Frankly I am more than a little confused about HL7. From what is claimed on some of these websites it seems to have the capability to tranlate between EMRs which is a pretty bold claim.
I have to agree. My opinion is biased, but I have been working in the EHR area for a while now. Problems HL7v3 has with EHRs:
- an EHR isn't just a collection of mesages. They also include GUI app generate content from online users.
- even when messages are put into to it, they need to be translated, massaged and built into the EHR information structures, versioning, links and so on. So in fact, no sensible EHR will look like a series of messages - it will look like a series of blocks of information, committed to a versioned, shared database.
- the message based approach is not suited to EHRs in my opinion. What gets sent from one EHR system to another cannot usually be known in advance as if it were a stock message structure - e.g. a referral - what goes in a referral could be decided 1 second before the "send" button is hit.
- what can be standardised is components of EHR extracts / communications - which is why we concentrate on archetypes of such components (think "medications", "problem list", "physical exam", "biochem lab result", "claims attachment" etc), and use templates to join them together at runtime.
- the EHR Extract approach of CEN and openEHR is much closer to what is needed for EHR - EHR communication and sharing. So is the HL7 CDA + templates approach.
SOme HL7 people will undoubtedly disagree with this, and I am perfectly happy to see other points of view posted - a good discussion is an excellent way to expose the strengths and weaknesses of various approaches.
- thomas beale
