“Simplicity is drawn from the complexity of problems. Once you ultimately deal with the complexity, the simple solution is no longer simple.” – Wes Rishel
This week I’m attending the SOA in Healthcare Conference co-sponsored by HL7 and OMG. As you might guess by the title, the meeting is focused on Service Oriented Architecture (SOA) use within healthcare applications. One excellent keynote session today was presented by Wes Rishel, who is a long term HL7 community member and (I and many others would argue) one of the founders of HL7.
During his presentation he talked in detail about SOA and how it fits into the architecture of current and future applications. In my opinion he struck an excellent balance between the “what is possible in the future” vs. “what is really possible today and in the near term.” While making many great points, these are the ones that stuck most strongly with me:
- Wes opened with a funny episode of Greg the Architect called “SOA this, SOA that“. This was a mockumentary video created by TIBCO that shows the challenges with SOA becoming “blah blah” to management and customers. This video was amazingly well received given that it was shown to a room of SOA pundits.
- There are few “real, accessible” SOA implementations within commercial healthcare applications. “Real” SOA implementations in Wes’ view means that the vendor provides a user-accessible product you can buy today that is backed by a help desk, a published WSDL, public training classes, etc. That is, a SOA product is provided that can be really used by healthcare customers and the proposed approach is not just something “supported at the architectural level” or living “within the product road map”. Those applications make the “real world SOA” list:
- Cerner Millennium Objects – Access via SOA to Computer-based Patient Record (CPR) and Revenue Cycle Management (RCM) functions
- Eclipsys ObjectsPlus/XA – CPR only
- Epic – CPR only
- IDX/GE HealthObjects – RCM only
- Siemens Soarian – “Backdoor” special — meets most requirements for “real world” SOA but not all
- SOA Platforms are different than SOA Interfaces. Large, complex platforms are really expensive and only a few buyers can afford them. Wes quipped, “You can spend as much installing dbMotion as some facilities spend installing an entire CPR.”
- Wes spent a long time discussing how SOA technologies fit into the Gartner Hype Cycle. The good news is that SOA is in/around the “Trough of Disillusionment” so sometime soon we’ll have something useful with SOA. Wes pointed out that the hype cycle can be used for many areas beyond technology and I think he is right. Wes wisecracked, “I had one guy tell me that, ‘The hype cycle is the best description of my marriage.'”
- Wes also discussed the broad challenge of “control” within the healthcare space. He referenced generic web mash-ups and the fact that they only really work when interface is controlled by a dominate party. That is, a vendor comes to the table and proposes an interface and the market follows. As an analogy, “Google and Amazon can do for generic, non-healthcare SOA what Wal-Mart did for X12 and supplier relationships — they can say ‘this is how we do it.'”
- The issues with healthcare are not simple. “Simplicity is drawn from the complexity of problems. Once you ultimately deal with the complexity, the simple solution is no longer simple.”
- Wes addressed the issue of legacy integration v. future approaches. He summed up with, “All new interfaces in green field areas should use web services.” I agree with the concept — if a healthcare data area is not currently integrated, do so via web services.
- Wes summed up with two critical points:
- “On the ground, connecting existing systems is a challenge due to differing levels of informatics sophistication.” Meaning that healthcare applications have huge differences in what level of data model and interaction they allow.
- This concept was built upon with this definition: “Frozen interface syndrome: Existing HL7 message interfaces are pretty much stuck at 2.2.” Meaning that application vendors don’t like to change for change’s sake — there needs to be a business motivation.
“Frozen interface syndrome: Existing HL7 message interfaces are pretty much stuck at 2.2.”- Wes Rishel
Latest posts by Dave Shaver (see all)
- HL7 ADT Q&A with Dave Shaver - July 2, 2014
- Health Standards Community Membership Archetypes: Who uses HL7? - August 6, 2013
- Note from the Field: Meditech 6.0 HL7 Integration - September 6, 2011