Advanced Methods Workshop

Causal inference is challenging with real-world data.  In many pediatric and perinatal epidemiologic studies, the first challenge is articulating a meaningful causal question.  In this workshop, we will briefly introduce attendees to the process (and challenges) of articulating a causal question and how this leads to a choice of statistical analysis.  We focus on settings where the question is about causal effects of interventions on a treatment or exposure that varies over time and review some statistical methods that may recover such an effect in longitudinal studies with time-varying confounding.  We illustrate an application of these ideas to data from Project Viva, a longitudinal pre-birth cohort study.