My experience in eHealth started as a suburban General Practitioner using an EHR application for prescribing and clinical notes, and then I moved sideways, becoming involved in building proprietary clinical software and a Personal Health Record. From 2000 I worked for 4 years in a single company that owned that PHR plus a primary care clinical application and a hospital application – the intent being that data created in one could be transferred between the systems, but we found it wasn’t that easy – for various reasons they all had different database structures, even within the same vendor! So if we have to do the same thing between disparate vendors in an environment that is more competitive than collaborative, the picture becomes infinitely more complicated.
In more recent years, I have had my world view shifted from the traditional application-drive EHR to a data driven health record (note the deliberate lack of capitalization) – see my previous blog posts. Once we focus on getting the data right, capturing it or displaying it in the applications are just one of the many things we can do with the data.
Why? I first heard about openEHR nearly 10 years ago. I didn’t understand openEHR at all initially, but there was something in the commonsense of getting the foundation data defined and standardized that resonated with me. Over time I have become convinced that openEHR provides an orthogonal approach to eHealth that has a very reasonable chance of success, and more importantly, of making a difference. I no longer believe that the traditional application-driven approach to electronic health information management is effective, economic or sustainable.
What is openEHR?
Think of openEHR as the open source health equivalent of the iPod/iPhone platform – a technical framework which will allow any compatible application, organization or provider to share ‘plug and play’ access to standardized data. This is openEHR’s innovation – the focus on ensuring that the underlying health data is correct, robust and trustworthy!
Rich health data definitions known as archetypes, are defined and agreed by the clinicians themselves to ensure that each piece of health information is unambiguously understood, ‘fit for purpose’ and can be dynamically used & reused to support wise and safe health choices, now and into the future. These same archetypes are also computable, so that when these common data definitions are shared, they act as a ‘lingua franca’, making it much simpler to capture, store, aggregate, query and exchange health information – effectively making the data ‘sing and dance’ and to flow according to privacy and access rules.
Developed over more than 15 years through international research, community input and implementations, openEHR is purpose-designed as a non-proprietary universal health record: application independent, yet supporting accurate and safe health information exchange between software programs, consumers, health care providers, organisations and researchers; and across the diverse requirements of private/public providers and regional, national and international jurisdictions.
- It is open source – break down the proprietary silos of data created by application vendors.
- It is the basis for the recently published ISO13606 standard for EHR extract. openEHR evolved and grew away from 13606 approx. 5 years ago as 13606 entered the CEN and ISO standards approval process; openEHR has subsequently progressed and developed as a direct result of implementation experience. At present openEHR is the commonest implementation pathway for nations mandated to adopt the 13606 standard.
- It is driven by an international open source community.
- It has been developed using a robust engineering process.
- The clinical content is driven by the domain experts – usually, but not limited to the clinicians themselves – through the Clinical Knowledge Manager.
- Archetypes are designed as maximal data sets for the universal use-case so the same data definitions can be used in any software application – whether a PHR, EHR, research project, clinical decision support system or running population queries.
- It is purpose-designed as a shared health record.
- The structured archetype definitions complement other standards-related work – for example the recent announcement of a collaborative work program between the openEHR Foundation and IHTSDO to explore how the SNOMED-CT and openEHR archetypes can be combined to provide a strong semantic solution for health information.
Who is using openEHR?
International momentum is building – current users noted on the openEHR website include commercial, government, academic, and non-profit organisations.
Current eHealth developments are progressing at a glacially slow rate because we are trying to develop interoperability by traditional and incremental methods. I’m increasingly sceptical that this is effective, economic or sustainable. And in fact, though direct experience I have become convinced that openEHR’s orthogonal approach can make a significant difference. I now work with openEHR every day, including:
- alongside an international group of like-minded clinicians collaborating in their spare time in a Web2.0 application to develop, publish and govern the archetypes;
- with national eHealth programs who are seeking to build a library of clinical content definitions as a ‘single source of truth’ to mandate for use by vendors;
- with vendors who no longer have to reinvent the wheel but can take the published archetypes and re-use them within their systems; and
- with researchers trying to aggregate and integrate disparate data from multiple sources into one cohesive repository on which they can query their valuable data.
I’m not trying to sell you a software application like everyone else claiming an ‘eHealth solution’; I’m trying to persuade you to take a look at an alternative approach to eHealth, to share a little of my vision and my passion!
Consider becoming part of an international open source community. Contribute and collaborate alongside other clinicians, informaticians and techies to build something bigger than all of us. Share a vision of how eHealth could work better and more effectively if we get the foundations for a universal health record consistent and solid.
It. is. all. about. standardizing. the. data.
Pingback: Jerry Fahrni » “What’d I miss?” – Week of February 28th, 2010
Pingback: Food for (lots of) thought
Very nice this post, I will study about the openEHR now! ’s!
