This document discusses Norway's national governance of clinical archetypes. It provides an overview of Norway's public hospital system and use of openEHR archetypes. A national archetype governance scheme was established in 2013 by National ICT Norway to develop high quality archetypes through a review and approval process. The scheme aims for semantic interoperability through shared archetypes. Key success factors include clinician involvement, appropriate tools, dedicated resources, and international collaboration. While progress has been made approving archetypes, continued challenges include translation efforts and aligning archetype development with review timelines.
3. Norwegian public hospital system
• Four Regional Health Authorities (RHAs)
• 24 Hospital Trusts
• 100 % EHR adoption
• No primary care!
• Two main EHR vendors
• One strategic coordinating
health trust for IT (National ICT)
4. openEHR and archetypes
• Specification for structured
electronic health records
• openEHR Foundation (openehr.org)
• Free use
• International community
• Archetypes are basic data models
• Templates are data sets made from
archetypes
Same set of
archetypes
Template 1 Template 2Illustration: Heather Leslie. Used with permission.
5. Archetypes
• Reusable data models each describing one clinical concept
• Maximum data sets that are combined and constrained to fit
each use case, making templates
6. openEHR in Norway
• Basic clinical use at Oslo University Hospital (DIPS)
• Several vendors are implementing openEHR
• National ICT archetype governance aims for
– high quality archetypes
– semantic interoperability through use of identical
archetypes
7. Real semantic interoperability requires identical data models
Clinical data modelling is difficult and expensive,
and should be done once
⇒ Archetypes should be shared
and strictly governed
(We can be a bit less strict with the templates…)
8. Archetype governance history
Early 2013: National ICT asks the Western Norway
RHA to develop a governance scheme for archetypes
Mid 2013: Bergen Hospital Trust develops the
governance scheme
Late 2013: Governance scheme is approved by
National ICT
Late 2013/Early 2014: Tools are procured. Scheme
is deployed. arketyper.no goes live.
10. Processes and tools
• Formalised processes for development, review and approval
– Archetype development is a “do-ocracy” (Sw: “görokrati”); not
centrally prioritised
• Online collaboration through two tools
– Clinical Knowledge Manager (CKM) arketyper.no (@arketyper_no)
