openEHR as a Selfless Platform

After reading It’s Time To Destroy Selfish Platforms by Daniel Siders, who is one of the architects of the Tent protocol, it seems pertinent to raise the profile of openEHR as a platform that is designed to be open, diverse and provided from different suppliers using different technologies. It is the second part of this excellent article that is important; where he draws attention to the difficulties of the current platforms on offer. Siders draws attention to the difference between email (where you can use whatever app you like and share information with other apps) and platforms which provide apps which are specific to that platform.

“Ultimately all platform companies fall prey to the same fallacy, that they can build the API to rule them all: a single platform for all the world’s users, applications, and data. This goal and their position of immense power incentivizes many platform companies to become bad actors in their ecosystems, hurting developers and users alike. Platforms act as intermediaries between users and developers, depriving both of a direct relationship with each other and an open market. This consolidation and potential for abuse of power is unsustainable for the ecosystem as a whole. Systems of significant value to society always becomes commoditised eventually. Initial development of technology is often centralized, closed, and proprietary, because this allows for more experimentation and faster iteration–which is beneficial to consumers. Once a product permeates the market, competing alternatives emerge. Eventually these systems need standardization, interoperation, and in some cases, disruptive intervention.”

To paraphrase Siders’ problem with current platforms, which he likens to file systems and protocols before the Web, it is unimaginable that we would have to have special handsets and usage plans to call our friends on other mobile carriers or all have to use the same email application to store and forward emails. These services are based on standard protocols. Current platforms behave quite differently, “Users are forced to be a citizen of only one platform cut off from their relationships, data and applications that are the exclusive citizens of another.”

Sider’s ambition is to create a protocol (called the Tent Protocol) which allows us to have our data on diverse platforms using the same apps to access the data wherever we choose to store it. Some readers may not remember that before the Web users had to dial up to notice boards and download what they wanted. The Tent Protocol aims to store and retrieve our data, synchronise devices and communicate with those we want regardless of the platform. There will be many providers.

My immediate thought is that this is the aim of openEHR although focussed on health records – we are a generous or selfless platform and although we have not developed a specific protocol as yet, the early specifications for the entire platform have begun. Many implementations are developing APIs that are necessary to support applications and the community is endeavouring to keep these aligned. It would not seem too far off to have a URL something like “https://openehr.tophealth.com/medicalimages” return all your radiology reports from one online health record provider and a little further in future a URL like “openehr://redcross.org/medicalimages” might return them from another provider. As a citizen I could consolidate these on another providers site or maintain a list of all my providers myself and log in to these sites individually. Clearly it is likely that money will change hands to provide these services without encumbrance but advertising or public funding may play a role.

More from Siders’ vision. “Applications need to be portable between service providers and accounts on different service providers need to be interoperable. It should not matter what provider you use–you should be able to communicate with anyone on any provider, using any app, and be able to take your apps with you when and if you change providers.” This should certainly be true for our health records. The basic tenets of the Tent Protocol are listed here.

It is the integrated openEHR stack that will put us firmly out in front of EHR developments: from ‘crowd sourced’ clinical modelling (to specify content) which is governed and transparent, to multiple applications working from the same EHR service, to those same applications (and new ones) working on different services, to secure sharing and consolidation of our health records and finally to a distributed set of services that support health care.

While this was a dream when we set out 10 years ago, the reality is now firmly within our grasp. Thank you to all who are contributing to this ground changing effort.

Sam Heard

Chairman, openEHR Foundation

Clinical Modelling

It is clear that specifying clinical (and other health) information for use in communications and shared health records is becoming an important enabler. Stan Huff, from Intermountain Healthcare has re-initiated efforts to get international consensus on how to do this under the banner of CIMI – the Clinical Information Modelling Initiative. openEHR and the Archetype Definition Language (ADL, ISO 13606.2) are seen as key technologies as are UML and OWL.

Of course, even if ADL is chosen in this forum there will be dissent: about the data types (Grahame Grieve is on  his 3rd iteration for HL7 and ISO); about the reference model (ISO 13606 is up for review, the England NHS has its LRA as does Singapore – although they differ); and the governance arrangements for such efforts. Actually, these disagreements are minor if ADL is seen as the technology of choice to do this.

What will be important will be to bring the current openEHR community headlong into the center of the international effort. Give people who are doing this work a high profile, help them be known and recognised. This will take some effort and some of those new to the effort or disenfranchised by such a decision will continue on other paths or develop new approaches.

