While packing my luggage for a family holiday in Oz I felt the urge to celebrate end of this year with you and share some news and wishes for 2015.
People keep asking me why such a great thing like openEHR hasn’t taken on the world. Fair enough – I hear repeatedly from people new to openEHR approach and mostly those coming from other sectors. I feel this is mostly akin to our community not being as commercially focussed as similar other SDOs and, to be honest, it has some learning curve. The best anti-example would be FHIR’s amazing success and adoption: simple to learn (I mean it!) and governed by HL7 (do I need any further qualifier here?). Well I have some good news, the openEHR community is bootstrapping itself and now preparing for first democratic elections. Under the hood a much more industry-driven and commercially sensible composure is being adopted and the four programs (specification, software, modelling and localisation) are putting in considerable effort to kick start some good work. Indeed some great work have already started mid year – thanks to funding from our industry partners there is intense Archetype development in progress. I think these are good first steps in the right direction.
What’s happening in New Zealand? I have previously posted (more than enough maybe!) quite a lot about the Gestational Diabetes Registry (an end-to-end openEHR implementation in collaboration with Counties Manukau DHB and Diabetes Projects Trust) and some modelling work with the National Cardiac Registry. Apart from these nothing much (that I’m aware of) really. Since the publication of the Interoperability Reference Architecture which depicted use of Archetypes to create a reference library of clinical concepts (called the Exchange Content Model) initial plans to share NEHTA CKM instance with Aussies didn’t happen. According to this any wire format, such as HL7 v2 message or CDA, would be derived from single source of truth: the Content Model. This would practically be a Kiwi CKM environment where we adopt existing international models and then adapt or create our own content. All I can say is the government is currently waiting to see what comes out of the collaborative effort (to work out a common Allergy/Intolerance Model) between FHIR and openEHR. Obviously the Exchange Content Model can be expressed by FHIR Resources and Profiles too. But I think the right way to approach this would be to use openEHR modelling for logical representation of the kiwi health record and then create corresponding FHIR Profiles (including Resources of course but also terminology bindings and other context constraints) from openEHR machinery. FHIR will be the wire format like v2 or CDA.
***Some FHIR musings below*** These are my personal thoughts
FHIR faces the same modelling challenges as openEHR but I think less equipped (from a modelling perspective) than openEHR. For example the distinct separation of Reference Model and Archetypes is traded-off for simplicity and in FHIR you’d see Resources (roughly corresponding to Archetypes) like Composition, List etc. which are expressed at RM level in openEHR. So you can still express stuff but when it comes to dealing with longevity and complexity of clinical information openEHR’s multi-level modelling is a more elegant approach. But having implemented FHIR myself this year at the HL7 New Zealand FHIR event I must say it is SO attractive! First of all the “implementer friendly” motto is not a lip service (which I thought it was before) and what is astonishing is any Web developer who’s comfortable with modern web development tools etc. can just do it. They indeed become fanatical about it. How cool is that?
So my take from this year is that FHIR is a great way to implement health information exhange and I think openEHR better suits to health IT environments where a high level of semantic precision is needed – such as advanced decision support or an integrated health information system or research/population health platform. This view has matured a lot since I got into Biomedical research – some of you may already have noticed I started a new role at Auckland Bioengineering Institute (while still part-time with NIHI). There is fascinating work going on, both in NZ and around the world, what I’d describe as Virtual Physiological Human. Basically to create a digital representation of human including ALL processes. I would strongly advise that you have a peek at this video. Digital Patient is also a related and popular topic where openEHR can be a good fit. Indeed the VPH-Share project funded by the European Comission has made extensive use of Archetypes to link biomedical domain with clinical data.
This brings the crucial importance of aligning content between FHIR and openEHR communities – having experienced the joint Allergy/Intolerance modelling effort and after a chat with Grahame Grieve this week it is clear it is not the technical differences of the two formalisms but the way their communities approach to modelling can be a real problem. So what I’m saying is if clinicians agree on a certain way to model things both FHIR and openEHR can accommodate it – no big deal. The issue is how to get the two organisations to the same requirements – this would be a challenge. Obviously CKM is a great tool for this and I hope FHIR community will realise it. I decided to put some effort in mapping existing Archetypes to FHIR resources in parallel with joint modelling efforts. If we fail to achieve this alignment, a golden opportunity will be lost – forever…Therefore my wish is to set openEHR on FHIR this year!
Merry Xmas and a very happy new year indeed…