A clinician seeing a patient for the first time in a new setting faces a familiar problem: the records exist, but they're somewhere else. A discharge summary sits in a hospital system that doesn't connect to the GP's medical record system. A referral arrives as a scanned PDF rather than structured data. Blood results from last month have to be re-requested because the laboratory system can't push data to the ward's clinical platform. These aren't edge cases. They're daily realities for clinicians across Europe, and they carry real consequences for patient safety, care quality, and the working lives of the people delivering that care. Poor healthcare interoperability isn't primarily an IT problem. It's a structural challenge embedded in the way European health systems have been built, procured, and governed over decades.

What healthcare interoperability actually means

Interoperability, in the context of clinical systems, refers to the ability of different medical record systems, diagnostic platforms, and digital tools to exchange, interpret, and act on shared data. The word is used loosely, which is part of the problem. A scoping review published in SAGE Journals identifies three distinct levels that matter in practice:

• Foundational interoperability: data can be transmitted between systems — the basic plumbing of connectivity

• Structural interoperability: data is formatted consistently enough that a receiving system can parse it

• Semantic interoperability: data carries shared meaning, so that a blood pressure reading in one system means exactly the same thing in another

A fourth dimension, organisational interoperability, concerns whether the governance, workflows, and institutional trust arrangements exist to enable data to flow even when the technical conditions are met. Research published in JMIR Medical Informatics confirms that operational health data interoperability requires not only technical standards but also enforceable governance arrangements. Systems can be technically capable of exchanging data and still fail to do so because no organisation has agreed to allow it.

This distinction matters when diagnosing why clinical systems fail to connect. A hospital and a GP practice may both use Fast Healthcare Interoperability Resources (FHIR)-compliant software and still not share data if the governance layer, covering agreements about consent, liability, and access, has not been established.

The four layers where interoperability breaks down

Each of the four interoperability levels represents a distinct failure mode in clinical settings.

At the foundational level, systems simply can't reach each other. A community mental health team and an acute hospital trust may run entirely separate network environments, with no technical pathway for data exchange at all.

At the structural level, data formats diverge. A referral letter generated in one medical record system may arrive in another as an unstructured PDF, stripping out all the coded clinical data that the sending system captured. The receiving clinician gets text, not data.

At the semantic level, the same clinical concept is coded differently across systems. Research on laboratory medicine under the European Health Data Space (EHDS) highlights how semantic coding inconsistencies, between systems using Logical Observation Identifiers Names and Codes (LOINC), Systematised Nomenclature of Medicine Clinical Terms (SNOMED CT), or local codes, make it impossible to reliably compare results across platforms, even when the underlying data has been successfully transmitted.

At the organisational level, institutions haven't agreed on who can access what, under what conditions, and with what liability. A qualitative study on affinity domains in digital health interoperability found that cross-enterprise document sharing requires structured organisational-technical models, not just technical readiness, and that these governance arrangements are frequently absent or inconsistently applied.

Why European healthcare has a particularly complex interoperability problem

Europe's interoperability challenge is compounded by structural features that don't exist in the same combination elsewhere. Health system organisation is a national competency across EU member states, meaning that medical record system procurement, data governance, and interoperability policy have historically been decided at country level, or lower. The result is a continent-wide patchwork of incompatible systems.

A Frontiers in Medicine analysis using Italy as a case study illustrates how even within a single member state, national health data ecosystems are fragmented across regional administrations, each with its own infrastructure, standards adoption, and governance arrangements. Italy's experience isn't unusual. It reflects the structural reality across much of Europe.

Several factors make this particularly difficult to resolve:

• Legacy system prevalence: Many hospitals and primary care networks run clinical platforms procured in the 1990s or early 2000s, built before modern interoperability standards existed

• Vendor lock-in: Analysis published in PLOS Digital Health documents how medical record system vendors have historically maintained proprietary data formats and closed application programming interfaces (APIs) as a commercial strategy, making it costly for healthcare organisations to switch systems or integrate with external platforms

• Mixed public and private care: In countries with significant private care sectors alongside public healthcare infrastructure, data sharing arrangements between the two are often absent or legally ambiguous

• Cross-border complexity: Patients who move between EU member states, or who receive care across borders, encounter systems that were never designed to communicate with each other

The role of standards: HL7, FHIR, SNOMED CT, and ICD

The technical standards community has produced frameworks intended to address these problems. Understanding what they do, and what they can't do alone, is essential for any healthcare leader evaluating system procurement or integration.

