In the Netherlands, approximately 1 million children (0-25 years) have a chronic disease. Above and beyond the ever-present challenges of growing up with an illness, these children have 40% chance to develop psychological problems, including depression, anxiety and loneliness. Throughout their life, this translates into decreased well-being and reduced social participation and generates additional costs for society. Early prevention of psychological problems is thus key to break this vicious cycle. Therefore, eHealth applications are promising. However, scientific knowledge is missing and validated tools are not yet available for this group and involved health care professionals. Our mission is to make scientifically validated eHealth tools that allow personalized and trans-diagnostic prevention of psychological problems widely available for this highly vulnerable group of chronically ill children and future adults, through an accessible, user-friendly, safe, and sustainable platform. To succeed in this mission, we present an iterative learning cycle approach in two four-year phases during which we gather the insights, and develop, evaluate, and implement the much needed eHealth tools: I. Development: Distil and validate the theoretical and game-design factors that make eHealth effective for chronically ill children. II. Evaluation: Evaluate trans-diagnostic and personalized eHealth tools for chronically ill children, using and developing state-of-the-art methods. III. Implementation: Study and remove the barriers that currently hinder implementation and uptake, and threaten availability of eHealth applications for chronically ill children. Our eHealth junior consortium includes (applied) researchers, pediatricians, psychiatrists, psychologists, patient organizations, knowledge centers, game designers, industrial designers, insurance companies, and business professionals. We will collaborate with the end-users (children, families, and professionals) in order to achieve both international scientific breakthroughs and optimal clinical and societal impact. Knowledge utilization is a crucial part of our project.

In the tsunami of new scientific knowledge, imagine developing guidelines for interventions, treatments, or working on any scientifically based study: there’s not enough time to read everything. Solutions from the field of artificial intelligence are typically closed-source, which is problematic in the era of transparency, especially if AI is making decisions. Moreover, most literature searches are biased toward western studies published in English. Therefore, the current project proposes to develop a fully open-source and real-time AI-aided pipeline using inclusive databases.

The aim of the research in this proposal is to extend the range of conclusions that can be drawn based on meta-analyses. The combination of meta-analysis and structural equation modeling with the goal of analysing structural models (MASEM) is a new and promising field of methodological research. Using MASEM, information from multiple studies can be used to test a single model that explains the relationships between a set of variables or to compare several models that are supported by different studies or theories. The state of the art approach to conduct MASEM is the two-stage approach of Cheung and Chan (2005a). The innovative feature of MASEM is that it facilitates the meta-analysis of complete models, instead of pooling separate effect sizes representing separate bivariate relations, which is done in ordinary meta-analysis. The research in this proposal aims to develop new methods to 1) model heterogeneity (differences in effect sizes) between studies in MASEM, and to 2) analyse mean structures across studies with MASEM. The two proposed methods will be evaluated using simulated data. Moreover, the research will lead to clear guidelines on how to apply the methods, so that the new methods will actually be used by substantive researchers. Keywords: Meta-analysis, research synthesis, structural equation modeling, heterogeneity, means.

Scientific research uses statistical models to describe and understand real-world social and behavioral phenomena. As more data is being collected, these statistical models become larger and more complex. I will develop novel methods so that researchers can use these increasingly large and complicated statistical models to better understand the world.

Synthetic data is a dataset with (more or less) the same properties as an original dataset but without privacy-sensitive information. By making synthetic data available instead of (or prior to) the actual dataset, scientists gain faster and easier access to confidential data. In this project, two tools for creating synthetic data are used to unlock existing datasets, including datasets archived at DANS.