Follow us on X
Follow us on Linkedin
Real-world-data enabled assessment
for health regulatory decision-making


The REALM project will provide a system-wide testing infrastructure, benchmarking data, and standardisation for the evaluation and certification of Medical Device Software in the healthcare industry. The lack of a framework for the certification and post-market monitoring of autonomous, self-learning, and potentially complex Medical Device Software, combined with the variance between performance in artificial testing environments and actual practice settings, makes it necessary to extend existing regulatory guidelines and provide data and technology to facilitate system-wide testing and certification of Medical Device Software across the EU. The REALM project aims to develop a federated cloud-based data resources catalogue, linking unstructured health data, clinical images, and real-world data expanded by synthetic data generation and digital twins, and organising this multimodal information in appropriate formats and standardised data structures.


Objective 1
Design optimised guidelines to improve the assessment of clinical/medical software, supporting heterogeneous and multimodal real world health data.

Objective 2
Develop a decentralised secure evaluation and certification framework of health innovations and medical algorithms under the full compliance with General Data Protection Regulation, considering Blockchain technologies supported by intelligent contracts offering repeatability, transparency, authentication, and traceability.

Objective 3
Implement advanced data pre-processing, natural language processing & (deep) machine learning tools and models to process and extract (meta-) information from real-world data supporting unstructured & heterogeneous Electronic Health Records extracting health information and improving data-centric decision-making and evaluation.

Objective 4
Develop AI-powered regulatory tools that will ingest the real-world health data and metadata, supporting multiple input interfaces, and employ the machine learning models to enable the evaluation of algorithms and models in terms of bias, privacy, accuracy, legal accountability, ethics, etc. in an optimised workflow, improving data driven decision making.

Objective 5
Develop and deploy explainable AI approaches to promote trust and adoptability of the machine learning solutions: The proposed framework will develop techniques for addressing concerns due to lack of transparency, accountability and explainability as well as due to potential biases induced by model predictions.

Objective 6
Enable synthetic data and medical/healthcare digital twins for medical data emulation and generation, for training and testing medical algorithms in realistic scenarios or scenarios with little-to-none available samples (rare diseases). This will enable data analytics and decision making on a vast medical data.

Objective 7
Adopt state-of-the-art European and Open Data standards and develop novel protocols for health data sharing and medical innovations and tools assessment.

Objective 8
Enable optimum selection of real-world data parameters in post-market surveillance of medical/healthcare software to achieve minimally invasive quality control.

Objective 9
Dissemination of knowledge among regulatory bodies across the EU in using state-of-the-art tools and best practices in evaluating medical/healthcare software under Medical Devices Regulation and In-Vitro Diagnostic Device Regulation using real-world data.