A systemization describes a set of assertions about the relationships between:
- Two entity descriptions.
- An entity description and a set of entity descriptions.
- The absence of a relationship between an entity description and any other entity descriptions.
The set of assertions within a given systemization are assumed to be consistent and cohesive and, in some cases, targeted for a specific use. As an example, the set of relationships in the SNOMED?CT relationships table are considered to be one systemization (assuming we ignore localizations). The mapping between SNOMED CT and ICD 9 CDM released by IHTSDO would be considered a second systemization.
A systemization may stand alone when asserted outside of the context of a code system. As an example, a third party mapping between SNOMED CT and ICD 10 would be represented as a systemization, as the mapping plays no role in making definitional or informative assertions about the intent or purpose of the specific entities themselves. A systemization may be associated with a code system when it directly or indirectly plays a role in the definition or description of the entities referenced by the code system itself. A systemization may be identified as the “native” systemization of a code system, meaning that it is the primary (or only) systemization used in the definition of the code system contents. There may only be one “native” systemization for any given code system.
A systemization may also describe a mapping between a source code system and one or more target systems. Note that the types of “mappings” that a systemization can describe are somewhat limited. The CTS2 specification, for instance, provides no mechanism for evaluating conditional mappings of performing complex transformations between two different post-coordinated expressions.
