The AID4RARE consortium has joined 5 European Reference Networks (the EC program networking leading experts in RDs) together with academic and industrial experts in computer science to develop novel AI-based solutions:

1. a Natural Language Processing (NLP) tool to extract and structure information from unstructured data from different clinical sources;
2. an interoperability platform to share data among EHR and RD registries; and
3. an AI-powered virtual assistant to detect patients at risk for RD.

Artificial Intelligence and Data Structuring for Rare Diseases- AID4RARE

Rare Diseases (RDs) affect 30M people in Europe. The process to a correct diagnosis and treatment for these patients is often long and cumbersome, and in many cases RDs are left undiagnosed or inadequately treated. The root cause for this can be attributed to under-representation in research and treatment development, due mainly to data unavailability. This requires the development of innovative artificial intelligence (AI)-based solutions to challenge patients? data in EHR to improve research, disease classification, and treatment.

Electronic Health Records (EHR) are generated in hospitals or RD registries, which could be used and reused to reduce the time to the right diagnosis, but data are currently unstructured and lack interoperability to allow their use among different IT tools. In this context, RDs need AI. To fill this gap, the AID4RARE consortium has joined 5 European Reference Networks (the EC program networking leading experts in RDs) together with academic and industrial experts in computer science to develop novel AI-based solutions:

1] a Natural Language Processing (NLP) tool to extract and structure information from unstructured data from different clinical sources;

2] an interoperability platform to share data among EHR and RD registries;

3] an AI-powered virtual assistant to detect patients at risk for RDs.

These tools will be tested in use-cases of selected rare metabolic, kidney, endocrine, blood and skeletal diseases in accordance with current legislation, ethics, personal data protection, and cybersecurity. The tools will also contribute to feeding data into ERN registries to demonstrate improved usability of health data and help the identification of patients to be enrolled in small population clinical trials to develop innovative treatments. Once tested and validated in the field of RDs, the novel solutions developed by the AID4RARE project could be applied to more common diseases to improve the care of many other patients in Europe.