Health Level 7 (HL7) is a set of messaging standards that defines how clinical information is structured and transmitted between systems. It has been in use for decades and underpins much of the existing clinical data exchange infrastructure in European hospitals. However, older HL7 v2 implementations are highly variable in practice, with local customisations that undermine interoperability.

FHIR, developed by HL7, is the current-generation standard. It uses modern web APIs to enable more flexible, granular data exchange. Research demonstrating FHIR-based interoperable medication records shows that the absence of interoperable medication data within and across hospitals creates administrative burden through redundant data entry and increases the risk of errors. FHIR is widely regarded as the most viable path to scalable interoperability, but a JMIR study on EHDS standards notes that the choice between HL7 FHIR and other frameworks such as Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) remains contested for different use cases, and that harmonising national data governance approaches adds further complexity.

SNOMED CT and the International Classification of Diseases, 10th and 11th revisions (ICD-10/11), provide the semantic layer, offering shared clinical terminology that allows a diagnosis coded in one system to be understood by another. Work on transforming legacy HL7 Clinical Document Architecture (CDA) documents to FHIR illustrates that semantic interoperability between older and newer systems requires active translation work, not just the existence of a shared standard.

The critical limitation is that having a standard doesn't guarantee adoption, and partial implementation is the norm rather than the exception. A system can claim FHIR compliance while implementing only a subset of the specification, creating interoperability in theory and fragmentation in practice. Better.care's analysis of EHDS infrastructure captures this gap directly, noting the difference between FHIR readiness in vendor presentations and FHIR readiness in production environments.

The European Health Data Space: what it is and what it changes

The European Health Data Space, established under EU Regulation 2025/327, represents the most significant legislative intervention in European health data governance to date. It entered into force in March 2025 and sets out a framework for both the primary use of health data (patient care) and secondary use (research, policy, and innovation).

For healthcare providers and medical record system vendors, the EHDS introduces several concrete obligations:

• Medical record system manufacturers must achieve certification demonstrating compliance with harmonised interoperability standards

• Systems that don't meet interoperability requirements won't be permitted to be marketed in Europe

• Member states must establish national contact points for health data to facilitate cross-border exchange via the MyHealth@EU infrastructure

• The European Electronic Health Record Exchange Format (EEHRxF) will define the mandatory data structures for cross-border patient summaries, ePrescriptions, discharge summaries, laboratory results, and medical images

A Frontiers in Medicine policy analysis describes the EEHRxF Standards Hub as a coordination mechanism for harmonising implementation across member states. The implementation timeline runs from 2027 to 2031, with different categories of health data becoming subject to mandatory exchange at different points.

The European Commission projects €11 billion in savings over the next decade from improved data sharing. However, legal analysis from Skadden notes that healthcare providers and digital health companies face significant compliance steps, including certification processes and mandatory interoperability testing, that many organisations haven't yet begun to prepare for.

The gap between policy ambition and clinical reality shouldn't be underestimated. Research on Italy's national health data ecosystem shows that even member states with relatively mature digital health infrastructure face substantial barriers to EHDS readiness, barriers that are technical, organisational, and political simultaneously. The EHDS creates the conditions for interoperability. It doesn't deliver it automatically.

How poor interoperability affects clinicians day to day

The human cost of fragmented clinical systems is measurable and well-documented. For clinicians, the most immediate consequence is increased documentation burden: data that exists in one system must be manually re-entered into another, consuming time that could otherwise go to direct patient care.

Research on medication data fragmentation found that the absence of interoperable medication records creates administrative burden through redundant data entry and increases the risk of errors through imprecise data transformations. This is a daily feature of clinical work in most European healthcare settings, not an abstract risk.

Beyond time, the cognitive load on clinicians is significant. During a ward round or outpatient consultation, a clinician who can't access a complete, up-to-date patient history must either work with incomplete information or spend time chasing records across systems. A nine-country survey of 417 European healthcare IT leaders found that 78 per cent cited interoperability as a top-three procurement driver, a figure that reflects how widely the operational pain is felt at leadership level.

The connection between documentation burden and clinician burnout is increasingly recognised. When administrative tasks consume a disproportionate share of clinical time, the result isn't just inefficiency. It's a structural contributor to the exhaustion and disengagement that characterise burnout in healthcare settings.

