Descripción del proyecto
Integrated, high-quality personal health data (PHD) represents a potential wealth of knowledge for healthcare systems, but there is no reliable conduit for this data to become interoperable, AI-ready and reuse-ready at scale across institutions, at national and EU level. AIDAVA will fill this gap by prototyping and testing an AI-powered, virtual assistant maximizing automation of data curation & publishing of unstructured and structured, heterogeneous data. The assistant includes a backend with a library of AI-based data curation tools and a frontend based on human-AI interaction modules that will help users when automation is not possible, while adapting to users? preferences. The interdisciplinary team of the consortium will develop and test two versions of this virtual assistant with hospitals and emerging personal data intermediaries, around breast cancer patient registries and longitudinal health records for cardio-vascular patients, in three languages. The team will work around four technology pillars: 1) automation of quality enhancement and FAIRification of collected health data, in compliance with EU data privacy; 2) knowledge graphs with ontology-based standards as universal representation, to increase interoperability and portability; 3) deep learning for information extraction from narrative content; and 4) AI-generated explanations during the process to increase users? confidence. By increasing automation of data quality enhancement, AIDAVA will decrease the workload of clinical data stewards; by providing high-quality data, AIDAVA will improve the effectiveness of clinical care and support clinical research. In the long-term, AIDAVA has the potential to democratise participation in data curation & publishing by citizens/patients leading to overall savings in health care costs (through disease prevention, early diagnosis, personalized medicine) and supporting delivery of the European Health Data Space.