SNOMED Documentation Search

 Other Documents
Skip to end of metadata
Go to start of metadata

Current Version - Under Revision

This guide does not specify a particular API . The services specified in this guide may be delivered using various types of interfaces based on a range of different technologies including:

  • Web services such using WSDL ( Web Services



    ) or REST (Representational State Transfer) interfaces;
  • Java


    such as JavaBeans or Eclipse plug-ins;
  • Microsoft .NET or Active -X / COM / DCOM in Microsoft Windows environments;
  • CORBA (Common Object Request Broker Architecture).

Decisions on which technologies to support depend on the intended functionality, performance, accessibility, ease of use and support requirements for maintenance or updates.

Over the past two decades there have been various efforts to specify standards for terminology servers and related APIs . The most recent development in this area is centered around the Common Terminology Server Release 2 ( CTS2 ). The requirements initially identified and documented within HL7 have now led to an OMG (Object Management Group) proposal. At least one of the responses to this proposal focuses directly on SNOMED CT related requirements. The OMG process is expected to result in a detailed specification and prototype implementation during 2011.

[1 Usually a single group is present in a context expression. Theoretical cases exist for multiple groups where different contexts apply to different aspects of a concept but these cases are beyond the scope of the normalization rules in this guide.

[2 A high-level primitive is a concept that is primitive and has no fully defined supertypes.

[3 An intermediate primitive is a concept that is primitive but which has fully-defined supertypes and subtypes.

[4 Only on very rare occasions will a single finding or procedure require separate lateralization of different sites in its definition. However, support for the direct approach does not preclude the nested approach if it is necessary to associate different laterality refinement with different structures.

[5 The words "attribute" and " attribute name " are used here as documented in the SNOMED CT guide to the "Abstract Logical Models and Representation Forms". In SNOMED CT distribution files a "defining relationship " is equivalent to this use of the word "attribute" and a " relationship type" represents an "attribute name".


The only valid context attributes are: 246090004 |associated finding| , 363589002 |associated procedure| , 2470590016 |finding context| , 2470591017 |procedure context| , 2470592012 |temporal context| and 2470593019 |subject relationship context| .

[7 This issue will gradually diminish in significance as more concepts are fully defined through addition of new defining relationships.

[8 The predicate long normal form can be used instead of the predicate short normal form . However, the short form is preferred as it reduces the number of steps required in testing each candidate expression.

[9 Active concepts that are related to ambiguous candidate -concepts should only be tested after deciding whether the prime objective of retrieval is "completeness" (in which case include these possible related concepts) or "precision" in which case they should be excluded unless the ambiguity can be resolved by the term selected with the original concept.