– Documentation wiki wiki.arketyper.no
• Both tools are critical for transparent online collaboration
11. Reviews and approval
• National editorial committee defines review requirements
• Review rounds last one week, several rounds per archetype
• Until clinicians reach consensus…
• If requirements are met, the archetype is «Published» in
arketyper.no
• Unpublished archetypes are unstable and should not be
used
12. Approach
• Supporting efforts by healtcare providers and
vendors
– Reviewing and approving specific archetypes
• Reviewing and approving basic concept archetypes
– Observations, diagnoses, symptoms, procedures, …
• Spreading the word
– Presentations, courses (next May 19 and 20 in Bergen),
workshops
13. Status
• During 2014 we only managed to approve 6
archetypes, 5 of them in December
– Clinician participation is crucial, and hadn’t
reached a critical level until late in the year
• As of April 2015 there’s 12 approved
archetypes, 10 more under review
15. More simple statistics
Clinical MD
33%
MD/IT
7%
Clinical RN
31%
RN/IT
16%
Other clinical
7%
Other
technical
6%
Technical
Professions
16. Success factors
• Clinician participation
• Good tooling
• Resourcing for coordinators
• Resourcing for training and consulting
• Collaboration with international
community
17. Pitfalls and difficulties
• Translation is difficult and
time-consuming
• Development projects don’t
take into account the time
needed to review and
approve archetypes
• Regional resource groups
are hard to get going
18. Summary
• Clinical data models (whether archetypes or other) must be
shared freely, and governed tightly
• Key success factors are clinician involvement, tooling,
resourcing and international collaboration
19. Have questions but don’t like microphones?
• Stop by the Q&A sessions by SFMI later today:
– 12:30-13:00 (23/4) Frågestund 1, med primärt fokus på norsk och
internationell arketypförvaltning, kliniskt engagemang och innehåll (Sal A7)
– 16:30-17:00 (23/4) Frågestund 2, fortsättning samt breddning till teknikfrågor
m.m. (Cambios monter, B05:01)
• Go to Erik Sundvall’s presentation in track Verktyg för entydig och
strukturerad dokumentation in A3/A4 at 14:30 - 15:00 today
• Come ask me in person (I’m easy to spot in a crowd, and not at all scary)
• Tweet me: @siljelb
• Email me: silje.ljosland.bakke@helse-bergen.no
Hinweis der Redaktion
Hei, jeg heter Silje og jeg er sykepleier.
Tusen takk for at jeg har fått komme og snakke om arketyper her i dag. Jeg jobber til daglig i Helse Bergen, som driver Haukeland Universitetssjukehus, men jeg er så heldig at jeg får bruke omtrent all tiden min på det gøyeste i verden, som selvfølgelig er arketyper. Det gjør jeg gjennom at jeg er utleid til Nasjonal IKT, som er en nasjonal, strategisk organisasjon for IT i spesialishelsetjenesten i Norge.
Før jeg begynner med selve presentasjonen vil jeg gjerne vite om det er noen som synes dialekten min er vanskelig å forstå. Jeg vil gjerne at dere skal forstå hva jeg sier, så hvis det er vanskelig så vift med armene nå så skal jeg bytte til engelsk.
four Regional Health Authorities
24 Hospital Trusts between them, each running one or more hospitals
All of these are using EHRs, with mainly electronic documentation
As opposed to the Swedish system, the Norwegian hospital system is organisationally separate from the primary care system. This has and is still causing a lot of problems with continuity of care.
All Norwegian hospitals use one of two EHR vendors; DIPS or Siemens. Siemens is used in the Central Norway Regional Health Authority area, and DIPS everywhere else.
National ICT
Jointly owned by the four Regional Health Authorities
National ICT was created as a proper health trust in 2014
has existed as a collaboration forum for the RHAs since 2003
main responsibility is strategic coordination of ICT in the specialist healthcare system.
Specification for structuring and querying data in electronic health records (EHRs)
Owned by the openEHR Foundation
Specification and artefacts are free to use, Creative Commons licenses
International community of implementers and modellers
Archetypes are the basic data model building blocks of openEHR data sets or templates
An archetype is a data model that describes all the data elements that could possibly be needed, for any clinical use case, to document one clinical concept
Archetypes are combined and constrained to make templates, which are the actual data sets used in systems
Oslo University Hospital A&E started using DIPS Arena in October 2014
Some structured information using archetypes
Mainly accident reporting
Several different Norwegian vendors are implementing structured EHRs based on openEHR
National ICT governs the national Norwegian archetypes
The main goals of the governance is
that the archetypes that are put to clinical use are of good quality through having been reviewed by many clinicians of a wide spectrum of professions and specialties
And facilitating semantic interoperability through enabling organisations and vendors to use identical data models
Once people get their hands on an archetype-based structured EHR, archetypes will sprout like wildflowers, and data models will start to diverge.
I don’t have any proof of this so you’re allowed to disbelieve me, but our experience is that clinical data modelling is hard. It takes a lot of time, and is therefore expensive.
Therefore it makes sense to do this modelling just once, and share the results. To avoid divergence, this set of shared models needs to be governed very strictly.
In 2012, National ICT realised archetypes needed to be governed to avoid the mentioned sprouting of wildflowers.
They then asked the Western Norway Regional Health Authority to develop a governance scheme
This scheme was developed by the Bergen Hospital Trust on behalf of the Western Norway RHA during the summer of 2013
Was approved by National ICT in October 2013
In late 2013 and early 2014 the tools were procured, and the governance scheme deployed. The arketyper.no site went online on February 7th.
