TREAT Interoperability Framework: a shared FHIR and RabbitMQ backbone for federated digital health ecosystems
- Project
- 22022 TREAT
- Type
- New system
- Description
The TREAT Interoperability Framework connects independent digital health platforms through one shared standard instead of one shared platform. Each partner keeps the platform that fits its clinical, organisational and national context, and connects once to a common messaging backbone to exchange patient data with every other connected partner.
The framework combines a RabbitMQ message broker with HL7 FHIR as the common data format. Vital parameters, early warning scores and clinical notes are exchanged as FHIR Observations with LOINC codes and UCUM units, so every connected platform interprets a message with the same meaning. An API client library lets partners publish to and consume from the backbone from their own technology stack.
In TREAT, the framework replaced the single central Clinical Command Center originally foreseen. It was proven in the cross-country TREAT Complete demonstrator, with a live two-way exchange between Well@Home (Belgium) and My Viva Plan (Canada), and with partner platforms in Turkey and The Netherlands connected to the same patient journey.
The result can be reused by healthcare providers, regions and technology vendors that want several platforms to work together around one patient without building and maintaining point-to-point integrations.
- Contact
- Seppe Neyens
- seppe.neyens@bewellinnovations.com
- Research area(s)
- Health data interoperability; semantic interoperability; HL7 FHIR; message-oriented middleware; federated health ecosystems; remote patient monitoring; cross-border digital health
- Technical features
RabbitMQ message broker with a dedicated account, exchange and queue per partner; publish/subscribe without direct point-to-point links between platforms. Asynchronous, persisted messaging: messages are kept until processed, so a temporary outage of one platform causes no data loss and does not stop the others. HL7 FHIR R4 Observation resources with LOINC codes and UCUM units for vital parameters, early warning scores and clinical notes or summary reports. Shared patient identifier to match the same patient across platforms. API client library so partner systems can publish and consume from their own technology stack. Integrate once, connect to all: one connection to the backbone gives access to every connected partner. Two-way flows proven: measurements from a clinical telemonitoring platform to a lifestyle coaching platform, and AI coaching summary reports back into the clinical record. Repeatable test procedure for publishers and consumers that checks end-to-end transport and that code, value, unit, time and patient are the same in the sent and received data. Vendor-neutral: built on open standards; any partner or third party can host the backbone.
- Integration constraints
The main requirement is on the data side: a connecting platform must be able to read and/or write HL7 FHIR R4 Observations following the TREAT profile (LOINC codes, UCUM units) and map them to its own data model. The message broker itself is not a constraint: it is based on open-source RabbitMQ, so any organisation or group of organisations can set up and host its own backbone, with an account, exchange and queue per connected platform. The connected parties need to agree on a shared patient identifier to match the same patient across platforms. Use with real patient data requires appropriate consent, data processing agreements, and security and access controls under GDPR and other applicable regulations; in TREAT, the framework was demonstrated with synthetic patient data only.
- Targeted customer(s)
Hospitals and hospital networks, home-care and primary-care organisations, regional and national health authorities, remote patient monitoring and digital therapeutics providers, health insurers, and health software vendors that need several platforms to share data around one patient journey, within one country or across borders.
- Conditions for reuse
The framework is built on open standards (HL7 FHIR, LOINC, UCUM, AMQP/RabbitMQ) and can be reused and hosted by any party. The TREAT data profile and test procedure may be reused under the conditions of the TREAT consortium agreement. Partner platforms connected to the framework remain the property of their owners; reusing or integrating them is subject to each owner’s licensing terms. Any use involving patient data must comply with GDPR and other applicable data protection and healthcare regulations.
- Confidentiality
- Public
- Publication date
- 27-11-2026
- Involved partners
- My Viva Inc. (CAN)
- BeWell Innovations (BEL)
- Almende BV (NLD)
- Eindhoven University of Technology (NLD)
- LiveWell Giyilebilir Saglik Urun Hizmet ve Teknolojileri San. ve Tic. A.S. (TUR)