Hospital Interoperability in 2026: FHIR, SNOMED CT and Ségur

The essentials in 30 seconds

Introduction

Every hospital runs dozens of software systems: the electronic health record (EHR), lab and imaging systems, the pharmacy, scheduling, billing. When these systems cannot speak to each other, clinical information is re-keyed, lost or trapped in silos. This is the problem that hospital interoperability exists to solve.

Interoperability is not a single feature you switch on. It is a stack of shared standards, HL7 FHIR for exchange, SNOMED CT and LOINC for meaning, and national frameworks like France's Ségur du numérique en santé, that let data move without losing its clinical sense.

The stakes are concrete. Poorly connected systems slow care, multiply data-entry errors and make secondary uses (quality reporting, research, AI) painfully expensive. Getting interoperability right is now a core responsibility of the hospital CIO (DSI) and the medical-information department (DIM).

Galeon builds a smart EHR designed for interoperability, exchanging data via the HL7 standard, developed with caregivers since 2016 and deployed across 19 hospitals including two university hospitals (CHU), covering more than 3 million patient records. The lesson from that experience is simple and worth stating plainly: hospital interoperability is far easier when data is structured correctly at the source.

What exactly is hospital interoperability, and why is it hard?

Hospital interoperability is the ability of different health information systems to exchange data and interpret it in the same way. It works on two levels that are often confused.

Technical (or syntactic) interoperability is about moving a message from system A to system B in a format both can parse. Semantic interoperability is harder: it means both systems attach the same meaning to the content, so "blood pressure" or "type 2 diabetes" is understood identically everywhere.

The difficulty is that hospitals accumulated software over decades, each with its own data model, vocabulary and interfaces. Connecting them one by one creates a fragile web of point-to-point connectors. Standards exist precisely to replace that web with a shared language.

What is HL7 FHIR and why has it become the default standard?

HL7 FHIR (Fast Healthcare Interoperability Resources) is a modern interoperability standard published by HL7 International. It breaks health data into standardised, reusable building blocks called "resources", a Patient, an Observation, a Medication, each exchanged over ordinary web APIs.

FHIR matters because it uses the same web technologies (REST, JSON) that power the rest of the internet, which makes it far cheaper to implement than older approaches. Its R4 release (2019) is widely adopted, and R5 followed in 2023.

FHIR does not replace everything overnight. Many hospitals still run older HL7 v2 interfaces, and the two coexist. But for new exchanges, FHIR has become the reference. In one sentence: FHIR standardises how data travels, not what it means.

What are SNOMED CT, LOINC and ICD, and how do they fit together?

If FHIR is the grammar, terminologies are the vocabulary. They give each clinical concept a stable, machine-readable code so that meaning survives the journey between systems.

SNOMED CT (Systematized Nomenclature of Medicine, Clinical Terms) is the most comprehensive clinical terminology, maintained by SNOMED International. It contains more than 350,000 active concepts covering diagnoses, procedures, findings and more. LOINC (Logical Observation Identifiers Names and Codes), maintained by the Regenstrief Institute, standardises laboratory tests and clinical observations, with close to 100,000 codes. ICD-10 (the WHO International Classification of Diseases, known in France as the CIM-10) is used mainly to code diagnoses for statistics and reimbursement.

These are complementary, not competing. A single lab result can carry a LOINC code for the test, a SNOMED CT code for the clinical finding and an ICD code for billing. Together they turn free text into structured, comparable data.

What is the Ségur du numérique en santé, and what does it require?

The Ségur du numérique en santé is France's national programme to accelerate and standardise digital health, launched in 2021 and steered by the Agence du Numérique en Santé (ANS). It has mobilised close to 2 billion euros to help providers upgrade to interoperable, secure software.

Rather than mandating a single product, the Ségur funds software that meets shared reference requirements. Central to it are the CI-SIS (the national interoperability framework maintained by the ANS), the INS (Identité Nationale de Santé, a unique national health identifier), and secure exchange services such as MSSanté.

The programme also feeds national services patients use directly. Mon Espace Santé, the digital health record opened for every person covered by French health insurance, is only as useful as the structured data hospitals send into it. The Ségur, in effect, makes interoperability a funding condition rather than an afterthought.

Why does structuring data at the source change everything?

Here is the core idea for any DSI or DIM: the cheapest place to structure data is where it is created. Coding a diagnosis or an observation once, at the point of care, is far less costly than reconstructing that meaning later from free text.