Part of the national governance scheme is the formal processes for development, review and approval of archetypes.
The development process is based on a model called a «do-ocracy». Who hasn’t hear of…?
A do-ocracy is similar to a meritocracy, only instead of the people with a proven track record making the calls, the people who actually spend the resources to do something make decisions. This only means that the development isn’t centrally prioritised. There’s no committee deciding what to do when, this is all decided by the actual requirements of whichever projects are going on at the moment.
Important to avoid losing momentum
To be able to do this we’re using two online tools, both of which are critical to be able to involve large numbers of people in a transparent manner
The national editorial committee defines the review requirements of each archetypes, which basically means, which professions and specialities should and must take part in the review
Review rounds last one week each, after which requested changes are made, and a new round started, until the clinicians agree………
If the requirements are met once the clinicians reach concensus, the archetype is published.
Unpublished archetypes can change drastically overnight, and should not be used in a production environment
Our main approach is to support (ie review and approve) all the archetypes needed by healthcare providers and vendors.
Secondly, we do initiate reviews of certain basic, ultra-reusable archetypes ourselves, because we anticipate they’ll be used almost everywhere
We also spend a lot of our time telling people about what we do, as presentations like now, courses, and workshops
This doesn’t sound like a lot, and it isn’t. But some archetypes are more important (in other words, reusable) than others. Once we have say 30 of these very reusable archetypes approved, we can probably represent at least 90% of primary care in a structured way. The rest will be mostly specialist stuff.
The arketyper.no clinical knowledge manager is an open tool where anyone can register for a user account and participate in reviews. They can also record their profession and health domain, which is the basis for our selection of who gets invited to each review. (you can also request to be invited to review a certain archetype by «adopting» the archetype)
As you can se, only about half the users have recorded their profession, which means that even if they wanted to participate, we won’t invite them to reviews because we don’t know what they know about.
All in all, only 43% of the users are actually eligble to be invited to a review, unless they actively adopt archetypes themselves.
This graph shows the proportion of users who are mainly clinical and actually work with patients on a regular basis, and the users who either have a purely technical background or a combined clinical/technical background and don’t see patients regularly. We’re very happy to have so many active clinicians participating, as this kind of thing often only attracts the ones who mainly work in IT.
We’ve identified some success factors to what we’re doing.
Obviously clinician participation is crucial, since the clinicians are the ones who both define and approve the data models.
Next, without the good tools we have we wouldn’t have been able to get the clinicians involved. Doctors and nurses don’t want to read through long PDFs with UML diagrams, and then write their response in a word document.
The coordination of this governance scheme takes a lot of time, and couldn’t have been done without dedicated resources.
This sort of thing is new to almost everyone, and without the means to do training and consulting with international experts, making good models without making too many mistakes would have been much harder.
Finally, since there’s a sizable international community also making and sharing archetypes, we haven’t had to make many ourselves, but mainly translating and adapting. This saves loads of time.
Then the things we could have done better.
Translation has proved to be a tough issue, since not many clinicians are trained to do clinical work in both Norwegian and English. Most of the review comments haven’t been about actual structure, but about language and wording.
Secondly, development projects aren’t used to having to get their information models reviewed and approved before they can be deployed. From our point of view, this would work much better if the models were developed early in the project, and then submitted for review asap
Finally, as I mentioned earlier, the regional resource groups have been hard to get started. This means the job of recruiting clinicians and supporting initiatives fall to the coordinators and regional representatives, which they’re not resourced to do. This can be a huge bottleneck to getting archetypes reviewed and approved.
Summing up, clinical models, which don’t have to be archetypes, should never be proprietary or vendor-specific, but be shared with everyone. At the same time, to enable semantic interoperability, they need to be strictly governed to avoid divergence.
Success factors do cost a little bit of money, but is probably much cheaper than the alternative, especially in the long run.