The results of well-designed observational studies (with either a cohort or a case-control design) do not systematically overestimate the magnitude of the effects of treatment as compared with those in randomized, controlled trials on the same topic.
No takes yet. Share an insight, caveat, or question.
Concato et al. (2000) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: