New analytical approaches to life course epidemiology

Abstract Recent decades have seen huge mega cohorts set up to collect complex, high-dimensional research data. These are often pooled with other studies or linked with routine data. The present chapter addresses the ensuing data analytical challenges for the field of life course epidemiology: First, multi-cohort designs are presented as efficient designs to extend the observation period or obtain more precise estimates. Practical issues such as data access and harmonization as well as approaches for the analysis of multi-cohort data are discussed. With complex data, the research questions also become more complex. The second part therefore introduces causal inference for life course epidemiology. The topics of causal effect estimation, causal mediation analyses, and causal discovery along with examples from life course epidemiology are covered. The chapter concludes with a discussion of machine learning approaches as promising tools for tackling the complexities of high-dimensional cohort data in future life course studies.