The study, conducted by researchers at SPMS, the shared services agency of Portugal’s Ministry of Health, combined a scoping literature review following PRISMA-ScR guidelines with a detailed regulatory analysis of the EHDS Regulation (EU) 2025/327. The team searched PubMed/MEDLINE, ScienceDirect, IEEE Xplore and Web of Science for publications from January 2016 to May 2026, deliberately spanning the period before, during and after the pandemic, and supplemented the search with grey literature from EUR-Lex, the European Centre for Disease Prevention and Control (ECDC), the World Health Organization, the OECD and EU-funded projects such as TEHDAS2 and EUCAIM. After screening 427 records, 35 scientific studies and six regulatory documents made the final cut.

The stakes are enormous. According to the WHO figures cited in the review, COVID-19 was responsible for approximately 7.0 million officially reported deaths and associated with around 22.1 million excess deaths globally between 2020 and 2023. Beyond the direct toll, the pandemic disrupted healthcare delivery, delayed diagnoses, and destabilised economies — and it exposed precisely how fragmented Europe’s health information systems really were. Surveillance infrastructures remain organised around disease-specific reporting systems, with fragmented data sources, limited accessibility, restricted reuse, and persistent difficulties in implementing electronic health records for anything beyond direct patient care.

The pandemic-era evidence assembled in the review is sobering. When researchers attempted to monitor COVID-19 across 31 European countries through primary care indicators, they found inconsistent recording, restricted public availability and poor comparability between nations. The ReCoDID project, which piloted linking laboratory data with clinical-epidemiological cohort data using a centralised approach, encountered significant delays tied to GDPR compliance requirements, limited alignment with FAIR data principles, and resource-intensive handling of centrally stored clinical data. Even real-world vaccine evidence — which proved invaluable for decisions on dosing regimens, booster strategies and responses to emerging variants — was constrained by poor data quality, uneven infrastructure and regulatory transparency gaps across countries.

These failures pushed the field decisively toward federated and decentralised architectures, in which data never leave their national or institutional environments but can still be analysed in place by authorised parties. This is the technical heart of the EHDS vision. The Regulation establishes health data access bodies (HDABs) as the central governance actors, mandates that health data holders make datasets available upon approved applications, and creates HealthData@EU, a common European infrastructure connecting national contact points, dataset catalogues and secure processing environments. Rather than shipping sensitive records across the continent, analysts run queries and algorithms where the data live — a model that preserves data sovereignty while enabling genuinely cross-border science.

Critically, the review highlights that the EHDS explicitly names surveillance and emergency preparedness as legitimate secondary-use purposes. Recital 2 of the Regulation states that COVID-19 demonstrated the importance of timely access to quality electronic health data for managing future pandemics, while Article 53 permits processing for public health activities including protection against serious cross-border threats. Article 51 defines the minimum data categories available for reuse: EHR data, determinants of health, pathogen data, population-based registries, aggregated healthcare resource data, and biobank holdings. Article 68 even shortens assessment timeframes for applications from public sector bodies pursuing epidemiological surveillance and emergency preparedness, and Article 71 allows national exceptions to opt-out rights when data use is necessary, proportionate, and no alternative exists.

The review also mines real implementation pilots for lessons. A study by Schneider and colleagues tested the HealthData@EU infrastructure through two use cases led by the European Medicines Agency and the ECDC — coagulopathy events in COVID-19 patients and antimicrobial resistance surveillance, respectively. The ECDC use case achieved high concordance between distributed analysis outputs and historical EARS-Net figures, proving the federated model can work. Yet decisions on data applications took roughly a year, hampered by variable national requirements and country-specific legal interpretations. Other examples reinforce both promise and friction: a Spanish regional hybrid federated data space shows the technical groundwork — OMOP terminology mapping, de-identification, cataloguing, quality pipelines — that HDAB-mediated access demands, while the EUCAIM cancer imaging infrastructure demonstrates pan-European federation preserving institutional data sovereignty. Federated models for automated infection surveillance, in which centrally developed algorithms run locally on EHR data, could deliver near-real-time, comparable indicators while source records stay under local control.

