In 2013, a group of HIT companies started an intiative called the Commonwell Health Alliance. The primary participants were a compendium of well known HIT companies, namely, Cerner, McKesson, Allscripts, Athenahealth, Greenway Medical Technologies® and RelayHealth. The goal – to enable data liquidity through the use of open standard adoption by the member companies.
As you read this list of vendor companies you’ll be wondering where Epic is. An excellent question. According to Chillmark Research, Epic was not invited to join Commonwell in the early stages and after some fairly direct remarks from Judy Falkner they went on to form a competing initiative called CareQuality.
Well, this is all old history.
Except for the fact that it isn’t. Just recently at on ONC meeting, the president of Epic, Carl Dvorak, suggested the need for a nationwide record locator service, preferably run by the Centers for Medicare and Medicaid Services. The usual rivalry ensued.
For Epic to come out and call for interoperability is a sign of the times. I’m starting to feel that the “I” word is now in disrepute. It’s kind of like “the internet of things” – everyone’s saying it and nobody really knows what it means.
I’ve commented previously on Epic’s interoperability approach using CareEverywhere where it’s really not easy to send out of the “epic” network. Typical Epic. Does anybody remember Epic’s approach to community physicians? Let’s give them all free Epic licenses.
Ok. Enough Epic bashing. Does the Commonwell technical approach make sense?
Let’s think about the technology Commonwell intends to use. It’s the “Integrating the Healthcare Enteprise” standard all the way baby, with cross community (XCA) gateways, HL7 PIX/PDQ feeds, and all the missing pieces being filled in. Hang on, did I just say the they are using a enterprise (i.e. Hospital) standard for a national implementation?
Now, don’t get me wrong, I sat next to the author of the XCA standard, Chris Lindop, for about two years when I was at Kodak Health Imaging. He’s a very bright guy and he really knows his stuff. But, he was authoring these specs for the Enterprise, not for a national system. That’s why the security model is a bit lacking, and the patient identity scheme language is awfully vague.
The idea of running a national network of XCA gateways and PIX/PDQ feeds seems unwieldy. My reasoning here is that it’s all about granularity. Fine gained system have a gateway at each hospital, whereas coarse grained systems basically work with vendor gateways. Fine is better if you care more about local control over data. Coarser systems lose some of that control.
Commonwell is now also talking FHIR as an alternate protocol. Having read the ebxml standard used in the IHE profiles I’m clapping loudly here. FHIR is simpler and gets the job done just as well in this case. Pulling data from systems is FHIR’s forte today.
Ok, that’s enough standards geeking out. The real question is “is it possible that either CareQuality or Commonwell will actually achieve their mission?”.
Now, there’s a tough question. I think the answer is yes, but only partially. Having dealt with HIE’s for a while, the major problems often come down to getting sites to participate, dealing with state level regulation, and keeping data secure.
You’d think participation would be easy. It’s not. eHealth Technologies has about 100 major imaging sites under management and we’ve worked hard to connect every one of them. You often have to go through security and legal review and it’s tough to get things going, even with the assistance of the HIE. Why would it be easier when your vendor turns up? The site has nothing to gain, and a lot of money to lose. They need to get paid for minor issues such as infrastructure and network bandwidth. Who’s cutting the cheque here?
Turning to regulation, every state has additional provisions over and above the federal law. You need to comply with them. I wonder how much thought has been given to this problem? Consent can get tricky.
Then, there’s security. This registry is going to contain PHI and it’s going to have a big red target painted on it. Who’s taking responsibility for the liability? RelayHealth? Hopefully they’ll do better than Anthem.
My final point is that even if either of the initiatives works it may not do much to fix healthcare. Having more records in the EMR is not the goal. Yes, collecting data is a necessary step. Sending that data directly to the EMR is a necessary evil today. Aggregating the data, removing the tremendous amounts of redundancy in the record, and extracting useful clinical information in summaries is the end game.