How poor interoperability affects patients

From the patient's perspective, fragmented systems translate into concrete harms and inconveniences that are often invisible to the individuals experiencing them.

Repeated diagnostic testing is among the most common consequences. When a GP can't access hospital laboratory results, or when a specialist can't retrieve imaging from a previous admission, tests are ordered again, adding cost, delay, and in some cases unnecessary radiation exposure. The European healthcare IT survey found that 55 per cent of organisations reported reduced duplicate testing following improvements in data sharing, which implies that duplicate testing remains prevalent where those improvements haven't been made.

Care transitions are a particularly high-risk point. When a patient moves from secondary care back to primary care, the discharge summary is a critical document. If it arrives days later in an unstructured format that can't be read by the GP's medical record system, the continuity of care is broken. Medication changes, follow-up requirements, and new diagnoses may not reach the clinician responsible for ongoing care in time to act on them.

Cross-border care exposes the most acute version of this problem. Research on telemedicine and the EHDS identifies how varying technical standards, data formats, and coding systems hinder cross-border medical record system communication, creating care gaps for patients who travel, live across borders, or seek specialist treatment in another EU member state.

Where AI fits in: can clinical AI bridge the gap?

Artificial intelligence tools, including AI medical assistants, ambient voice technology (AVT), and real-time transcription systems, are increasingly being deployed in clinical settings. Their relationship to interoperability is frequently misunderstood.

AI tools don't fix broken infrastructure. A system that can't exchange structured data between two medical record systems won't be repaired by adding an AI layer on top. However, AI can reduce the burden that fragmentation places on clinicians in specific, meaningful ways.

AVT and real-time transcription can capture clinical encounters and generate structured notes without requiring the clinician to manually enter data into multiple systems. This addresses documentation burden at the point of care, even when the underlying systems remain disconnected. When an AI medical assistant can extract relevant information from unstructured text, such as a scanned letter or a free-text note, and surface it in a usable form, it partially compensates for the absence of structured data exchange.

Research on EU AI Act compliance within the MyHealth@EU framework notes that AI-based clinical decision support systems are automatically classified as high risk under the EU AI Act, and that integration with cross-border health data infrastructure introduces additional regulatory complexity. AI tools deployed in interoperability-sensitive contexts carry compliance obligations that go beyond standard software procurement.

The honest limitation is that AI tools operating on fragmented, incomplete, or inconsistently coded data will produce outputs that reflect those limitations. AI can reduce friction and surface information more efficiently, but it can't substitute for the semantic consistency and structural completeness that genuine interoperability provides.

What good interoperability looks like in practice

It's worth being specific about what a well-connected clinical environment actually delivers, because the gap between current reality and the achievable benchmark is often larger than it needs to be.

In a genuinely interoperable clinical setting:

• A GP receiving a referral from a specialist receives structured, coded data that auto-populates the relevant fields in their medical record system, not a PDF that must be read, interpreted, and manually re-entered

• A discharge summary generated at the end of a hospital admission is transmitted to the patient's GP in a format that updates the patient record automatically, including medication changes, new diagnoses, and follow-up requirements

• An Advice and Guidance (A&G) workflow between a GP and a secondary care specialist shares the relevant patient context, including current medications, recent results, and active diagnoses, without the GP having to retype it into a messaging system

• A clinician seeing a patient who has recently moved from another region or country can access a patient summary that their system can read, using shared terminology that requires no manual translation

The EHDS Regulation sets patient summaries, ePrescriptions, discharge summaries, laboratory results, and medical images as the priority data categories for mandatory cross-border exchange, a list that maps directly onto the most clinically consequential failure points in current systems.

Key barriers still standing in the way

Despite regulatory momentum and growing clinical demand, substantial obstacles remain.

Vendor lock-in and commercial incentives are among the most persistent. PLOS Digital Health analysis documents how medical record system vendors have historically maintained closed systems as a competitive strategy. The EHDS attempts to address this through mandatory interoperability requirements and pre-market testing, but enforcement mechanisms and the pace of vendor compliance are yet to be tested at scale.

General Data Protection Regulation (GDPR) and data residency requirements create genuine complexity for cross-border data flows. While GDPR isn't inherently incompatible with health data sharing, the interaction between data protection obligations, national health data legislation, and EHDS requirements creates a compliance landscape that many organisations find difficult to navigate. Legal analysis of EHDS obligations identifies certification requirements and mandatory interoperability testing as areas where healthcare providers and digital health companies need early legal and technical preparation.

