Edition of 23 April 2024

A new article has been published in the Orphanet Journal of Rare Diseases analysing the current landscape of European data sources which could be used for machine learning (ML) in the context of rare disease screening. The authors also propose a set of recommendations to provide a better understanding of the health data ecosystem, and to help inform future work of the Screen4Care project.

Because patient populations in rare diseases are small and geographically scattered, data sharing is crucial for conducting quality clinical and epidemiological research. It is also the only way to obtain enough data to properly train ML tools, which can be a powerful resource in supporting clinical decision-making and improving the quality of healthcare available to people living with a rare disease.

Between March-December 2022, researchers administered a semi-structured questionnaire to representatives from European rare disease databases to determine to what extent the databased adhered to FAIR (findable, accessible, interoperable, reusable) principles, as well as the legal and business aspects of database management. They found that there are significant barriers to the secondary use of health data in Europe, due to insufficient adherence to FAIR principles and low willingness of databases to share patient information.

Based on the study findings, the authors outline a number of recommendations for how EU databases can better support the development of ML tools and rare disease screening protocols. These are primarily centred around better adherence to FAIR principles, as well as some which pertain to legal and business concerns. Among others, some specific actions from these recommendations include:

• Databases with standardised data sharing practices and data formats ensuring consistency and interoperability across different databases should be used as a primary data source for ML applications;
• Collaborative research initiatives should be sought by researchers to identify databases with standardised data sharing practices and adherence to common data exchange protocols;
• A solid understanding of data protection law should be obtained to guarantee that IT security standards are strictly followed;
• Improved collaboration with ERNs and Healthcare institutions on country and EU level could accelerate EU and local initiatives to bring increased data sharing and accessibility for sustainable innovation in rare disease diagnosis and treatment.

Recommendations on data FAIRness and legal & business considerations (open in new tab to enlarge)

Overall, as new technologies continue to be developed which can benefit the rare disease community, the need for robust, high-quality and FAIR data sources will only grow. In order to continue improving the quality of care available to people living with a rare disease, it is vital that European health databases increase their adherence to best practices and their willingness to share patient information with researchers. In light of incoming legislation such as the EU Data Governance Act and the European Health Data Space, it is hoped that data sharing can be stimulated in a way that promotes accessibility and sustainable, ethical innovation in rare disease care.

Screen4Care was established in 2021 as a public-private-patient consortium, and takes an innovative approach to accelerating rare disease diagnosis through research. The five-year project explores how both genetic newborn screening and digital technologies can be leveraged to shorten the diagnostic odyssey, and ensure people living with a rare disease can achieve timely access to a diagnosis, and consequent treatment and care. It is funded under the Innovative Medicines Initiative 2 Joint Undertaking, which receives support from the EU’s Horizon 2020 program and the European Federation of Pharmaceutical Industries and Associations (EFPIA).