[Risk and odds--how to calculate with events]

Abstract

Categorical scientific data consisting of counts of events are frequently reported as risk or odds figures. Given a certain set of data, odds always differs upwards from risk. The relation between two risk figures may be expressed in various ways, one of which is the risk ratio. In the case of two odds figures, the choice is almost always the odds ratio. With a certain set of data from two groups, the odds ratio is not identical to the risk ratio (except when both are equal to 1). The odds ratio always magnifies the intergroup difference. When assessing published data, one must take care to observe whether reported ratio figures denote risk ratios or odds ratios.