Will FHIR Solve Current Clinical Data Quality Problems?
Adapted from an Integrated Healthcare Executive Interview with John D’Amore, Diameter Health’s President and Chief Strategy Officer (July 22, 2019)
Integrated Healthcare Executive: Can you explain why the FHIR guidelines do not address unstructured data?
John D’Amore: There are two things to talk about with FHIR. One is the way that APIs have unlocked tremendous information flows on the World Wide Web and
the Internet. FHIR is going to unlock even more data and enable it to flow through the healthcare ecosystem. It’s clearly a better and faster way to get information. However, FHIR doesn’t necessarily clean up the information that flows.
When you’re using APIs with technology companies like Google you always get data back in a completely consistent format. No matter where you are in the world, you can always find your longitude and latitude using a Google API service. EHRs are going to be able to respond in the same way using the FHIR protocols.
However, the way that you get that information may still be very divergent. I’ll give you one example. When you look at a medication list, you see medications that are encoded using the National Library of Medicines’ RxNorm vocabulary. You can also use the FDA’s method to refer to that same medication which is called the National Drug Classification.
You can also use proprietary libraries, like Medi-Span, or First Databank, or Multum to record that information. There are many different ways that that information could be codified in a medication list as well as also having the opportunity for a clinician to write down that the patient is taking baby aspirin (for example) without using any codes at all.
When you get that information back as part of an API request from FHIR, it’s not necessarily going to turn that information into a usable information asset. FHIR’s going to make it available but won’t necessarily make it fully usable.
The analogy with languages is that FHIR will make it so that you can hear all the people around you and what they’re saying. It’s not going to be a Star Trek universal translator where everyone speaks the same language.
There’s still a need for technology companies to provide that middle layer service of cleaning up and making the clinical information semantically interoperable for FHIR to achieve its true potential.