Periodic Reporting for period 1 - WISDOM (Well-being improvement through the Integration of healthcare and reSearch Data and models with Out border for chronic iMmune-mediated diseases)
Período documentado: 2023-12-01 hasta 2025-05-31
Chronic immune-mediated diseases (CIMDs) such as multiple sclerosis, rheumatoid arthritis, and myasthenia gravis affect nearly 10% of the global population and place an enormous burden on patients, healthcare systems, and societies in Europe. Despite advances in therapies, diagnosis is often delayed, treatment choices remain insufficiently personalised, and the unpredictable course of disease leads to both over- and under-treatment. At the same time, vast amounts of clinical, registry, and research data remain fragmented and underused due to technical, regulatory, and ethical barriers.
WISDOM brings together leading European universities, SMEs, and patient organisations to address these challenges. By harmonising health data across borders and developing secure, explainable AI-driven predictive models, WISDOM will enable earlier diagnosis, better therapy selection, and more effective disease monitoring. Its co-design approach ensures that tools are patient-centred, transparent, and accessible, empowering individuals to take an active role in managing their health and supporting equity of care across diverse populations.
Strategically, the project contributes to EU ambitions to strengthen digital health sovereignty, foster trust in AI, and reduce inequalities in access to healthcare. It directly supports Horizon Europe’s goal of “Unlocking the full potential of new tools, technologies and digital solutions for a healthy society,” aligns with the European Health Data Space initiative, and advances GDPR-compliant frameworks for data sharing. By setting ethical, legal, and technical standards, WISDOM will promote safe and responsible use of digital health technologies.
The project’s expected impacts are significant: improved outcomes and quality of life for patients; reduced healthcare costs linked to avoidable disability; and greater citizen trust in data sharing and digital tools. By ensuring equitable, patient-centred access to innovation, WISDOM not only addresses CIMDs but also creates a scalable blueprint for managing other chronic diseases, strengthening Europe’s leadership in responsible, inclusive healthcare innovation.
WISDOM brings together leading European universities, SMEs, and patient organisations to address these challenges. By harmonising health data across borders and developing secure, explainable AI-driven predictive models, WISDOM will enable earlier diagnosis, better therapy selection, and more effective disease monitoring. Its co-design approach ensures that tools are patient-centred, transparent, and accessible, empowering individuals to take an active role in managing their health and supporting equity of care across diverse populations.
Strategically, the project contributes to EU ambitions to strengthen digital health sovereignty, foster trust in AI, and reduce inequalities in access to healthcare. It directly supports Horizon Europe’s goal of “Unlocking the full potential of new tools, technologies and digital solutions for a healthy society,” aligns with the European Health Data Space initiative, and advances GDPR-compliant frameworks for data sharing. By setting ethical, legal, and technical standards, WISDOM will promote safe and responsible use of digital health technologies.
The project’s expected impacts are significant: improved outcomes and quality of life for patients; reduced healthcare costs linked to avoidable disability; and greater citizen trust in data sharing and digital tools. By ensuring equitable, patient-centred access to innovation, WISDOM not only addresses CIMDs but also creates a scalable blueprint for managing other chronic diseases, strengthening Europe’s leadership in responsible, inclusive healthcare innovation.
In WP 1 we have:
• Completed State of the Art Review on legal, ethical, and policy barriers to data integration and AI use.
• Conducted analysis of biases, discrimination risks, patients’ and disability rights, identifying six key ethical themes. Report on ELSI aspects nearly complete.
• Ethical insights structured into themes; early guideline framework trustworthy AI applications initiated.
In WP2 we have:
• Developed and delivered Data Management Plan aligned with FAIR principles.
• Collected metadata and conducted dataset mapping.
• Data harmonisation and anonymisation workflows developed, with tests on FinRegistry and Karolinska datasets. Harmonised Core Dataset delivered
• Successful harmonisation of multimodal core dataset for downstream modelling
• Designed and deployed the WISDOM Warehouse IT platform with GDPR-compliant secure upload, storage, and federated access .
In WP3 we have:
• Developed disease-risk stratification models for MS, with RA and MG in preparation.
