Real-world-data Enabled Assessment for heaLth regulatory decision-Making

Projektinformationen

REALM

ID Finanzhilfevereinbarung: 101095435

DOI

10.3030/101095435

EK-Unterschriftsdatum 24 November 2022

Startdatum 1 Januar 2023

Enddatum 31 Dezember 2026

Finanziert unter

• Health

Gesamtkosten

€ 6 659 650,00

EU-Beitrag

€ 6 659 650,00

Koordiniert durch

UNIVERSITEIT MAASTRICHT
Netherlands

Periodic Reporting for period 2 - REALM (Real-world-data Enabled Assessment for heaLth regulatory decision-Making)

Berichtszeitraum: 2024-07-01 bis 2025-12-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

The REALM project addresses the growing need for a standardized, transparent, and collaborative framework for evaluating and certifying healthcare software. As medical software increasingly plays a critical role in healthcare delivery, the lack of consistent regulatory guidelines and evaluation processes poses significant challenges. REALM aims to bridge this gap by creating a regulatory sandbox environment where developers, regulatory authorities, healthcare professionals, and policy officers can collaborate. This platform will standardize tools and data resources, ensuring that healthcare software meets high-quality standards across the European Union and beyond.

The project's impact is expected to be substantial, fostering trust in digital health technologies and enabling faster, more efficient certification processes. By integrating real-world data, synthetic data, and a living lab environment, REALM will enhance the reliability and safety of medical software. The inclusion of social sciences and humanities in the project ensures that legal, ethical, and human-centered considerations approaches are integral to the software development and evaluation process, further aligning the project with EU policy priorities.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

During the second phase of the REALM project, significant progress was made in technically integrating the core components of the REALM framework. Major efforts were dedicated to finalising the Federated REALM Architecture (D3.1) advancing the Regulatory Toolbox Methodology (D3.2) implementing the blockchain-supported model card lifecycle tracking (D3.3) and operationalising the federated cloud-based data repository and data catalogue (D4.1). In parallel, the Standard Data Quality Framework (D4.4) the Multimodal Synthetic Data Generator (D4.6) and the intelligent analytics and post-market surveillance tools (D5.1) were developed and progressively integrated into the platform workflow. Demonstrator-specific pipelines were constructed, interoperability across components was tested, and a dedicated end-to-end demonstration session supported technical alignment and identification of remaining integration gaps. The platform now supports federated data access, OMOP-based structured querying, synthetic data generation, explainable AI methods, bias and fairness analysis, and blockchain-backed model documentation in a coordinated evaluation workflow. In parallel, regulatory and ethical alignment was strengthened through a detailed analysis of the interplay between MDR, IVDR, AI Act, and GDPR (D2.1) and with engagement in standardisation and policy discussions at EU level. These components collectively constitute a robust platform that will enhance the reliability, transparency, and efficiency of healthcare software certification. Testing and integration of these tools has begun, marking significant technical milestones in the project.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

REALM has delivered a portfolio of KERs that strengthen the European framework for evaluating, certifying, and governing AI-driven medical device software (MDSW). The project combines data quality, interoperable infrastructure, certification transparency, regulatory guidance, and capacity building into one coherent approach.

The Multimodal Standard Data Quality Framework (SDQF) provides a regulator-oriented method to assess the quality and robustness of real-world and synthetic data used for MDSW evaluation, supporting reproducible evidence generation. The REALM Data Catalog enables harmonised use of real-world, synthetic, and digital twin data for benchmarking and regulatory-grade validation, while highlighting remaining needs around data governance and licensing.

REALM’s federated infrastructure supports distributed monitoring and evaluation of MDSW performance across multiple nodes, demonstrated across five use cases and designed to support post-market oversight. The blockchain-based certification process embeds certification steps in smart contracts to enable traceable and auditable workflows; further maturation and alignment with standards are needed for broader deployment.

These KERs form the technical basis for a planned spin-off led by VITO and UM, targeting medical AI developers requiring regulatory-grade validation services. An EIC Transition proposal (submitted September 2025) supports further TRL and market validation. In parallel, the REALM Academy translates project methods into training for competent authorities, notified bodies, HTA agencies, and manufacturers, with sustainability options under exploration. Horizon Results Booster activities are supporting exploitation planning and next steps