The odds ratio: some considerations

Abstract

Over the last few years a growing number of retrospective and controlled cross-sectional studies using the odds ratio as a measure of intensity of the association have been published. The objectives of this article are: to compare the odds ratio with the prevalence ratio; to study the behavior of this measure with the sampling variation of the prevalence of the risk factor either in cases or in controls; and to give relevance to the expression of it in terms of confidence interval.