Page tree

Introduction

SNOMED CT is a dynamic and evolving terminology in which the content must remain concurrent with current clinical practice and our evolving understanding of disease processes and treatments. As a result, content may change, become outdated, or as a result of improved understanding needs remodeling. However, individual patient records are created at a point in time and may use concepts that later become inactivated and replaced. For legal reasons, it is important that the concepts used at the time the record was made should persist over time so that records made prior to a change in the status of that concept remain available to extraction routines for population analysis and more importantly to ensure that individuals do not get lost to routine follow-up for ongoing care etc.

Managing and making transparent the history of a given concept over time is the responsibility of SNOMED International and this is achieved through the concept inactivation process. In order to make the process transparent, it is necessary to clearly document the reason for a given change in the status of a concept and provide vendors and users with the information necessary to understand those changes. Vendors can then use this information in such a way as to protect the meaning encapsulated in the individual's clinical record over time and ensure that decision support, routine analysis, and population studies can continue consistently and without risk to clinical safety.

There are two key elements to the inactivation process, firstly allocating an appropriate reason for the concept inactivation and secondly providing a clear understanding of the level of semantic equivalence through the historical association link between the inactivated concept and its replacement where it exists. To further improve the quality of our own audit trail additional information and links to reference sources can be added within text fields that accompany each inactivation.

The vendor can use the historical association 'type' to understand the level of semantic equivalence between the inactivated concept and its replacement and as a guide to what action should be taken by the end-user in updating their patient records and routine extraction routines in response to each inactivation.

Below is a summary of the inactivation reasons and the meaning of each of the historical associations and their implications for vendors/users. The link to the detailed guidance for each inactivation reason is provided.

Inactivation Reasons

Inactivation ReasonAssociation TypeSummary Definition and Usage

Ambiguous Concept

POSSIBLY_EQUIVALENT_TO
  • The inactivated concept is inherently ambiguous.
  • It is important to differentiate between inherent ambiguity and "vagueness". If the FSN is vague consider using Meaning Unknown.
  • Identify all of the POSSIBLY_EQUIVALENT_TO concepts, the sum of which should be semantically identical to the inactivated concept.
  • Only in exceptional circumstances will it be acceptable to have only one POSSIBLY_EQUIVALENT_TO target and that is where the second target concept represents a concept that is considered to be of little or no clinical usefulness.
  • Note: Non-synonymous synonyms should be inactivated and resigned before inactivating the ambiguous concept.

Classification Derived Concept

REPLACED_BY

PARTIALLY_EQUIVALENT_TO

  • The inactivated concept originates from the “closed world” classification paradigm and as such is inappropriate content for use within the clinical record.
  • These concepts often contain the expressions "not otherwise specified (NOS)", "Unspecified", "Otherwise specified", conjunction/disjunction, etc.
  • REPLACED_BY should be used where the target replacement concept is of the same semantic category and have the same or less specificity than the inactivated concept. It will often be the same as the inactivated concept minus the "classification tail" of "other" or "NOS", etc.
  • PARTIALLY_EQUIVALENT_TO should be used where there is a conjunction, (with or and) and must include all of the elements of the conjunction e.g. has a cardinality of 2 or more.

Concept Moved Elsewhere

ALTERNATIVE 

No suitable replacement identified

  • Applies where jurisdictional control of a concept passes between extensions, or between the international edition and an extension (e.g. veterinary extension)
  • Where the meaning of the concept to be inactivated is strictly jurisdictional e.g. relates to specific forms, branded products, legal entities, etc, no replacement is required and, therefore, no association type is allocated.
  • ALTERNATIVE should be used where there is concern that the concept may have been used by members of the community of practice. The ALTERNATIVE substitute concept is offered as semantically "sufficiently similar" to the original concept for some use cases that should be specified by the end-user. 

Duplicate Concept

SAME_AS

  • The inactivated concept semantically represents exactly the meaning of the remaining active concept and therefore the association is "SAME_AS".
  • Note that the meaning of the concept is based on the FSN but does not imply that the FSNs are identical.

Erroneous Concept

REPLACED_BY

  • The inactivated concept has an FSN which contains an error that when corrected potentially changes the semantic meaning of the concept.
  • NOTE: Where the error is grammatical or a spelling mistake which when correct does not change the meaning, the description rather than the concept should be inactivated. 

Meaning Unknown

No association required

  • The meaning of the inactivated concept cannot be determined.
  • It will normally be necessary to search the clinical literature to establish that this is truly an unknown concept rather than an outdated clinical concept.
  • This inactivation reason may be used where the meaning of the FSN is considered to be vague.

Non-Conformance to Editorial Policy

REPLACED_BY

ALTERNATIVE

No suitable replacement identified

  • The inactivated concept does not comply with the current or updated Editorial Policy.
  • Where conformance to editorial policy potentially changes the meaning of a concept and it is possible to replace this with a concept that is semantically very close to the inactivated concept the association should be REPLACED_BY. e.g. a radiological procedure which in practice almost exclusively uses contrast but this wasn't included in the FSN would be inactivated and REPLACED_BY the new concept in which the FSN clearly states the use of contrast.
  • Where the change to editorial policy results in a change in scope of SNOMED CT then it may be possible to provide a target replacement that is similar to but less specific than the inactivated concept and therefore the association would be ALTERNATIVE. e.g. where branded products were considered out of scope for SNOMED CT an ALTERNATIVE would be the generic product.
  • In some instances, a suitable replacement cannot be identified e.g. should a decision be taken to count occupations as out of scope for SNOMED International Edition.

