Who is the Audience for HL7?

When I implemented my first HL7 interface, I printed out about 50 pages from the Version 2 specification, read it, and wrote, in about 4 hours, a working (but not fully functional) ADT feed for my application. That was six years ago, and I had been involved with HL7 for about 2 months. If I hadn’t needed it quickly, I could have handed the task off to the most junior of engineers on my team, and gotten something back in a few weeks that met the very minimal requirements that I had.


I’m very glad that I at that time I didn’t need a Version 3 interface, because there is no way it could have been completed in the same time frame. There are a number of reasons why it takes longer. Here are a few of the bad ones (there are also some good ones, but that isn’t the focus of this posting):

1. HL7 doesn’t provide any WSDL with its V3 specifications, but that’s the first thing my engineers ask for.
2. HL7 schemas often need manual tweaking for some of the more complicated models, but there are no instructions anywhere on what those changes are or why they are needed.
3. The XML has needless variation across different domain models.
4. What HL7 uses for modeling is only used in and by HL7.
5. Examples are notably absent from many of the specifications.

All of these are barriers to implementation of HL7 Version 3. Many of them completely stymie the engineers I’ve worked with across the healthcare landscape. I get to work with some pretty bright people. Many of them are the more senior staff associated with numerous different implementers of healthcare IT. If they cannot figure it out, then the problem is not with them, the problem is with us.

 

In order for HL7 Version 3 to be not just the best, but also the most widely used standard in healthcare we need to consider who our audiences are. Ten years ago that audience included junior and senior engineers as well as application architects. These days HL7 Version 3 seems to mean something only to application architects. We need to figure out how to enable the other consumers our work products, or we’ll never get there.