EDIFACT messaging in healthcare

EDIFACT predates FHIR by two decades, and the infrastructure built on it is still in production across Europe. The pattern repeats country by country - a national or regional mailbox network connects general practitioners, laboratories, hospitals, and pharmacies, and the messages flowing through it are EDIFACT profiles defined by a national standards body.

These networks are store-and-forward - a lab drops a result into the recipient's mailbox and the network guarantees delivery. There is no live acknowledgment loop as with HL7 v2 over MLLP, which simplifies the transport side of EDIFACT integration, as transport patterns explains.

One call - parse_edifact - handles the UNA service string advice, the UNB-UNZ envelope, and every message inside the interchange. Each message resolves to a typed Python class from its UNH identity, and message types without a registered class parse into generic, fully navigable objects instead of failing.

Dialects are plain Python. The platform ships the Dutch healthcare profiles as a complete example - typed segments, composites for names, addresses and dates, and the message structures - and your own dialects follow the same pattern, whichever country or sector they come from.

The service below receives a laboratory result and walks its repeating groups - each specimen with its determinations and their values. The message in this example is a Dutch MEDLAB, but the shape of the code is the same for any dialect - parse, navigate by name, act on the values.

# -*- coding: utf-8 -*-

# Zato
from zato.server.service import Service
from zato.edifact import parse_edifact

# The example dialect - import your own instead
import zato.edifact.nl

class LabResults(Service):
    name = 'demo.edifact.lab-results'

    def handle(self) -> 'None':

        interchange = parse_edifact(self.request.raw_request)
        msg = interchange.message

        # The patient and the sending department
        patient = msg.pid.patient_name.married_name
        department = msg.department.department

        # Each specimen and its determinations
        for material in msg.materials:
            for bep in material.determinations:
                self.logger.info('%s: %s = %s %s',
                    patient,
                    bep.determination,
                    bep.result,
                    bep.unit)

From here the values go wherever the integration needs them - a database, a REST API, a FHIR server following the same bridging pattern as the HL7 v2 to FHIR mapping, or a JSON representation via to_json for any downstream consumer.