Outdated Concept

REPLACED_BY 

POSSIBLY_REPLACED_BY

No suitable replacement identified

  • The inactivated concept is an outdated concept that is no longer considered to be clinically acceptable or semantically interoperable internationally.
  • A REPLACED_BY association would be used where there exists a concept that is semantically similar to or more general than the inactivated concept for the purposes of data healing and analysis of historical data.
  • Where there has been an update to a disorder classification, a substance or organism originally believed to be a single entity has, following further study, been reclassified as 2 or more substances or organisms. Where 2 or more potential replacements exist use POSSIBLY_REPLACED_BY for each of the target concepts.
  • In some circumstances, an outdated concept simply falls into disuse without any appropriate replacement.

Historical Associations

AssociationSummary Definition
SAME_AS

For the Author:

  • The SAME_AS association type exists to declare when any pair of different concept identifiers in fact represent exactly the same semantics. They are true semantic duplicates, for all time.
  • Because of this statement of unambiguous and exact, bidirectional semantic equivalence, (A) SAME_AS (B) explicitly implies that everything ever subsequently said about B is, by definition, also true of A. (And, technically, also vice versa though, since A is inactive, nothing new should ever be said about it)

Implications for the vendor/user:

  • Given that the inactivated concept is semantically equivalent to the retained concept there should be no requirement for the user to review the replacement.
POSSIBLY_EQIVALENT_TO

For the Author:

  • The POSSIBLY_EQUIVALENT_TO association is to be used where an inactive concept was previously the conflation or disjunction of two or more distinct possible meanings, at least one of which is also now represented as a current SNOMED concept.
  •  The implication of a set of POSSIBLY_EQUIVALENT_TO associations is that all possible EHR instances previously coded to the inactivated code can be re-coded to one or other of the listed association concepts, without exception and also without either adding or losing any semantics not already expressed by the original coding. No other active concept exists that would more precisely capture what was expressed by a clinician originally selecting the inactivated code.

Implications for the vendor/user:

  • Where a POSSIBLY_EQUIVALENT_TO association is encountered this implies that a decision is to be made by the end-user as to which of the available POSSIBLY_EQUIVALENT_TO associations correctly represents the original meaning stated by the clinician.
  • The vendor must provide adequate tooling to enable the end-user to view all of the POSSIBLY_EQUIVALENT_TO associations and select the most appropriate one to replace the inactivated concept.

PARTIALLY_EQUIVALENT_TO

For the Author:

  • The historical association PARTIALLY_EQUIVALENT_TO explicitly states; when a concept that contains conjunction using “WITH” or “AND” is inactivated each of the individual elements of the conjunction is replaced by semantically equivalent clinical concepts using an association type of “PARTIALLY_EQUIVALENT_TO”. Where an element does not currently exist within SNOMED CT it must be created to ensure exact semantic equivalence between the inactivated concept and the sum of its replacements.

Implications for the vendor/user:

  • "PARTIALLY_EQUIVALENT_TO"  signifies to the implementer and end-user that the single inactivated concept MUST be represented within the clinical notes by ALL of the PARTIALLY_EQUIVALENT_TO targets.

REPLACED_BY

For the Author:

  • The purpose of the REPLACED_BY association type is to indicate that the suggested target replacement concept is NOT semantically equivalent to the inactivated concept (i.e., SAME_AS cannot apply). The target concept chosen should normally be of the same semantic category and have the same or less specificity than the inactivated concept.
  • The stated REPLACED_BY target represents an alternative which, although not semantically exactly the same, is considered to be clinically useful and - in the judgment of the author/editor - represents the intended original meaning of the inactivated concept closely enough for some unspecified data processing use cases, most likely data repair in the context of data retrieval.

Implications for the vendor/user:

  • Given that the REPLACED_BY association target is not semantically equivalent the end-user should be offered the opportunity to provide an alternative replacement that better represents their intended meaning.

POSSIBLY_REPLACED_BY

For the Author:

  • The purpose of the POSSIBLY_REPLACED_BY association type is to indicate BOTH that none of the suggested target replacement concepts are semantically equivalent to any meaning implied by the inactivated concept (ie neither SAME_AS nor POSSIBLY_EQUIVALENT_TO can apply) AND that more than one potential replacement exists, only one of which will be appropriate.
  • The stated POSSIBLY_REPLACED_BY targets represent all of the alternatives identified as a result of the reclassification process, one of which, although not semantically exactly the same as the inactivated concept, is considered to be clinically acceptable for each individual patient previously coded to the inactive concept.