• MS outcome prediction developed and validated across Italian, Swedish, and German datasets.
• Framework for personalised interventions and drug-response modelling (multi-task learning, AlphaFold, transformers) in progress
In WP4 we have:
• Confirmed predictive performance of MS models in an external (German) population (cross-border validation)
• Established infrastructure and preparatory datasets for future validation and transferability testing. Prepared AutoML-Med tool for external dataset applications.
• Initiated synthetic dataset generation using VEIL.AI for secure sharing.
Set up federated learning environment (SAPU) at Estonian Biobank data.
• Planned cross-disease transfer of MS frameworks to RA and MG.
In WP5 we have:
• Began prospective clinical validation of risk-stratification and outcome-prediction tools. Successful early testing of MS outcome prediction tool at 24 months post-diagnosis.
• Launched the PREDICT cohort in Germany, with expansion planned in Sweden and Italy.
• Collected patient data and biosamples for validation studies.
• Tested outcome-prediction tools on an existing prospective MS cohort.
In WP6 we have:
• Conducted innovation activities: idea catalogue, use case scenarios, prototypes.
• Assessed stakeholder needs for clinical adoption of AI tools.
• Progressed with innovation roadmap towards clinical demonstrators.
• On track to deliver demonstrator validation (TRL 5–6) in lab and real-world environments.
In WP7 we have:
• Facilitated knowledge transfer (summer school scheduled for Sep 2025).
• Effective project management, with timely reporting and communication.
• Scientific dialogue and partner coordination well established.
• Baseline KPI have been identified and reported.
• Completed State of the Art Review on legal, ethical, and policy barriers to data integration and AI use.
• Conducted analysis of biases, discrimination risks, patients’ and disability rights, identifying six key ethical themes. Report on ELSI aspects nearly complete.
• Ethical insights structured into themes; early guideline framework trustworthy AI applications initiated.
In WP2 we have:
• Developed and delivered Data Management Plan aligned with FAIR principles.
• Collected metadata and conducted dataset mapping.
• Data harmonisation and anonymisation workflows developed, with tests on FinRegistry and Karolinska datasets. Harmonised Core Dataset delivered
• Successful harmonisation of multimodal core dataset for downstream modelling
• Designed and deployed the WISDOM Warehouse IT platform with GDPR-compliant secure upload, storage, and federated access .
In WP3 we have:
• Developed disease-risk stratification models for MS, with RA and MG in preparation.
• MS outcome prediction developed and validated across Italian, Swedish, and German datasets.
• Framework for personalised interventions and drug-response modelling (multi-task learning, AlphaFold, transformers) in progress
In WP4 we have:
• Confirmed predictive performance of MS models in an external (German) population (cross-border validation)
• Established infrastructure and preparatory datasets for future validation and transferability testing. Prepared AutoML-Med tool for external dataset applications.
• Initiated synthetic dataset generation using VEIL.AI for secure sharing.
Set up federated learning environment (SAPU) at Estonian Biobank data.
• Planned cross-disease transfer of MS frameworks to RA and MG.
In WP5 we have:
• Began prospective clinical validation of risk-stratification and outcome-prediction tools. Successful early testing of MS outcome prediction tool at 24 months post-diagnosis.
• Launched the PREDICT cohort in Germany, with expansion planned in Sweden and Italy.
• Collected patient data and biosamples for validation studies.
• Tested outcome-prediction tools on an existing prospective MS cohort.
In WP6 we have:
• Conducted innovation activities: idea catalogue, use case scenarios, prototypes.
• Assessed stakeholder needs for clinical adoption of AI tools.
• Progressed with innovation roadmap towards clinical demonstrators.
• On track to deliver demonstrator validation (TRL 5–6) in lab and real-world environments.
In WP7 we have:
• Facilitated knowledge transfer (summer school scheduled for Sep 2025).
• Effective project management, with timely reporting and communication.
• Scientific dialogue and partner coordination well established.
• Baseline KPI have been identified and reported.
Results Beyond the State of the Art
• Provided comparative analysis of EU, US, and Canadian AI healthcare regulations.
