PREPARE Preterm aims to develop early personalised risk prediction and patient stratification after very preterm birth (VPT, 1-2% of births, >50.000 annually in Europe) using expert-guided AI methods and the world’s most comprehensive data platform of children and adults born VPT. Survivors of VPT birth face high risks of multiple health and developmental problems compared with children born at term, leading to significant health, social and educational costs. Early personalised prediction is needed for patient stratification to optimise known effective interventions to prevent and mitigate adverse outcomes. Work on prediction models for individuals born VPT is very limited despite growing knowledge about risk and resilience factors and innovation in analytic and computational methods. Our mission is to mobilise expertise and innovation to shift research from describing the consequences of VPT birth to acting on and improving them. The project develops 4 components needed for early personalised prediction, effective patient stratification and societal impact: (1) a theoretical framework integrating lived experience that identifies life-course resilience factors to predict health, development, participation in society and quality of life; (2) integrated and dynamic data from population cohorts, e-cohorts and Big Data registers; (3) predictive models based on explainable and responsible AI (domain expert involvement, participatory research; protocols for sharing model performance, federated learning) and (4) tools that promote use of data for optimal healthcare decisions and health literacy. The project leverages a H2020 project that created an unrivalled platform of 23 European VPT cohorts and the Nordic registers and fosters Open Science (RECAP Preterm, project 01/03/2017-30/09/2021. To achieve its aims, this project will expand this platform to include data from population-based and clinical cohorts, including routinely generated real-life data, to integrate new computational capabilities and to create innovative tools for research and clinical use. These tools will facilitate follow-up using the platform which can allow inputs from geographic areas which have been underrepresented in research on VPT birth and expand and update prediction models to reflect advances in care for VPT babies. Finally, we will explore the opportunity for joint activities by networking with other European research platforms using birth data as well as infrastructure initiatives to promote population health research and data science. The development and validation of personalised prediction models for patient stratification based on AI and Big Data technologies by a leading interdisciplinary team with a proven record of working together will lead to clear improvement of outcomes for individuals, care systems and the wider society.