We’ve got a plan for that: Methods to avoid oversimplification and address bias and validity in epidemiologic research