• Identified six key ethical themes and developed training modules for ethical self-assessment, aligned with GDPR, AI Act, MDR, EHDS, HIPAA, and FDA.
• Delivered a harmonised multimodal dataset, interoperable with ELIXIR/EuroHPC, and established Warehouse for anonymised, federated data access.
• Developed and validated predictive models for early MS detection (C-index 0.92) and outcomes, with demonstrated transferability across national health systems.
• Introduced novel findings on comorbidity network dynamics in MS prodrome and created a VAE for HLA imputation with cross-disease potential.
• Advanced multi-task AI frameworks for drug–protein interaction and established infrastructure for federated/distributed learning.
• Enrolled 94 high-risk individuals in the PREDICT cohort, generating longitudinal clinical, imaging, and biomarker data for predictive modelling.
• Defined an innovation pipeline from idea to validated demonstrator (TRL 5–6), and actively disseminated results via multiple channels.
• Demonstrated scientific, economic, societal, and clinical benefits, including model generalisability, reduced redundant development, equitable access, and improved personalised MS treatment.
Potential Impacts
• Enhances integration of ethical/legal principles in AI tool design, supporting compliance with EU regulations and public trust.
• Provides a reusable data foundation for predictive AI, advancing cross-disease modelling and collaboration.
• Accelerates SME and industry product development, supports pharma R&D, and enables secure, equitable data sharing.
• Facilitates earlier diagnosis and personalised treatment, strengthening EU leadership in AI-driven personalised medicine.
Key Needs for Further Uptake
• Ongoing engagement with regulators and health systems for clinical deployment and alignment.
• Standardised ethical self-assessment protocols and regulatory integration guidance.
• Demonstration of scalability, sustainability, and integration with national infrastructures.
• Commercial exploitation strategies for WISDOM Warehouse and predictive tools.
• Expanded validation across diseases, industrial engagement, and completion of long-term cohort follow-up.
• Regulatory support for biomarker integration, stronger commercialisation efforts, and a sustainability roadmap beyond project lifetime.
• Provided comparative analysis of EU, US, and Canadian AI healthcare regulations.
• Identified six key ethical themes and developed training modules for ethical self-assessment, aligned with GDPR, AI Act, MDR, EHDS, HIPAA, and FDA.
• Delivered a harmonised multimodal dataset, interoperable with ELIXIR/EuroHPC, and established Warehouse for anonymised, federated data access.
• Developed and validated predictive models for early MS detection (C-index 0.92) and outcomes, with demonstrated transferability across national health systems.
• Introduced novel findings on comorbidity network dynamics in MS prodrome and created a VAE for HLA imputation with cross-disease potential.
• Advanced multi-task AI frameworks for drug–protein interaction and established infrastructure for federated/distributed learning.
• Enrolled 94 high-risk individuals in the PREDICT cohort, generating longitudinal clinical, imaging, and biomarker data for predictive modelling.
• Defined an innovation pipeline from idea to validated demonstrator (TRL 5–6), and actively disseminated results via multiple channels.
• Demonstrated scientific, economic, societal, and clinical benefits, including model generalisability, reduced redundant development, equitable access, and improved personalised MS treatment.
Potential Impacts
• Enhances integration of ethical/legal principles in AI tool design, supporting compliance with EU regulations and public trust.
• Provides a reusable data foundation for predictive AI, advancing cross-disease modelling and collaboration.
• Accelerates SME and industry product development, supports pharma R&D, and enables secure, equitable data sharing.
• Facilitates earlier diagnosis and personalised treatment, strengthening EU leadership in AI-driven personalised medicine.
Key Needs for Further Uptake
• Ongoing engagement with regulators and health systems for clinical deployment and alignment.
• Standardised ethical self-assessment protocols and regulatory integration guidance.
• Demonstration of scalability, sustainability, and integration with national infrastructures.
• Commercial exploitation strategies for WISDOM Warehouse and predictive tools.
• Expanded validation across diseases, industrial engagement, and completion of long-term cohort follow-up.
• Regulatory support for biomarker integration, stronger commercialisation efforts, and a sustainability roadmap beyond project lifetime.