FHIR for Laboratory Integration
Saint-Petersburg government started project, whose goal is to create united integration bus for exchanging laboratory orders and results for state ambulatories, hospitals and laboratories.
A lot of objectives are expected: paperless orders, quick access to results for physicians and patients, resources planing & optimization, information for analytic & data mining by state healthcare organizations.
The first iteration was analyze current processes and create working prototype. This iteration was completed by creating custom SOAP API and integrate some organizations from focus group.
At the end of pilot project it became clear, that some international standard should be used as foundation. We've been invited to help make choice between standards (we have experience in integration, but in USA realm).
We have considered HL7 v2, HL7 v3 & FHIR and compared them with custom API.
At the first sight, HL7 v2 looks the best candidate. That’s may be true for United States, because there are a plenty of working systems (HIS, EHR, LIS) already exchanging HL7 v2 messages. But picture in Russia is different: most laboratory information systems are home-made and do not use HL7 v2. So key argument for HL7 v2 do not valid, but all birth injuries of v2 come into play.
HL7 v2 protocol was evolved ad-hock and we have archaic messages format, no built-in extensibility & machine-readable profiling/constraints and no future (standard in only-support/non-development stage).
Archaic messages format requires to implement awkward parsers and builders almost from scratch. Only Java & .Net developers could reuse HAPI and nHAPI libraries, but other popular platforms (node.js, erlang, php, python, ruby, etc) do not have convenient & robust libraries.
Absence of extensibility and profiling/constraints could be fixed using approach like LRI. But at least half of LRI is just fixes & patches for HL7 v2 bugs without bringing real value.
In summa: HL7 v2 requires a lot of meaningless work to deploy an obsolete standard ☹
HL7 v3 & CDA
Another candidate is HL7 v3 & CDA. There are two pro-arguments: some of HL7 v3 standards are now normative in Russia and HL7 working group released in 2013 DSTU for laboratory orders & results.
HL7 v3 has a solid theoretical foundation and designed for real semantic interoperability, but missed one important property — implementability. This could be illustrated by fact, that even until now there are no good implementation libraries. It also has heavy learning curve and complex for implementers. HL7 v3 is not one standard, but deep stack of standards, so even localization of this standard looks time consuming.
Project has deadline about half year and it does not look realistic use HL7 v3 & CDA to bootstrap it. All of this time will be spend on developers education ☹
Next option is FHIR. This is modern standard for fast interoperability. Standard is designed to find right balance between needs of semantic healthcare information exchange & implementation complexity.
It has built-in extensibility & profiling and good toolbox of libraries & services. FHIR allows implement only required part of standard, which could dramatically simplify start and makes possible incremental addition of new features.
We've compared custom API developed in first iteration with FHIR API and found almost one-to-one match. To be FHIR conformant current API should re-implement only few resources: Organization, Practitioner, Patient, Encounter, DiagnosticOrder, DiagnosticReport & Observation. Required business rules could be expressed as FHIR extensions. Laboratory, Hospital & Ambulatory systems could communicate with bus by REST api, using available reference implementation & client libraries. Public FHIR services could be used as sandboxes for development. HealthSamurai with support HL7 Russia already localized most of FHIR documentation.
Surely there are a lot of small technical questions, but we have to resume, that FHIR perfectly fits the needs & requirements of this project.
We hope, that managers and decision makers will do right choice with FHIR and will achieve their goals.