A €20 million European project will test six cardiovascular AI applications across 13 countries. Its 53 partners will assess their clinical value and safety without centralising patient data.
Hospitals in 13 countries will assess the tools in clinical settings while keeping personal health data under local control.
The European Society of Cardiology announced CHAIN on 7 October 2026. The €20 million project is co-funded through the EU4Health programme and coordinated by the ESC. It will assess six applications for cardiovascular care, examining where AI can support prevention and early detection. Treatment is another focus, alongside the conditions needed for safe and useful clinical practice.
CHAIN has 53 partners from 20 countries. The consortium includes hospitals and universities, along with research centres. Technology companies and health authorities are also involved, as are patient organisations. Spain’s Centro Nacional de Investigaciones Cardiovasculares Carlos III, or CNIC, is among the partners. The 20-country membership is separate from the 13 countries where the clinical scenarios will be assessed.
Algorithm performance is only part of the test.
CHAIN will examine clinical usefulness and fairness. It will also assess transparency and ease of use, then check whether the tools work consistently across patient groups and healthcare settings. The use cases include cardiovascular risk prevention and stratification. Others cover early disease detection and clinical decision support. A further focus is managing complex cardiovascular conditions. The project runs for 36 months and aims to build a shared European infrastructure for evaluating AI in cardiovascular care.
That infrastructure, CardioData@EU, is designed as a federated ecosystem aligned with the European Health Data Space. It will connect hospitals and clinical registries in different countries for evaluation. Other health-data sources will also be linked. Personal health data will remain under local control rather than being gathered in a central database, allowing researchers to compare tools across populations without moving the underlying records.
CHAIN also plans a European repository of validated AI solutions. Each tool will have an «AI passport» describing its performance and safety. The passport will also set out its suitability for clinical use. Hospitals, clinicians and health authorities could use that evidence to judge which systems fit their needs. The project aims to support safe and effective AI use. It is not intended to replace clinical decision-making.
For CNIC, the work starts with the foundations needed to compare systems. The centre will contribute to shared data standards and interoperability. Data quality is another focus. Coordinated by Héctor Bueno and Fátima Sánchez-Cabo, the CNIC team will join a work package led by the University of Luxembourg. It will develop common clinical definitions and data-quality indicators. The package will also set technical and security specifications for connecting institutions.
Inconsistent records can undermine comparisons between hospitals or patient groups. Common criteria should also help researchers identify bias. They will be used to test whether algorithms produce consistent results across different populations. CHAIN plans to carry out its assessments in varied clinical settings across the participating countries.
The ESC says CHAIN is intended to make cardiovascular care safer and more equitable through reliable AI. CNIC scientific director Borja Ibáñez has said the centre will bring scientific rigour to evaluating tools designed to support prevention. Patient care is also part of that work.
Outcome measures are also under debate in Spain’s thyroid cancer care. AECAT has called for assessments to include lasting symptoms and patients’ ability to return to everyday life.
CHAIN will build on existing ESC initiatives such as EuroHeart and contribute to the EU’s Safe Hearts Plan. Its three-year programme includes a shared roadmap for evaluating and adopting AI in cardiology. Clinical assessments will take place across 13 countries.
