Unified Medical Language System (UMLS)

Key facts

Produced by

National Library of Medicine (NLM)

Components

Metathesaurus, Semantic Network, SPECIALIST Lexicon and Lexical Tools

Release cycle

Twice a year, in May and November

Access

Free UMLS license; some source vocabularies carry added restrictions

Concept identifier

CUI: the letter C followed by seven digits

What is the UMLS?

Health data is recorded in dozens of vocabularies, from billing code sets to research thesauri. The Unified Medical Language System (UMLS), built by the National Library of Medicine (NLM), connects them: when terms from different vocabularies mean the same thing, the UMLS assigns them a shared Concept Unique Identifier (CUI).

The UMLS is a knowledge resource for developers and researchers, not an end-user application or a code set for claims. Its sources include SNOMED CT, ICD-10-CM, LOINC, RxNorm, MeSH and CPT, along with vocabularies in languages other than English.

How the UMLS is organized

The UMLS has three knowledge sources:

NLM distributes the files through the UMLS Terminology Services, which also offers a browser and an API. Users can build a local subset containing only the vocabularies they need with the MetamorphoSys tool.

• Metathesaurus: the core database of concepts, names, codes and relationships gathered from the source vocabularies. Each concept has a CUI, and each source term keeps its original code.
• Semantic Network: broad semantic types, such as Disease or Syndrome and Pharmacologic Substance, assigned to every concept, plus relationships among those types.
• SPECIALIST Lexicon and Lexical Tools: an English lexicon and programs that handle spelling variants, inflections and word order in biomedical text.

Why the UMLS matters

The UMLS is infrastructure for anyone linking or parsing health vocabularies:

• Crosswalks: a shared CUI helps link a SNOMED CT problem, an ICD-10-CM code and a MeSH heading that describe the same condition.
• Natural language processing: tools that extract diseases, drugs and procedures from clinical notes or articles often map text to UMLS concepts.
• Licensing gateway: U.S. users obtain the full RxNorm files, SNOMED CT and other vocabularies through a UMLS license.

Common misconceptions about the UMLS

A shared CUI means NLM editors judged terms to be synonymous. It is not an official billing crosswalk, and a link can be broader or narrower than a coder would accept for a claim.

The UMLS license is free, but it does not grant rights to every vocabulary inside it. Some sources carry additional restrictions, and certain uses of vocabularies such as CPT or MedDRA require separate licenses.