by Eliot Muir

June 2011 – HL7 was originally a grassroots movement. Vendors and providers got together and looked at the typical content of the real world interfaces they were implementing and put together a rough, yet relatively practical, standard which we know today as HL7 Version 2.X.  Implementers played a big part of that early HL7 organization.

HL7 has always been agnostic as to how the data was sent over the wire. The irony is that the standardization it brought was actually the source of most of its value in the early days. HL7 increased standardization by allowing the exchange of data via TCP/IP using the simple “LLP protocol” with a standard EDI format.

Suddenly far less money had to be spent in order to get two different vendor applications to talk to each other.

The HL7 standard has also always been a fairly loose standard, with a lot of options and ambiguity of where data should go.  But it was better than anything else in its time.

Culturally now, however, the HL7 organization is in a bit of a funk.  Government funded organizations are starting to excessively dominate the agenda and the standards dialog has become increasingly isolated from broader trends in technology outside of health care.

The organization got caught up in a time warp of what I would describe as the “XML gold rush” of the last decade. Vendors and standards committees were outdoing themselves to be the king of data plumbing in the enterprise, tossing out new complicated XML technologies faster than anyone could learn them.  Vendors like Microsoft, IBM, Sun and all the usual suspects were doing their best to dominate the brave new XML world.

Fundamentally it didn’t work.

The new XML standards degenerated into becoming too complex for their own good. The new XML protocols like SOAP became so convoluted that, unless you used the same implementation or limited yourself to a tiny subset of features, it made interoperability harder rather than easier.

I remember at the time sitting off to one side and thinking, “I don’t love where this is going. Hmmmm – maybe it’s time to give up on technology and go become a math teacher.  I have always had better trust of technologies I could actually understand myself.”

Fortunately rationality prevails.  The market gravitates towards simplicity and efficiency. We’re seeing the gradual rejection of the excesses of that decade.  The big vendors have moved on to bigger things.  We don’t see Steve Jobs and Larry Page standing up at keynote presentations getting all excited about the latest XML standards.  The center of gravity in the computer landscape has shifted. From a revenue perspective, enterprise integration is a much smaller part of the overall technology market.

Application integration is shifting to where it should be, which is just a means to an end. Organizations and developers are voting with their feet and gravitating to much simpler integration solutions – namely RESTful web services.  It promises a new era of simplicity.

The problem is that HL7 still has to wake up to this fact.  The organization is still caught up in the aftermath of XML fever of the last decade.  The Version 3.0 HL7 (V3) standard uses complex protocols with very difficult to comprehend XML.

All this would be harmless enough if the market could be left alone to choose whether or not to use version 3.0.  The answer would be a resounding NO.  There is no market in the world where version 3.0 would get any type of traction if industry were given the choice whether to use it.

But the problem is that at this stage the HL7 organization has far too much influence with governments, which are attempting to impose top down standards on the industry.  This is really hurting the progression of interoperability within healthcare. Vendors are now being forced in multiple jurisdictions to leap through too many hoops in order to become ‘’certified’’ as compatible with these complex standards.

This is bad news.

The healthcare industry already suffers from excessive centralization of decision making. Overall product quality and interoperability always suffers. Resources that could be going into improving product quality are instead going into meeting various conformance standards. If there is not an active push back by the healthcare IT industry or some rationality within the HL7 community to re-examine and try to simplify these standards, then we can look forward to yet another decade of disconnected systems in healthcare.  The new HL7 standards should be forced to compete in their own right with competing solutions. A more creative, market driven approach is needed. For example:

1. Every vendor must provide some type of open integration layer that gives access to 100% of the data within their system.
2. These interfaces must be publicly documented.
3. These interfaces must be part of the standard package or at least turned on a published price.

Governments and regulating bodies should not be in the business of specifying what these interfaces should be – that should be left to natural market processes to resolve what really works in practice.

About the Author

Eliot Muir is the CEO of iNTERFACEWARE, the company that makes the healthcare integration engine Iguana.  Iguana is used both by vendors and healthcare providers.  iNTERFACEWARE’s direct customers include MD Anderson, Mayo Clinic, GE Medical, McKesson, Phillips and over 300 other leading healthcare organizations.  For more information please see http://www.interfaceware.com