When data is captured as free text and coded retrospectively, hospitals pay repeatedly (in re-keying, in cleaning, in mapping projects) and quality still suffers. When data is structured and coded at the source, interoperability, quality reporting and research reuse become almost a by-product.

This is why the standards above are not just IT plumbing. They are the difference between data that flows and data that has to be rebuilt every time someone needs it. Structure once, reuse everywhere.

What does hospital interoperability mean for the CIO (DSI)?

For the DSI, hospital interoperability decides whether the information system is an asset or a liability. Standardised FHIR exchanges reduce the number of bespoke connectors to maintain and lower the cost of adding or replacing an application.

It also reframes procurement. A CIO can now ask, before signing, whether a solution natively supports FHIR, the relevant terminologies and the Ségur reference requirements, instead of discovering integration costs after go-live.

What does it change for the DIM and medical coding?

For the medical-information department (DIM), interoperability and coding are two sides of the same coin. When SNOMED CT, LOINC and ICD codes are applied close to the clinical act, coding quality improves and the burden of retrospective abstraction falls.

Better-structured data also strengthens activity reporting and opens the door to caregiver-validated datasets that are genuinely usable for research, without the months of cleaning that unstructured records usually demand.

Traditional interfacing versus a standards-first smart EHR

The contrast below is deliberately fair: traditional integration works, and many hospitals run it well. The point is where the effort lands, and how reusable the resulting data is.

Limits and challenges to be aware of

Interoperability is a direction of travel, not a switch. A few honest caveats matter for any realistic plan.

• Standards keep evolving. FHIR versions and national profiles change over time, so implementations need ongoing maintenance rather than a one-off project.
• Terminologies require governance. SNOMED CT is powerful but large; using it well means licensing, mapping choices and a team to maintain them. In France, national access is coordinated by the ANS.
• No standard removes human coordination. Aligning vocabularies and workflows between departments and institutions is an organisational effort as much as a technical one.
• Legacy data still needs migrating. Structuring new data at the source does not, on its own, clean up years of historical free text; migration and mapping remain real work.
• Benefits materialise progressively. Semantic interoperability pays off as coverage grows, so gains accumulate over months, not on day one.

FAQ

Is FHIR replacing HL7 v2?
Gradually, for new exchanges. HL7 v2 is still widely used and the two standards coexist in most hospitals today. FHIR is the reference for modern, API-based integrations.

Do hospitals in France have to use SNOMED CT?
Not universally for every use. SNOMED CT is increasingly central to structured clinical data, and France coordinates national access through the ANS. Its adoption grows as interoperability requirements mature.

What is the difference between the DMP and Mon Espace Santé?
Mon Espace Santé is the patient-facing national service that includes the shared medical record (DMP) along with messaging and other features. In practice, Mon Espace Santé is the space patients open; the DMP is the record inside it.

Does the Ségur du numérique make interoperability mandatory?
It does not impose one product, but it ties public funding and software certification to interoperability and security requirements, which makes standards a practical condition for eligible solutions.

What is the INS?
The Identité Nationale de Santé is a unique national health identifier that reliably links data to the right patient, a prerequisite for safe data exchange across systems.

How does a smart EHR help with interoperability?
By capturing structured, coded data as care happens and by supporting FHIR and reference terminologies natively, a smart EHR reduces the downstream cost of exchange, reporting and research.

In summary

Hospital interoperability rests on a clear stack: HL7 FHIR to move data through modern APIs, SNOMED CT, LOINC and ICD to preserve its clinical meaning, and, in France, the Ségur du numérique en santé and services like Mon Espace Santé and the INS to align the whole system. None of it delivers value if data is captured as free text and coded after the fact. The most reliable lever is to structure data correctly at the source, where it is created and validated by caregivers. That is the principle behind Galeon's smart EHR, built with caregivers since 2016 and now used in 19 hospitals covering more than 3 million patient records, where HL7 support and caregiver-validated structuring are designed in rather than bolted on. For a DSI or a DIM, interoperability is less a project than a property of how data is captured, and it is worth building in from the first keystroke.

Smart EHR vs traditional EHR: why hospitals are switching to AI in 2026

Sources

• HL7 International, FHIR specification
• SNOMED International, SNOMED CT
• Regenstrief Institute, LOINC
• World Health Organization, ICD
• Agence du Numérique en Santé, Ségur du numérique en santé and CI-SIS
• Mon Espace Santé, national digital health record
• Galeon, AI and health data: what every hospital must master in 2026
• Galeon, predictive medicine in hospitals: why AI depends first on data quality

‍