Inconsistent FHIR adoption means that the existence of a standard doesn't translate to uniform implementation. Better.care's infrastructure analysis notes that Europe's healthcare landscape is fragmented across hundreds of medical record system vendors with varying architectures, and that vendor compliance claims frequently outrun production reality.

Legacy system replacement is underfunded across most European health systems. Replacing a hospital medical record system is a multi-year, multi-million euro programme that most healthcare organisations undertake infrequently. In the interim, legacy systems must be bridged to modern interoperability infrastructure, a technically complex and resource-intensive task.

Organisational alignment remains perhaps the hardest problem. Even where technical standards exist and systems are capable, institutions must agree on governance, data access policies, and liability arrangements before data can flow. Research on affinity domains in cross-enterprise document sharing confirms that enforceable governance arrangements are a prerequisite for operational interoperability, and that these arrangements are frequently the missing element.

What clinicians and healthcare leaders can do now

Healthcare decision-makers aren't passive observers of the interoperability problem. There are concrete steps that clinical and operational leaders can take within their current contexts.

When procuring or renewing medical record systems, ask vendors specific questions about FHIR implementation scope. Not just whether they support FHIR, but which FHIR profiles, which resource types, and whether their API has been tested in production environments with external systems. Request evidence of interoperability with the systems most relevant to your clinical workflows.

When evaluating new clinical tools, including AI medical assistants, documentation tools, and clinical decision support systems, assess whether they're designed to work within existing medical record system infrastructure or require parallel data entry. Tools that reduce documentation burden without integrating into the medical record system may shift work rather than eliminate it.

When engaging with national EHDS implementation processes, healthcare organisations have the opportunity to influence how member states implement the regulation's requirements. National contact points for health data and eHealth coordination bodies are the relevant forums. Clinical input into these processes is valuable and frequently undersupplied.

When reviewing internal workflows, map the points at which clinicians are manually re-entering data that already exists in another system. These are the highest-priority interoperability failure points, both for patient safety and for staff time, and they represent the strongest business case for investment in integration.

Interoperability is a clinical issue, not just an IT issue

The framing of interoperability as a technical problem to be solved by IT departments has historically kept it away from the clinical leadership attention it deserves. The evidence is clear that fragmented clinical systems affect patient safety, increase diagnostic error risk, delay care, and contribute to admin burden and clinician burnout. These are clinical outcomes, and they require clinical leadership engagement.

Progress is real. The EHDS Regulation represents a structural shift in how European health data governance is approached. Survey data shows that 64 per cent of European healthcare IT leaders accelerated FHIR and API modernisation programmes in the twelve months to late 2025, a meaningful shift in procurement behaviour. Technical standards have matured, and the regulatory framework is now in place to mandate their adoption.

The gap between policy and practice remains wide. Implementation timelines run to 2031. Legacy systems won't be replaced overnight. Governance arrangements that enable data to flow across institutional boundaries require sustained organisational effort that no regulation can substitute for. Closing that gap requires both the technical standards that are now largely available and the organisational will that is still, in many settings, the limiting factor.

Frequently asked questions

▶ What does healthcare interoperability actually mean?

Healthcare interoperability refers to the ability of different medical record systems, diagnostic platforms, and digital tools to exchange, interpret, and act on shared data. It operates across four levels: foundational (data can be transmitted), structural (data is formatted consistently), semantic (data carries shared meaning), and organisational (governance and institutional trust arrangements exist to allow data to flow). Systems can be technically capable of exchanging data and still fail to do so if the organisational layer, covering consent, liability, and access agreements, hasn't been established.

▶ Why is poor interoperability a clinical problem, not just an IT problem?

Fragmented clinical systems directly affect patient safety, increase diagnostic error risk, delay care, and add to documentation burden and clinician burnout. When a clinician can't access a complete patient history during a ward round or outpatient consultation, they must either work with incomplete information or spend time chasing records across systems. Research on medication data fragmentation found that the absence of interoperable medication records creates administrative burden through redundant data entry and increases the risk of errors. These are clinical outcomes that require clinical leadership engagement.

▶ Why does Europe face a particularly complex interoperability problem?

