Recently, I received a question regarding the HL7 ADT message and thought that it, along with the answer, warranted a larger audience.
Question: Is there a national standard to what the baseline structure of an ADT (A28, A31) message should include? Are there national specifications/schematics that say what the HL7 baseline skeleton message should look like/contain?
Answer:
At its core, HL7 is a data model and an encoding standard for representing that model.
The HL7 V2 standard has been (re)defining the broad attributes of a shared, collaboratively-built, global, and “generic” model since the late 1980s. Much of the V2 model is “implied” in the message encoding—a bit messy but functional.
HL7 V3 – especially the RIM and related artifacts – is much more of a formal data model. The CDA standard is the basis of CCD, which is mandated by Meaningful Use. CDA is a core part of HL7 V3.
The latest work by HL7 is FHIR. It, too, is mostly-backed by the formal model from HL7 V3, but FHIR is also deeply focused on ease of implementation. The technology stack is “modern” (REST + JSON/XML), but that is not as critical, IMO, as the standardization of an easy-to-use data model.
FHIR is at the DSTU stage and is being “rapidly implemented” by a number of projects, yet the number of commercial products that will support it in the next two years is unclear. To be clear, we have a long way to go before 1) FHIR is “done”, and 2) the normative edition of FHIR is released (in about two years).
Read: Healthcare Informatics named FHIR as one of the top ten trends for 2014
In summary, ultimately there is not any magic technology or approach—two independent systems that need to exchange data need to agree on a model for exchange. When they almost agree an interface engine is used to make adjustments. (See my recent post on comparing interface engines.)
I recently answered a similar question in the HL7 group on LinkedIn (you may need to request to join).
