medication reconciliation across pharmacy benefit managers and EHRs.

The Medication Reference Terminology maintained by the Veterans Health Administration. It provides a formal ontology of a drug's mechanism of action, physiologic effect, and pharmacokinetics. RxNorm's integration with MED-RT allows clinical NLP systems to infer that two drugs with different RxNorm ingredient codes share a therapeutic class, enabling drug class-based allergy checking and formulary substitution logic.

The most granular RxNorm term type representing a fully specified clinical drug with ingredient, strength, and dose form. For example, 'Acetaminophen 500 MG Oral Tablet' is an SCD. This concept is the target for NLP extraction pipelines that must map a raw text mention like 'Tylenol 500mg PO' to a structured, computable identifier for drug-drug interaction checking and clinical trial eligibility screening.

The RESTful web service provided by the National Library of Medicine for programmatic access to RxNorm data. It supports functions like getApproximateMatch for fuzzy string searching, getRelatedByType for navigating the RxNorm graph, and getNDCProperties for reverse-mapping an NDC to its clinical components. This API is the runtime dependency for real-time medication normalization in clinical AI systems.