TOPIC 2: Global Promotion of Content
Assumptions
Content is currently created locally due to the perceived delay in getting content through the process into a published International Edition
There will continue to be data which is not available in the International Edition (will this be initially large volumes but level off once those have been tackled)?
SNOMED International does not have the capacity to author all the content requested by members and the wider community
Some SNOMED CT extensions contain the same content as other SNOMED CT extensions.
SNOMED International
The global challenge is, as always, a capacity one. If too much continues to be promoted, the central team does not have the capacity to review everything, whilst doing other International work.
Can templates be used more locally to enforce editorial and modeling guidelines?
All identifiers (concept, description and stated relationships) would be kept in the promotion.
The workflow still needs to reflect the one currently used by CRS, which would imply that any solution needs to use CRS as the entry point. This is technically feasible, but the functionality would need to be added (import RF2 for example).
Following the CRS workflow still means a concept by concept approach... is this efficient or the right way to do it? What are the other options?
The better modelled the data, the easier to review and include
What training or accreditation should be required by terminology authors who request content be promoted into the International Edition?
NRC
Currently only concept ID is kept, creating overhead and duplication for descriptions and relationships in the local extension. Being able to promote all identifiers will resolve this issue.
If this was done, does this make the international 'upgrade' an easier, less painful process?
How many members have the ability to provide an RF2 of only content to be promoted? Is this a solution only applicable to a minority?
This is also functionality that should be built into the Managed Service by SNOMED International anyway... a concept promotion button directly creating a CRS request.
Implementors
Organizations in countries have to currently go through their NRC, and this is unlikely to change outside of identified bodies who take ownership (whatever this means) of a domain of knowledge.
Is there a more efficient way for local organizations to participate in getting content promoted? In fact, is there even a lot of this very local content?
Are these exclusions to what can be promoted, including interface terms?
Other Challenges
If multiple SNOMED CT extensions contain overlapping content, then how do we decide which version of the content gets promoted?
Will this lead to greater overhead for those NRCs whose content is not promoted, who need to modify implementations to use the identifiers of the content that was selected for promotion
Note: This challenge already exists when new content is added into the international release, which was modelled independently in an extension - However will the scale of this challenge increase? And if NRCs don't resolve these overlaps, then will the existing issues just continue?
Should additional tools and/or mechanisms be put in place to support those countries whose content was not chosen for promotion - e.g. mapping tools, history mechanisms, guidance etc.
Comments
From an NZ perspective and like the US NRC, we intend to have as small an extension as possible, limited to things that are genuinely local. Most new content we believe should be added directly to the International Edition or promoted if it already exists elsewhere.
I support distributed or collaborative authoring, and see more value in centring this on the International Edition rather than using national extensions as feeders.
Would like to have joined you in London but unfortunately won't be there this time. I look forward to following the discussions.
A few thoughts:
Based on a study done a couple of years ago, it was surprising to find that the content overlap amongst three large extensions was very small. So while duplication is a real issue, it is not as pervasive as one might have thought.
From a SNOMED International point of view, issues related to promoted content involve a number of dimensions:
Is the concept valid (i.e. is it represented by an authoritative source in the literature, from a professional society, etc.)?
Is it modeled according to editorial guidelines (i.e. proximal primitive parent, full set of stated relationships, FSN terming, template adherence, etc.)?
This is key as there are multiple "valid" approaches to modeling
Is it international in scope?
Does it represent content that would better be represented by the information model (e.g. state diagrams)?
From an US NRC point of view, the goal is to maintain as small an extension as possible, so any mechanism that facilitates the actual injection of the promoted content into the content development stream would be of great benefit.
WRT "who wins" in a overlapping promoted content contest, as stated above, the current overlap is very small, so the sooner this process is established (i.e. before more overlapping content is developed by extensions) the less likely this scenario will occur. However:
From an NRC perspective, reconciliation of duplicated content (i.e. content promoted to core and duplicating content in another extension) has been a minor burden from a maintenance perspective in RF2.
The much larger burden would be on implementers who have either mapped extension content that is superseded or have a lot of historical data that uses the inctivated extension concept. Tools to provide more efficient use of the historical relationships or data migration would be of benefit. The whole notion of what to do with content that uses a now inactivated concept is a workshop in itself.
Content that is NOT chosen for promotion, should remain local and is highly unlikely to duplication content in the international release (unless it was not accepted due to a duplication with exiting international content).
This is a highly interesting and challenging area for which tooling can provide great benefit. Adherence of promoted content to International editorial guidance prior to uploading would eliminate the vast majority of content development work currently performed on content originating from extensions and allow author to concentrate on clinical validity.