FHIR to help drive down the cost of interoperability
The first Australian FHIR Connectathon will be held in conjunction with the International HL7 Interoperability Conference (IHIC 2013) later this month, offering Australian-based integration specialists the opportunity to see the new standard up close.
Representatives from NEHTA and the PCEHR national infrastructure partner Accenture are planning to be involved in the event as part of their investigations in to how FHIR will fit into future PCEHR functionality.
FHIR – or fast healthcare interoperable resources – is a framework standard devised by healthcare standards specialist Grahame Grieve that promises to help create interoperability solutions at a much reduced cost than current standards such as HL7 v3.
Mr Grieve told the recent eHealth Interoperability Conference in Sydney that FHIR had come out of intense frustration with the development of HL7 v3, which while it captures a huge amount of knowledge about how healthcare works around the world, is too detailed and difficult for implementers and vendors to understand.
There are good things about it, Mr Grieve said, “but the trouble was you have to spend months consuming the standards before you understand them. Vendors just don't have the money for it.”
Instead, FHIR has been developed to make standards and implementation as simple as possible, taking the best from other applications and allowing vendors to make interoperability cheaper and easier.
The spark for it came from a company called 37Signals which develops cloud-based applications for integration.
“They live or die by their interoperability,” Mr Grieve said. “They have a particularly highly regarded API called Highrise, which is a customer relationship manager. So what I did was I took a copy of their spec and made as few changes as I could to make it work for healthcare.”
FHIR's resources are tailored to be used with a RESTful interface, a design model for developing web APIs.
“It doesn't matter if you don't know what a RESTful interface is, but it's the way Facebook, Google and so forth work,” he said. “It is tailored to work that way but it can also be used in the classic ways that we use for v2 messaging or documents.
“FHIR is designed for implementers. Every resource has a section for standard data elements and a section for narrative so that a human can read it even if the system doesn't have a clue what it is.
“It is written to be understood and implemented and we test that constantly by connectathons. As much as we possibly can we describe the content to be exchanged in the natural language of the people who use the resource.
“We've stolen the best ideas we could from HL7 and DICOM and IHE as well as from the web – it's really a mash up of other people's ideas. We don't do anything new – we steal other people's ideas. Our mantra is simplicity, implementability, does it work?”
FHIR was first conceived in 2011 and already has a draft standard for trial use (DTSU) open for comment. Mr Grieve expects the DTSU to be finalised early next year, which he described as “an awesomely rapid path”.
What role FHIR will play in the PCEHR is as yet unclear, as it was designed on an HL7 v3 offshoot, clinical document architecture (CDA). This standard has been used to develop the templates for clinical documents in the PCEHR such as shared health summaries and event summaries.
While Mr Grieve described CDA as reasonably flexible – and he is the CDA expert for the PCEHR – he said it was still too hard for average developers to implement. With FHIR, the goal is that an average developer or systems integrator can read the FHIR specs and understand them in half an hour.
Unfortunately, the development of FHIR came a couple of years too late for the PCEHR, and at the time there was no other option, he said.
“(CDA) doesn't work overly well in that documents are not data,” he said. “I wouldn't want you to go away and say CDA was a bad choice – CDA was the only choice, given the program’s stated goals – but there are problems with narrative versus data.
“The bottom line is that development of this is too hard. CDA is too simple for requirements but too hard for implementers. That is a terrible combination and in the PCEHR, it is playing out. Again, it's not that it was a bad choice, it was the only choice.”
FHIR is a free resource with no limitation on its use or distribution, and all tutorials and documentation are published under open licences. Mr Grieve said this, combined with its ease of use, should enable widespread use and therefore drive down the cost of making vendor products interoperable.
“These factors – easier to develop, easier to troubleshoot – mean you can find far more people to do the work, which will drive down the cost of integration and interoperability. You won't need experts to get anything done.
“At the moment you have v2 and CDA and you need someone really good at that, and you need someone to mediate between the two and it's really expensive. The big thing about FHIR is that it has the same content models that support messaging or documents or services, but the crown jewel is the RESTful API.
“Overall, the cost of integration will go down and the cost of integration prevents better healthcare. Fixing integration won't fix healthcare, but it will stop getting in the way of people who want to fix healthcare.”
While it is too late to rebuild the PCEHR using FHIR, there are possibilities in the future, he said.
"FHIR may be relevant and both NEHTA and Accenture are looking at it. That doesn't mean they are going to use it, they are just evaluating it, and that means connectathons.””
The 14th International HL7 Interoperability Conference will be held in Sydney on October 28 and 29, with the FHIR Connectathon taking place on Sunday, October 27.
Posted in Australian eHealth