Implications for the vendor/user:

  • Where a POSSIBLY_REPLACED_BY association is encountered this implies that a decision is to be made by the end-user as to which of the available POSSIBLY_REPLACED_BY associations most closely represents the original meaning stated by the clinician.
  • The vendor must provide adequate tooling to enable the end-user to view all of the POSSIBLY_REPLACED_BY associations and select the most appropriate one to replace the inactivated concept.
  • Because the offered replacements are not semantically equivalent to the inactivated concept the end-user should be given the opportunity to provide an alternative replacement that better represents their intended meaning.

ALTERNATIVE

For the Author:

  • The identified ALTERNATIVE substitute(s) are offered as semantically "sufficiently similar" to the original inactive concept, for certain unspecified use cases.

  • In most circumstances, the ALTERNATIVE concept is likely to be a grouper concept a few levels higher up the hierarchy.

Implications for the vendor/user:

  • The user should decide whether the alternative provided is appropriate for their use case and if necessary or appropriate replace it with a concept of their choosing

Editorial Guidance

  • No labels

6 Comments

  1. "Concept moved elsewhere" may also be moved between hierarchies (semantic roots of SNOMED) as in the past when Clinical findings were moved to Events or as planned now for Procedures to Observables.  Although they may remain SCTID, they should be tracked historically.

    "Note the meaning of the concept is based on the FSN" should be revised: "The semantic meaning of the concept are interpreted from the FSN AND all defining relationships including the supertype relationship 'IS_A'." 

    "The inactivated concept is an outdated concept that is no longer considered to be clinically acceptable or semantically interoperable internationally." This should also support advance of the state of scientific knowledge which ambiguates or changes the concept definition

    Define acronym 'EPR'

    "No other active concept either does - or could ever - exist..."  That is rather arrogant given the evolution of clinical science.   I would strike "or could ever".

    Based upon the discussion, I assume that the cardinality of SAME_AS is 1:1 and EQUIVALENT_TO is 1:*  ???  I would add the cardinality in the advice to authors as it is sometimes important in the selection of an ASSOCIATION.

    1. James R. Campbell, I don't think we have ever used the MOVED_ELSEWHERE when we moved a concept from one hierarchy to another as in Events and situations.  In those cases, the concepts were not inactivated, so there would be no inactivation reason.  For example: 276507005 |Fetal death (finding)| in 2002-01-31 was moved to the event hierarchy in 2006-01-31 as 276507005 |Fetal death (event)| with no inactivation.

      As for "Note the meaning of the concept is based on the FSN", my impression is that the modeling has been derived from the interpretation of the intended meaning of the FSN, so it all begins with the FSN, even when the terming of the FSN is not completely clear (but is used clinically).

      EPR = Electronic patient record

    2. Hi James R. Campbell,

      Many thanks for your comments. Changes have been made and I believe most of your queries answered.

      With respect to the cardinality issue you raise, this applies to a number of the inactivation reasons and our proposed way of dealing with this is to provide prompts and advice as part of the tooling driven by QA rules. Therefore, in the example of EQUIVALENT_TO, we would have a rule that says "there must be at least 2 EQUIVALENT_TO target concepts". The author will not be allowed to complete the inactivation without providing at least 2 target concepts and there will be a prompt to ask whether all EQUIVALENT_TO targets have been identified.

  2. Summary definition for Concept moved elsewhere – Use international edition instead of international core. Our Glossary notes the usage has been deprecated: https://confluence.ihtsdotools.org/display/DOCGLOSS/SNOMED+CT+core

    SAME_AS – Vendor/User implications. There should not be a requirement to review but I question if we should we advise 'there is no requirement to make the end user aware'?

    The EQUIVALENT_TO association is not listed in the first table of inactivation reasons to show with which inactivation reasons it should be used with.

    POSSIBLY_REPLACED_BY

    • The statement “The stated POSSIBLY_REPLACED_BY targets represent all of the alternatives identified as a result of the reclassification process, one of which, although not semantically exactly the same as the inactivated concept, is considered to be clinically acceptable for each individual patient previously coded to the inactive concept”. Suggest remove ‘for each individual patient previously coded to the inactive concept’. We don’t and won’t have access to these patient records to make that determination. Suggest change the end of this sentence to read ‘would be considered a clinically acceptable alternative’
    • With the sentence “Because the offered replacements are REPLACED_BY the end-user should be given the opportunity to provide an alternative replacement that better represents their intended meaning” is this intended to say that when the replacement is chosen by the end user that the historical association would then by REPLACED_BY?

    Agree with Jim’s comment on removing “or could ever” on the sentence - No other active concept either does - or could ever - exist..."

    With EPR - we have EHR in our glossary. Should be consistent. 

    1. Cathy Richardson, many thanks for your comments - all of the suggested changes have been made.

  3. Basis for a concept's meaning: I strongly agree that the FSN should be the source of a concept's meaning. Concept definitions may change based on the modeling philosophy at a specific time. Additionally, subtypes and supertypes may change as concept definitions change or based on concept additions and/or retirements. Therefore, relying on a concept's definition, supertypes and/or subtypes to infer its meaning is less reliable than relying on the FSN.

    Minor Typo: Equivalent is misspelled (POSSIBLY_EQIVALENT_TO) in the Historical Associations section.