Readers should systematically assess methods used to identify, measure and analyze confounding in observational cohort studies.