Health system organisation is a national competency across EU member states, so medical record system procurement, data governance, and interoperability policy have historically been decided at country level or lower. The result is a continent-wide patchwork of incompatible systems. Many hospitals run clinical platforms procured in the 1990s or early 2000s, built before modern interoperability standards existed. Medical record system vendors have historically maintained proprietary data formats and closed application programming interfaces as a commercial strategy, making integration costly. In countries with significant private care sectors, data sharing arrangements between public and private providers are often absent or legally ambiguous.

▶ What are FHIR, SNOMED CT, and HL7, and what can they actually deliver?

Health Level 7 (HL7) is a set of messaging standards defining how clinical information is structured and transmitted between systems. Fast Healthcare Interoperability Resources (FHIR), developed by HL7, is the current-generation standard using modern web application programming interfaces for more flexible data exchange. Systematised Nomenclature of Medicine Clinical Terms (SNOMED CT) and the International Classification of Diseases (ICD-10/11) provide the semantic layer, offering shared clinical terminology so a diagnosis coded in one system can be understood by another. The critical limitation is that having a standard doesn't guarantee adoption. A system can claim FHIR compliance while implementing only a subset of the specification, creating interoperability in theory and fragmentation in practice.

▶ What is the European Health Data Space and what does it require?

The European Health Data Space (EHDS), established under EU Regulation 2025/327, entered into force in March 2025. It sets out a framework for both patient care and secondary use of health data for research and policy. Medical record system manufacturers must achieve certification demonstrating compliance with harmonised interoperability standards, and systems that don't meet requirements won't be permitted to be marketed in Europe. Member states must establish national contact points for health data to facilitate cross-border exchange. The European Electronic Health Record Exchange Format will define mandatory data structures for patient summaries, ePrescriptions, discharge summaries, laboratory results, and medical images. The implementation timeline runs from 2027 to 2031.

▶ How does poor interoperability affect patients directly?

Fragmented systems lead to repeated diagnostic testing when a GP can't access hospital laboratory results or a specialist can't retrieve previous imaging. A nine-country survey of 417 European healthcare IT leaders found that 55 per cent of organisations reported reduced duplicate testing following improvements in data sharing, which suggests duplicate testing remains common where those improvements haven't been made. Care transitions are a particularly high-risk point: if a discharge summary arrives days late in an unstructured format the GP's medical record system can't read, medication changes, follow-up requirements, and new diagnoses may not reach the responsible clinician in time. For patients receiving care across EU borders, varying technical standards and coding systems create further care gaps.

▶ Can AI tools fix interoperability problems in clinical settings?

AI tools don't fix broken infrastructure. A system that can't exchange structured data between two medical record systems won't be repaired by adding an AI layer on top. However, ambient voice technology and real-time transcription can capture clinical encounters and generate structured notes without requiring clinicians to manually enter data into multiple systems, which addresses documentation burden at the point of care even when underlying systems remain disconnected. AI tools operating on fragmented, incomplete, or inconsistently coded data will produce outputs that reflect those limitations. Research on EU AI Act compliance also notes that AI-based clinical decision support systems are automatically classified as high risk under the EU AI Act, adding regulatory complexity for tools deployed in interoperability-sensitive contexts.

▶ What are the biggest barriers still blocking interoperability progress?

Vendor lock-in remains persistent: medical record system vendors have historically maintained closed systems as a competitive strategy, and enforcement of the EHDS's mandatory interoperability requirements is yet to be tested at scale. General Data Protection Regulation (GDPR) and data residency requirements create genuine complexity for cross-border data flows. Inconsistent FHIR adoption means vendor compliance claims frequently outrun production reality. Legacy system replacement is underfunded across most European health systems, and bridging older platforms to modern interoperability infrastructure is technically complex. Organisational alignment is perhaps the hardest problem: even where technical standards exist, institutions must agree on governance, data access policies, and liability arrangements before data can flow, and these arrangements are frequently the missing element.

▶ What can clinicians and healthcare leaders do now to improve interoperability?

When procuring or renewing medical record systems, ask vendors specific questions about FHIR implementation scope: which profiles, which resource types, and whether their application programming interface has been tested in production environments with external systems. When evaluating clinical tools including AI medical assistants, assess whether they integrate into existing medical record system infrastructure or require parallel data entry. Map the points in internal workflows where clinicians manually re-enter data that already exists in another system: these represent the strongest business case for integration investment and the highest-priority patient safety risks. Healthcare organisations can also engage with national EHDS implementation processes, where clinical input into eHealth coordination bodies is valuable and frequently undersupplied.