Minutes

Initial comments from GRE:

-          How will the refactoring of substances affect concepts that need them – allergies

-          Role group 0 discussion is controversial, need a more detailed document for discussion in London.  He does not feel the proposal is persuasive.

Drug model update

-          Second version of editorial guidelines distributed

-          Third version will be distributed in mid-April

-          Final determination of changes for July 2017 and Jan 2018 under development

-          Several new concept model attributes added for July; several new semantic tags also added

-          Two new hierarchies: number and “Unit of presentation”

KCA had a question about the use of “virtual” in the semantic tags.  It was suggested that if better terms were available, we would consider them.  GRE said that it is not clear that “actual” products would be direct descendants of “virtual” products.  The MAG will be addressing this issue.  The main point is that the SNOMED International release will not be containing the actual products.

Items for discussion in London (GRE):

-          The potential approaches for resolving “roles”.  It impacts what extensions will be allowed to do.

• Necessary but not sufficient set
• Technical approach needs to be discussed with the editorial group

-          Dose forms and concept model for dose forms – needed for the January release

-          Routes of administration – how it is used in the International release and how it might be expected to be used in extensions

Substances

-          Scope of content changes for the July 2017 release have been determined

-          Review of groupers underway

-          New attribute HAS DISPOSITION developed to allow for definition of groupers

-          Editorial guidelines will be available for review when test plan completed

Discussion on levels

-          It is unclear what is meant by the term “X level”.

-          PAM expressed the current status in the UK that ICD is general is not a practical approach to recording injuries.  The UK has produced their own injury concepts that emergency physicians consider more useful.  Concern that these concepts do not represent what is needed and what is the value of adding these as opposed to what is needed by physicians.

-          KCA are the semantics understandable and representable and is it worth the effort to do it.  If level meant “region” than is it worth the effort to add these?

• Might be useful to develop the modeling pattern but not add the concepts until they sre recognized as being useful

-          Do not want to have the classification drive the clinical content of SNOMED CT

-          Anatomic regions have some use, but is this the right way to represent these injuries?

-          BGO: for amputations there is a need to designate where the amputation occurred

-          GRE: “Level” is ambiguous.  These could be added as primitive concepts if they look to be useful.

-          JCA proposed that these not be added based on their lack of clinical usefulness. 

-          GRE: These could potentially be added as an extension.

-          KCA suggested that Adding these as primitives defeats the value of adding them to the terminology as we would still need to do mapping because the terminology cannot be used to create the linkage.

-          Conclusion: Because of the perceived lack of usefulness of these terms for data entry and the need to create and maintain a mapping to ICD-11 limits the value in adding them to SNOMED CT.  They may be good candidates for an extension or another moduleID.  They will not be added to the International release.

Allergy model

-          BGO reviewed the most recent iteration of the Hypersensitivity concept model.

-          GRE proposed more testing to be done to ensure that there are not unexpected results.  He asked for access to the task in SCA that would allow him to review this.  What is needed is to determine the amount of maintenance and the resulting classifications.  What does the transitive closure look like between this model and the current state.  The addition of intermediate primitive groupers is not desirable but may be necessary in this case.

• How these terms might be used in clinical practice needs to also be considered.

Complications and Sequellae

-          BGO went over definitions of “Complications” and “Sequellae” as well as the differentiation surgeons make with regards to sequellae.

-          All of these can be modeled using the patterns developed for combined disorders

-          Initial proposal is that “Complications” and “Sequellae” be retired and descriptions that use the terms be retained for retrieval purposes.

-          GRE recognizes the problem, but is concerned that the solution may cause more problems.  Making something a synonym of something because the terms are not always able to be differentiated.  This is a language specific issue. 

• Is a complication a role?
• Is it always unexpected?
• How do we determine the relationship?  Is a new attribute needed?

-          The attempts to model using temporal and causal relationships is elegant, but may not be totally satisfactory.

-          In general, there was not consensus that this proposed solution would solve the problem.

-          The source of many of these comes from ICD.  Many of these are judgment calls as to what is a complication or sequella. 

-          It may be difficult to do this in the core due to these issues.  The proposal needs to be thoroughly reviewed by the consultant terminologists.  This needs to be developed and tested more fully before a final review and approval by the EAG.

Patient-friendly terms

-          This is an English-only issue.  Translations can add whatever terms they would like to their extensions.  The initial proposal was not specific to English, but it is implied

-          The extension mechanism was designed to address this type of issue.

-          KCA asked why “dialect” was not suitable to resolve this issue.  There are mechanisms in place to handle this, and why are these not suitable to meet the need?

-          GRE: Two aspects, 1) the recommended mechanism to handle this and 2) will the International edition handle these terms. 

• The patient would need to have a separate language refset in order to always present the patient friendly terms to the user.
• In the Netherlands, they added these in a separate extension in a language refset. These may contain specialized consumer terms and some commonly clinically understood terms.
• It is important to be able to separate them from the core, especially when the temrs are less specific than the core concept.
• There was a mechanism in the original RF2 structure (i.e. context). But that was not included in the final specification
• A partial solution does not meet the needs and is not a suitable solution.  However, this only affects the English editions.  What types of support do we have in the tooling to be able to keep these “separate” from the core concepts
• We need to be able to recommend to extensions how to handle these.

-          PAM: This is both language and culturally complex.  Different local descriptions are possible even within the English language.

• Not sure what is really being asked for and how this proposal might meet that requirement.
• The potential number of “preferred” terms is relatively unbounded and would not be maintainable.
• Is it possible to “crowd-source” a set of PFTs that would help to meet the requirement?  What is the scope of the PFT coverage?  All of SNOMED? 

-          A new version of this proposal will incorporate these comments and be reviewed again in London

We need to propose the recommended way that PFTs should be handled using the existing mechanisms. We cannot violate the guidelines for adding descriptions to the core.