From this evidence, the authors construct a conceptual three-layer ecosystem. At the foundation sit distributed health data holders, retaining control in line with subsidiarity and data sovereignty. The middle layer is the EHDS itself — governance rules, interoperability standards, and the institutional federated system connecting data holders, HDABs, HealthData@EU and EU surveillance authorities through secure processing environments and data permits. The top layer is public health intelligence and action, where epidemiologists, statisticians and decision-makers convert EHDS-enabled data flows into continuous surveillance outputs. The model is designed to complement, not replace, existing systems such as TESSy and the Early Warning and Response System, extending their reach into non-communicable disease monitoring, One Health surveillance linking human, animal and environmental data, and AI-supported predictive analytics.

To make this operational, the authors propose three non-exclusive governance options grounded in specific articles of the Regulation. Option A designates a European health surveillance platform as an authorised participant in HealthData@EU under Article 75, letting bodies like ECDC, HERA and EMA connect directly and coordinate federated analyses across Member States — the most scalable option, but one demanding heavy governance and interoperability investment. Option B, based on Article 50(3), creates national health surveillance data intermediation entities that prepare, standardise and harmonise surveillance-relevant data on behalf of data holders, easing the burden on national systems but contributing less to EU-wide coordination. Option C designates a dedicated HDAB for public health intelligence under Article 55, building specialised expertise and faster access decisions at national level. The authors stress these can be combined — for instance, pairing a national surveillance HDAB with a European authorised participant to cover both domestic readiness and cross-border crisis response.

One caveat looms large: time. Even under accelerated EHDS procedures, HDAB decisions may take two months, extendable by one, with data holders allowed up to three months to make data available — timelines measured in months when emergencies demand answers in days or hours. The authors argue that crisis-ready procedures should be specified within the EHDS implementation guides, and that surveillance and emergency preparedness should be treated as priority use cases in future EU pilots, regulatory sandboxes and simulation exercises before the next pandemic arrives. If those tests succeed, Europe’s fragmented health records could finally cohere into something resembling a nervous system for the continent — one that senses threats early, thinks in federated harmony, and acts before an outbreak becomes a catastrophe.

Subject of Research: The role of the European Health Data Space regulation in modernising public health surveillance and emergency preparedness in Europe

Article Title: Modernising public health surveillance in an EHDS-enabled Europe: building a coordinated European public health intelligence ecosystem

Article References: Ferreira-Pinto, M., P. Silva, N., Silva Marques, S., Cruz, N., Gouveia Anjo, E., & S. Pinto, C. (2026). Modernising public health surveillance in an EHDS-enabled Europe: building a coordinated European public health intelligence ecosystem. Frontiers in Public Health, 14, Article 1974886. https://doi.org/10.3389/fpubh.2026.1974886

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1974886

Keywords: European Health Data Space, public health surveillance, emergency preparedness, HealthData@EU, secondary use of health data, electronic health records, federated data, ECDC, pandemic preparedness, health data governance, interoperability, public health intelligence

The European Cardiovascular Health Data and AI Network, known as CHAIN, has been launched with €20 million in funding from the EU4Health Programme and is coordinated by the European Society of Cardiology (ESC). The 36-month project brings together a consortium of 53 partners from 20 European countries, spanning universities and research centres, hospitals and clinical networks, technology companies, national ministries of health and patient organisations. Its central ambition is to build a federated ecosystem for the safe, effective and scalable adoption of AI in cardiovascular care, aligned with the European Health Data Space, the EU’s ambitious framework for cross-border health data sharing.

The problem CHAIN seeks to solve is well documented. AI models capable of detecting heart disease earlier, stratifying patient risk more accurately and supporting clinical decisions have proliferated in the scientific literature, yet their uptake in routine care remains patchy and inconsistent. Standardised validation mechanisms are largely lacking, meaning hospitals and health authorities often have no reliable way of knowing which algorithms perform as advertised, for which patient populations, and under which clinical conditions. The result is a landscape in which promising tools stall at the pilot stage while unvetted ones occasionally reach clinics without adequate scrutiny.

CHAIN’s technical answer to this challenge rests on two pillars: a federated data infrastructure and a system of validation credentials. Rather than centralising sensitive patient records, the project will securely connect hospitals, health data hubs and disease registries across Europe in a federated architecture, allowing AI solutions to be tested, validated and deployed at scale while data remain within their originating institutions. This approach, which aligns with the privacy-preserving principles underpinning the European Health Data Space, enables algorithms to be evaluated on diverse, real-world populations drawn from multiple healthcare systems, a critical step for detecting the biases and performance gaps that often emerge when models trained in one setting are applied in another.

