From Clinfowiki
Jump to: navigation, search

Systematized Nomenclature Of Medicine Clinical Terms (SNOMED CT) is a reference terminology standard from the Unified Medical Language System (UMLS). SNOMED that consists of concepts, terms, and the interrelationships between them. It standardizes the way healthcare terminology and data is recorded. It aims to facilitate the coding, retrieving, analysis, aggregation, indexing, and the exchanging of clinical information across health care entities.


SNOMED was designed for use in software applications to represent clinically relevant information in a reliable and reproducible manner. It currently has 370,000 pre-coordinated concepts. SNOMED-CT allows post-coordination, meaning that concepts can be developed by combinations of other concepts, as opposed to strictly pre-coordinated terminologies where all concepts are given express codes beforehand. Post-coordination is governed by a Description Logic. Despite the Description Logic, some primitive concepts can be expressed in terms of other concepts, and there are semantically duplicate concepts.


Systematized Nomenclature Of Medicine Clinical Terms (SNOMED CT) was created by the merger of Systematized Nomenclature Of Medicine Clinical Terms Reference Terminology (SNOMED RT), maintained by the College of American Pathologists (CAP), USA, and the UK’s National Health Service’s Clinical Terms Version 3 (CTV3), or Read Codes. In 2007, the International Health Terminology Standards Development Organization (IHTSDO) [1], a non-profit organization, bought the licensing rights to SNOMED CT from the CAP.

The College of American Pathologists (CAP) developed the Standard Nomenclature of Pathology (SNOP) in 1965, basing it on the New York Academy of Medicine’s Standard Nomenclature of Diseases and Operations (SNDO). The Systematized Nomenclature of Medicine (SNOMED) was developed in 1974. This was expanded in 1979 into SNOMED II, followed by the Systematized Nomenclature of Human and Veterinary Medicine, SNOMED International, or SNOMED Version 3, in 1993. These versions were multi-axial. Coding was done by post-coordination of terms from multiple axes to represent complex terms.

SNOMED changed from a multi-axial terminology to a more logic-based structure in May 2000 with the release of SNOMED-RT, for Reference Terminology. The CAP, working with Britain’s National Health Service (NHS), merged SNOMED, with its strong support for terminology in specialty medicine, with Clinical Terms Version 3 (formerly READ codes), with its strong support for terminologies in general practice, to create SNOMED-CT, the Systematized Nomenclature of Medicine - Clinical Terms. This was released in January 2002.


Systematized Nomenclature Of Medicine Clinical Terms is comprised of concepts and relationships among those concepts. The relationships formally define the concepts. The concepts in SNOMED CT exist in a hierarchical structure, which contains 19 top-level hierarchies, or axes. Each top-level hierarchy contains sub-hierarchies that further specify a concept.

SNOMED CT is multi-hierarchical; a single concept can exist in multiple sub-hierarchies. However, a single concept can exist in more than hierarchy. For example, the concept "aspirin (substance)" is the substance acetylsalicylic acid, but the concept "aspirin (product)" includes all drug products that contain aspirin.


For SNOMED CT, a concept is some clinical definition or entity that is associated with a unique ConceptID. A ConceptID is a unique, numeric identifier given to a concept. A ConceptID is permanent and has no implicit meaning. A standard term for the concept is a human readable description of a concept at some level of granularity. There are three types of descriptions for a SNOMED CT concept: Preferred Term, Fully Specified Name, and Synonym.

The Preferred Term is some word or phrase that is used by clinicians to name a clinical concept.

The Fully Specified Name is essentially the Preferred Term, along with a “semantic tag” as a suffix to indicate the type of concept and to eliminate ambiguity. For example, a particular concept has the Preferred Term of “apoptosis”, the ConceptID of “20663007”, and the Fully Specified Name of “Apoptosis (morphologic abnormality)”. Synonyms are additional terms that may define the concept at the same level of granularity. There are four types of relationships that a concept can have in SNOMED CT: Defining, Qualifying, Historical, and Additional. Every concept in SNOMED CT has a defining, hierarchical relationship called IS_A, to a slightly less granular parent concept (except the grand-daddy Root Concept). The IS_A relationship is basically a parent-child relationship. For example, the concept apoptosis IS_A “morphologically altered structure” IS_A “body structure”.

List of domains

The 19 top-level domains:

  1. Body structure
  2. Clinical Finding
  3. Context-dependant category
  4. Environments and geographical locations
  5. Event
  6. Linkage concept
  7. Observable entity
  8. Organism
  9. Pharmaceutical / biologic product
  10. Physical force
  11. Physical object
  12. Procedure
  13. Qualifier value
  14. Record Artifact
  15. Social Concept
  16. Special concept
  17. Specimen
  18. Staging and scales
  19. Substance


A new version of SNOMED CT is released every 6 months, in January and July. There are over 300,000 concepts, 800,000 descriptions, and almost a million relationships in the latest release of SNOMED CT.

SNOMED-CT Spanish Edition was released in April 2002, and the German Edition was released in April 2003. The International Healthcare Terminology Standards Development Organization (IHTSDO) acquired SNOMED-CT in April 2007. The National Library of Medicine (NLM) is the US member of IHTSDO and distributes SNOMED-CT in the US.

Additional stuff

SNOMED-CT cross maps to multiple other terminologies, including ICD-9-CM, ICD-10, and LOINC. It supports ANSI, DICOM, HL7, XML and ISO standards. The US Government has accepted it as a standard for the National Health Information Infrastructure.

The thesaurus is a list of terms created from almost 2 million free text inputs extracted from the clinical data repository. The terms included in the thesaurus are divided into concepts (real clinical entities) and descriptions (different ways of naming these clinical entities)


  1. Edward H. Shortliffe and James J. Cimino, Biomedical Informatics: Computer Applications in Health Care and Bioinformatics, Third Edition (New York: Springer, 2006) 285-288.
  2., “Historical Perspectives,” 01/26/2008,
  3., “Requirements Document,” 06/03/2006, 01/26/2008,