FHIR projects in the NHS
Presented by
Richard Kavanagh
Head of Data Standards – Interoperability Specifications
HL7 in NHS within England
NPfIT introduced HL7v3 at scale in to the architecture
of the NHS.
National applications deployed at scale
Implementation demonstrably possible, though at a
The “challenge” of HL7 V3
Creating HL7v3 specifications uses niche skills not readily
Implementing HL7v3 specifications uses niche skills not
readily available.
Expensive to specify, expensive to build.
“Cost” and “Benefit” are they proportional?
Semantic rigour often not used or even required due to the
“templating” processes
Set the data free
Access to data via APIs is a strong design
requirement for future systems.
#Code4Health desire to make access to data simple
to promote innovation
The age of the “App” is upon us. Smaller devices
requiring simpler technology stacks
Web Service vs. REST
Technology Refresh
Choose and Book to be redeveloped and transformed
to eReferrals Service.
Supplier consultation (via Intellect)
• not positive about using HL7v3
• JSON and REST technology skills commonly available in the
“IT Development” workforce.
• APIs are prevalent ( lists
s10,394 APIs)
• HL7v3 discounted, supplier engaged to build API
The “FHIR” opportunity
What we need is some thing that supports
Web Services and REST
Allows fine grained APIs
Is more accessible to innovators
Is extensible
Getting buy in
Initial scepticism – some antibodies to HL7
Reviewed by supplier – positive feedback
Internal HSCIC governance
• Interoperability Design Authority
• Architecture Governance Group
FHIR now an accepted architectural solution for future
use by HSCIC
Setting Expectations
FHIR is a “work in progress” showing lots of promise
but not yet a “standard”
The painful “bleeding” edge
The alternatives to FHIR, bespoke solution or early
The development has started
Stability has yet to materialise
Managing the risk of going first
We will accept deviation
We will be pragmatic
• New resources
We will report back
• JSON serialisation
• QUERY resource
It will not be 100% FHIR
Development through collaboration
No proven process for profiling FHIR
FHIR is online and freely available – don’t reinvent the
FHIR is a standard not an implementation
Evolve the process to meet demand
Not another “MIM”
What to expect from eReferrals
The delivery will be a mix of existing HL7v3 messages
and FHIR
The APIs will be released as per the eReferrals
delivery schedule
The resources will be for a mixture of clinical and non
clinical concepts.
What we expect so far
Custom resources will be published
profiles for all resources will be published
Examples in XML and JSON for resources
FHIR System namespaces
FHIR valueSets
Validation rules and generated data
Other projects underway
Two projects in discussion, too early to publish.
Training underway in the central team
Sharing the learning and early feedback
FHIR is not the only way
GP2GP changes being discussed using HL7v3
ETP changes being discussed using HL7v3
Public Health England messages using CDA
Feedback is welcome
[email protected]

similar documents