Extension Management Proposal

Reaction of the problem bullets in chapter 5:

Bullet 1-3: We already do this, so agree a NRC must be able to perform this actions.

Bullet 4: Can be useful, it allows us to create intermediate primitive concepts and not just leave primitive concepts.

Bullet 5: When bullet 4 is possible, this also should be possible, to ‘remove’ unnecessary ISA-relations . But this only should be done by the classifier!

Bullet 6: When you want to add a relation to a concept to add definition (and not change definition), the first path is to request the addition with IHTSDO. I see 2 exceptions: 1) when bullet 4 is allowed, a extra parent is added, the inferred relations should be added by the classifier. 2) When a country creates his own attribute and wants to use it in a core concept. No idea if any country already has done something like that J. The additions are stored in the national extension.

Bullet 7: We don’t think you want to do that. When a concept contains an error or is inappropriate, request a change with IHTSDO and don’t use it in your refsets, ignore the descriptions. If IHTSDO takes too long to decide, create a new concept with the right content and use that concept. The only point is that you can’t create a concept with the same name.

We think it’s not the right way to inactivate a concept that others still can use.

So in summary: We agree with this paper except for the last bullet, retiring core concepts that are inappropriate. It must be stated that most additions or removing actions only should be done by the classifier (by adding newly created concepts as parent).

I must say that most of it is already be done by us, by using the classifier, and I don’t know whether we are violating the license with that. It is clear that clarifying the license text  is necessary.

I’ve raised this issue with our NRC, interested parties via a SNOMED CT users group, and directly with individuals I know are working with SNOMED. There has been little response received from these areas; I suspect due to the lack of current extension development activity in New Zealand.

From what I did glean there is recognition that international core content should not be tampered with, and nervousness regarding collaborative editing unless very well managed.

There is interest by extension developers to be able to create fully defined terms within their extensions and use a classifier; therefore, the increased flexibility raised by Matt would be welcome.

Overall there is support for Matt’s proposal.

We have discussed the proposal in the Danish NRC and can of course recognize the issues.

The Danish recommendation for making changes to the international core of SNOMED CT is that changes should be limited to the terminology team in IHTSDO and that IHTSDO should increase its terminological capacity to match the demand. We think that strict governance of the terminology is essential to obtain the quality that we all expect SNOMED CT to have. High quality of the content of SNOMED CT has always been one of our main issues when discussing new project or making suggestions for the budget in Member Forum.

To support a more agile development of the content we suggest that new tools are developed to make it easy for Members to create small models of new terminological content that can be submitted to the IHTSDO terminology team. This could e.g. be using graphic elements from the Single Content Authoring Tool to illustrate how the content should be modeled and members should be able to submit this through the request submission system. Such a tool would also make it possible to make cooperative authoring possible. This way it would be much easier and faster for IHTSDO to pick up and implement the suggestions.

If single Members are allowed to add changes to the core content we fear that focus would be on issues that are of interest to the single Member and not have a more holistic dissemination that we hope going through IHTSDO will entail.

Anyway, a much more detailed discussion of the implications of the different paths is necessary in the wider IHTSDO community to discuss pros and cons as well as governance (for instance: who will decide if two or more Members wish to model core concepts in different ways?)

Just a few high-level comments.

1) The international release should always have precedence and be the reference.

2) Any changes made in an extension that conflict with the international release are not acceptable and issues need to be resolved with the international release first.

3) In general, allowable changes made to concepts from the international release through an extension should be submitted to the international release, even if they do not create conflicts.