The aim is to achieve interoperability. It is taking far too long. When I was talking about the rise of modelling approaches within Health IT with an international expert, he thought me grossly impertinent to suggest that this was a potentially more productive approach than HL7 version 3. “How could it be better when we have had 100s of experts working for 15 or more years?”. I think time will show that is not the way to go.

If things do move towards the openEHR approach, I am very keen that the Foundation does not grow into a bureaucracy and does not seek SDO status. Rather, the funding for work is channeled through the organisation when cooperation is required but generally goes to those doing the work. Also, any solutions we come up with as a community are then put to one of the myriad of standards bodies already working hard with overlapping aims.

We need new ways of doing things working with the agencies that are already in place.

A further thought on licenses

There is more concern about licenses and posts on the subject on the openEHR lists. People of the more technical persuasion have more to say, and have more experience with discussions about licensing. They are more concerned on the whole about any negative impact of licenses.

Archetypes and Templates represent clinical agreement on what to store, forward and share. This is a fundamental aspect of what an EHR is and what it can do. So we need to be careful to ensure that the licensing meets the needs and aspirations of clinicians and other experts who are going to author and review these.

The Board discussed these issues in detail the other day. Our collective view was that we are only interested in ensuring that the ‘semantic expression of content’ was freely available. Any derived artefacts that are not the fundamental expression of content are not. But there are many problems even when we focus here. What of a derived artefact in OWL which captures the semantics? This could become the normal expression in future and thus be a major concern.

Our conclusion was that the authorative source is important for interoperability – either national or international. The fact that archetypes and templates came from an authorative source offered the key element of safety – it provides a ‘quality assurance’ stamp, assuming the source has suitable processes in place. The CKM currently offers an MD5 Hash that enables a user to test that the archetype is unchanged. This approach seems to be appropriate.

So if we can tell where it came from and that it is still the same as it was when at that source, we have some protection. All that remains is probably to ensure that there is no restriction on use and a public domain stamp may be the best way to do this. This overcomes any copyright restrictions – without assignment to openEHR.

It has a way to run yet but I think we are getting close to a solution that will meet everyone’s needs.

At the Outset – SA or No SA

This is not a question about if I have a say or if South Australia should exist. It is a question about the most appropriate license for openEHR ADL artefacts held on the Clinical Knowledge Manager.

The specification of the content by openEHR uses ADL; a language created for the purpose and part of the ISO 13606 standard. The specifications are known as Archetypes and Templates. There is a healthy debate captured by Eric Sundall on the openEHR Wiki which illustrates a gap between technical views and clinico-legal views. Although Eric tends to personalise the argument on this page, the tension is broader than individuals. I guess it will be my first task as Chair to settle this argument to everyone’s satisfaction.

The issue is the balance of concerns. On one hand to not to limit uptake of Archetypes and Templates by commercial and other operations. We could, for example, just make them public domain and let people do what they like. This was our initial intent, balanced by keeping copyright of the Archetypes to the openEHR Foundation, by making it clear that people could do what they want with them, and avoiding any complexity.

This approach raised concerns that openEHR might claim some sort of ownership in the future, and also, “Why should openEHR have the copyright, a license is a better solution”. We came to the Creative Commons site and it seemed to offer a license that meant we could use the archetypes created by others who wanted to retain their copyright but agree to a license for use.

There are two choices that are relevant:

• To allow people to use the ADL artefact and derive works as they wish but to give attribution to the person or organisation that licensed it. This license is known as the CC-BY license (the BY means who it was created by I guess) – there is no restriction on how the derived works are used or licensed.
• To allow people to use the ADL artefact and derive works but there is now a requirement to publish these derived works under the same or similar license. This approach is known as CC-BY-SA or ShareAlike

The latter is of concern because it is viral and means that potentially everything derived would have to be available free to others as well – a concern for software companies building softwares that use these artefacts. On the positive side, the SA will prevent people deriving work and copyrighting it to limit others using it or to sell it. Clearly we only want this to apply to derived ADL artefacts; templates and archetypes. The ability to express extra permissions is addressed in the creative commons license framework as a fundamental part.

I have put up a page on this blog to show what the CC-BY-SA licence for ADL artefacts might look like.

I hope this moves the discussion along.