The second pillar is a European repository of validated AI solutions, each accompanied by what the project calls an ‘AI passport’. These passports will document a model’s performance, safety and suitability for clinical use, giving healthcare authorities, hospital administrators and practicing clinicians clear, standardised evidence on which solutions work, for whom and in which settings. In an era when clinicians and patients alike are wary of opaque algorithmic tools, the passports are designed to function as a trusted seal of approval, helping decision-makers distinguish genuinely useful innovations from overhyped ones. Professor Folkert Asselbergs of Amsterdam University Medical Centre, CHAIN’s scientific coordinator and chair of the ESC AI Gateway, emphasised that in an era of misinformation, patients and healthcare professionals are looking to medical societies to provide a framework for implementing trusted solutions that add real value beyond what is currently available.

To prove the framework works in practice, CHAIN will implement six real-world use cases across 13 countries, deliberately chosen to cover the full continuum of cardiovascular care. The use cases span risk stratification, early disease detection, clinical decision support and the management of complex cardiovascular conditions, ensuring that the validation and deployment mechanisms are tested across different healthcare settings, from large academic centres to smaller regional hospitals. This breadth matters: an AI tool that performs well in a well-resourced Dutch university hospital may fail in a rural clinic in eastern Europe, and the project’s multi-country design is intended to surface such disparities before tools are scaled.

The initiative does not start from scratch. CHAIN builds on the ESC’s EuroHeart registry network, which already unifies 18 ESC member countries with standardised data on cardiovascular care and outcomes, providing tools that help countries monitor results, compare practices and drive sustainable improvements. It will also seek synergies with a constellation of existing EU projects, including EUCAIM, COMPASS AI, AI4HF, DataTools4Heart, EHDEN, JACARDI, TEHDAS2 and Xt-EHR, an effort to maximise complementarities and avoid duplicating infrastructure that Europe has already invested in. The project is explicitly positioned as a key implementation vehicle for the EU Safe Hearts Plan, in particular its flagship initiative on innovation and integration of AI and digital technologies in cardiovascular healthcare.

Timing is a central theme of the launch. CHAIN arrives just as the European Health Data Space, the EU’s landmark AI Act, and an evolving regulatory framework for medical products, spanning pharmaceutical, medical device and in vitro diagnostic legislation, are reshaping how AI is developed, evaluated and deployed across European healthcare. Navigating this regulatory landscape is one of the project’s implicit challenges: any AI tool validated through CHAIN will need to satisfy not only clinical evidence standards but also the transparency, safety and human-oversight requirements that the AI Act imposes on high-risk medical applications. By building governance, evaluation and scale-up mechanisms into its core design, the consortium hopes to create a template that regulators, industry and health systems can adopt more widely.

The political weight behind the project is considerable. Olivér Várhelyi, the EU Commissioner for Health and Animal Welfare, framed the challenge bluntly, saying the question is no longer simply what AI can do, but how to make it work safely and effectively for patients and health professionals in everyday healthcare, and that this is where CHAIN can make a real difference. Professor Cecilia Linde, President of the ESC and a cardiologist at the Karolinska Institute in Sweden, struck a similar note, arguing that the need to act is critical given that one in three deaths in Europe is caused by cardiovascular disease, and describing CHAIN as an effort to turn the promise of AI into better cardiovascular health for everyone by helping trusted innovations reach patients safely and equitably.

Over the next three years, the consortium hopes to deliver a strategic roadmap for scaling AI adoption in cardiology across Europe, one that Professor Asselbergs says will benefit research and industry through shared infrastructure and standards, while validated decision-support tools help reduce the burden on overstretched healthcare systems. The ultimate measure of success, he argues, will be clinical: trustworthy tools that translate into earlier, more accurate diagnosis and safer, more personalised care for patients. If CHAIN succeeds, it could offer something Europe has lacked so far, a coherent, continent-scale pathway for moving AI in medicine from promising publications to measurable patient impact, and a model that other fields of medicine, from oncology to neurology, may soon seek to replicate.

Subject of Research: A European initiative to validate and scale artificial intelligence for cardiovascular disease prevention, detection and treatment

Article Title: From promise to patient impact: New European initiative aims to accelerate AI in cardiovascular care

Article References: From promise to patient impact: New European initiative aims to accelerate AI in cardiovascular care. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: artificial intelligence, cardiovascular disease, CHAIN, European Society of Cardiology, EU4Health, European Health Data Space, AI validation, federated data, clinical decision support, health data infrastructure, AI passports, digital health