Pingback: Google Health – The Sacrificial Lamb « Youth Health 2.0
Indeed, it is all about standardizing the data. Content is king. The rest is technology, and we have working solutions to transport the data. We have defined/built the railroad system and now we need to build some wagons to it.
I would like to find out how OpenEHR concepts and all the achievements of HL7, starting with the RIM, go together such that they can benefit from each other (and btw, what is “orthogonal” about the OpenEHR approach?)
The orthogonal part is the two level modelling, with separation of technical and clinical knowledge domains. Added to that, the collaborative focus on the content doesn’t is quite unique to the tradition approach of building siloes of proprietary data. At HIMSS12 last month I had people laughing and saying that the vendor in the booth down the hallway didn’t want to share data structure – they wanted to keep the vendor lock-in scenario and their solution to interoperability is ‘buy more of my system’.
Interesting. By coincidence, I am about to evaluate a clinical archiving solution that claims to be open in that they deal with clinical information in any form (HL7 CDA, PDF, etc.) as objects (original bit pattern preserved, no transformation of any sort), separate technical from clinical data and allow to generate metadata to each object (automatic analysis of known formats or custom made filters). That seems like a good way to avoid vendor lock-in. Assuming that OpenEHR offers another means to structure data, it would be just another animal in the zoo that the mentioned solution can handle. I probably miss some important facts about OpenEHR, do I?
Interested to hear more about your alternative.
I promised to report about my findings from my evaluation of a clinical archive. I am aware that this may sound like a sales pitch, but what the heck, I found my preferred solution and until someone else shows me a better concept – this is it!
It is the Hitachi HCR (Hitachi Clinical Repository)
Why do I like it so much – (just my personal two main points)?
1. I fully trust the underlying technology, you can start small and scale up inifinitely, reuse existing storage and administrate the virtualized storage with comfortable tools (even remote via an iPad app) => I won’t have to worry anymore about storage technology for years, end of the story.
Hitachi has been an OEM for IBM mainframes and for storage components since decades, so they know how to store data and, probably less known, how to build computers.
2. any information system in a hospital can store its results, docs, images etc. in one archive
(causes fear in the minds of traditional hospital IT managers, I admit, but look at the advantages)
now that is the crucial point: not only can you safely archive clinical documents of any kind (structured and unstructured, HL7 CDA, DICOM, PDF, Word, you name it) and retrieve it in its untransformed original state in x years, but the HCR produces custom meta data to each object it stores. (and you can teach the HCR to recognize more formats, or analyze unstructured stuff ..)
Figure all your docs with meta data residing in one repository – now search over all docs, or their metadata, respectively, in one place …
In this context OpenEHR would be another way to structure data and enable the generation of useful metadata in the HCR and therefore help to make data better searchable?
challenge me 🙂
I’m just trying to get a hang of openEHR ans Archetypes…. What i wanna know is what exactly can be done with these archetypes…. How are they different from creating a database and storing information in them…. How are these records saved and how does one use them ???
You ask a question that is impossible to answer easily here:-). The underlying premise of openEHR is separation of the clinical and technical domains – the notion of two level modelling. The full openEHR technical specs can be found at http://www.openEHR.org. The focus of this blog is largely on the clinical side – ie the archetypes that are computable representations of the clinical content. I’m not a techie, and don’t need to understand the ins and outs of building a fully functional EHR but I can build archetypes that will be consumed within an EHR; similarly the techies don’t need to understand clinical medicine expressed within the archetype.
Governed archetypes eg those found at http://www.openEHR.org/knowledge, shared between systems become the common pattern for storing clinical data – whether it be in an openEHR repository or another approach. By using the same computable patterns it makes it orders of magnitude easier to share health information unambiguously.
I suggest that you engage with the openEHR community via the email lists – technical, clinical, implementers etc – http://www.openehr.org/community/mailinglists.html. There is a large community ready and willing to help, and many sharing experience from systems they have built.
Pingback: “What’d I miss?” – Week of February 28th, 2010 « RX Informatics.com
Pingback: Setting up a MEDrecord Open Source openEHR server, for DUmmIeS. | opengpsoc
Heather, I think it would be good to add the main purpose and objectives of openEHR on this page: http://www.openehr.org/what_is_openehr because it is not explicitly stated, maybe that can be based on this post. What do you think?
Very happy for others to adapt and reuse this, Pablo. The openEHR website is still largely impenetrable
Yes, I’m thinking for a newcomer it is difficult to understand the basics. Do you know who is in charge of updating the site? Maybe we can propose to add more basic descriptions of the openEHR aims in the What’s openEHR? page. Also I’ve translated the siet to spanish some time ago but don’t remember why the translations where not deployed. I would love to help on that part also. Thanks!