Making big data meaningful for a promising start

Background The first 1000 days of a child are of fundamental importance for healthy physical, cognitive and social-emotional development. A poor start can have lifelong consequences, for physical and mental health, and economic and social participation. Many children lead a less flourishing life due to adverse conditions in-utero and postnatally, and the structural drivers of these conditions outside the health sector. Early preventive interventions and policies can have huge, long-run benefits for individuals and society. This requires 1) support to professionals to better identify vulnerable (expectant) parents and children and provide tailored care, and 2) support to municipalities for evidence-based policy making, 3) involving and informing (expectant) vulnerable families. New opportunities for linking administrative data from across societal sectors for the entire Dutch population, combined with advances in data analytics, hold potential for supporting prevention in the first 1000 days. Such efforts should: - meet the needs and concerns of families, professionals, and policy makers, requiring stakeholder participation throughout; - actively reduce inequalities and stigmatisation; - be based on an understanding of vulnerability as a complex problem, requiring cross-sectoral action in which tailored interventions and precision public health policies are combined. This is exactly what we will do. Aims By combining stakeholder participation with data science methods applied to linked administrative data (e.g. Statistics Netherlands, Netherlands Perinatal Registry, Youth Health Care, Vektis), and by building on our extensive experience with vulnerable families in practice and policy making, we aim to: 1. develop guidance on the use of big data for practice & policy in the first 1000 days 2. develop a dynamic tool to support professionals across the care chain in the first 1000 days to identify vulnerable families and tailor interventions 3. develop a roadmap for tailoring municipal policy making for vulnerable families, and implement this in our ‘living lab’ Rotterdam Methods We will: 1. use complex systems mapping to visualize evidence on the web of factors leading to vulnerability in the first 1000 days to support meaningful identification and policy making, and consult on this map with stakeholders 2. conduct data dialogues with vulnerable families, practitioners and policy makers to define the conditions under which big data can be used from their perspective 3. use consultations with vulnerable families and practitioners to develop an overview of problems in current practice of identification and referral and what is needed to address these problems 4. develop prediction models for adverse outcomes across the life course based on risk factors in the first 1000 days, comparing traditional prediction modelling and machine learning (ML) techniques 5. develop vulnerability profiles using traditional and ML-based clustering techniques to support tailoring of interventions and policies 6. integrate our models in existing digital personal health environments as a dynamic tool for identification of vulnerable (expectant) parents and children across the care chain throughout the first 1000 days; this gives parents control over their data 7. develop municipal-level forecasts of the burden of adverse outcomes due to vulnerability in the first 1000 days to support policy making 8. develop a blueprint for a municipal level vulnerability dashboard, and implement this in Rotterdam 9. develop a roadmap for municipal tailoring of policy-making, using Rotterdam as living-lab, by comparing the complex systems map of vulnerability to the empirical and policy reality in Rotterdam, combining this with our newly generated evidence, to highlight entry points for policy tailoring, and integrate this in ongoing policy processes. Our consortium consists of professionals and organisations with long-standing experience in working with vulnerable (expectant) families and children in clinical practice, policy and research. In the Netherlands, municipalities coordinate services for children and families in need. Our consortium member City of Rotterdam, is at the forefront of identifying these families and providing tailored care. Our work will provide evidence on the added value, limitations, risks and conditions under which big data and analytics can be used for a promising start in the first 1000 days, guiding future work in this area. It will provide tools for professionals to provide more timely and tailored support to vulnerable (expectant) parents and children and to policy makers to tailor policies to needs. By addressing the needs and concerns of parents, practitioners, and policy makers, and by embedding our work in regional care networks and a municipal government living-lab, and supporting roll-out to other municipalities, our work provides a strong base for improving the first 1000 days of Dutch children.