This paper shows that case-control studies can produce unbiased estimates of relative risk and attributable fraction without assuming that the studied disease is rare. It also discusses the still current myth that it is necessary to assume that a disease is rare in order to assure that odds ratio estimates are closer to relative risk measures in case control studies, showing that it comes from the beginning of the methodological development of case-control studies, in the mid 20th Century.
Case control study; Estimability; Relative risk; Odds ratio; Attributable fraction
