Causal inference in nutritional epidemiology
13 juni kl. 13:00
SVEP and NEON (the Swedish Network in Nutritional Epidemiology)
“Causal inference in nutritional epidemiology – optimising the use of observational data to strengthen the evidence between diet and health”
Preliminary speakers include Daniel Ibsen, Århus/Copenhagen University, Georgia Tomova, Leeds University and Anton Nilsson, Lund University
Randomized controlled studies of long-term dietary adherence and disease incidence are most often not feasible and sometimes unethical to conduct. Disentangling causal effects from associations in nutritional research is therefore a difficult task and most often, the field relies on the use of observational data to answer such questions. However, nutritional epidemiology faces many challenges such as confounding bias, selection bias, and measurement error from self-reported dietary intake. Are we stuck with only associations or can we use causal inference methods to better separate pure associations from actual causal effects?
In this seminar, we will learn how to; identify biases using graphical tools, critically think about the concept of adjusting for total energy intake and what comparator nutrients/foods we are contrasting, as well as how to analyze observational data as randomized controlled studies.