From HL7 v2 to FHIR without rewriting the HIS
A regional hospital network needed unified clinical access across hospitals, labs and imaging — without replacing a 12-year-old HIS. This is how we delivered it.
The starting point
The HIS had been in production for more than a decade. The lab ran on a separate system. Imaging sat on its own PACS, with its own protocols. Clinicians wanted unified access on mobile and desktop — without breaking anything that already worked.
What we didn't do
We didn't replace the HIS. We didn't swap the lab vendor. We didn't push every dataset into FHIR overnight. The cost and risk of rewriting a production HIS in a live hospital almost always exceed the short-term benefit.
The canonical clinical model
On one side: HL7 v2.x messaging — ADT (admissions, transfers, discharges), ORM (orders), ORU (results). On the other: FHIR R4 resources — Patient, Encounter, Observation, ImagingStudy, DiagnosticReport. Between them, a canonical clinical model. Transformation happens in one place, not scattered across twenty integrations.
IHE profiles where they earn their cost
Patient identification across systems via PIX/PDQ — one patient, one identity, no matter how many systems know them. Clinical document sharing via XDS — one registry, multiple repositories, explicit consent policy. Not every IHE profile is worth the effort; PIX/PDQ and XDS are.
DICOM routing with an audit trail
Between modalities (CT, MRI, X-ray) and PACS: Q/R SCU and C-STORE SCP, with per-transit logs. Every DICOM study passes through one auditable point. When a clinician asks "why is this image here and not there?", there is an answer.
The measurable outcome
Clinician response time measured in seconds, not minutes. Zero message loss when a source system reboots — every message is acknowledged, correlated, replayable. No lost clinical referrals.