[Effect measures and confidence intervals].

Abstract

In order to exemplify the use of effect measures and confidence intervals in epidemiology, it is assumed that a group of patients treated with antiflogistic drugs were compared by repeated gastroscopy with a control group. The effect measures "incidence rate difference" and "incidence rate ratio" are used to compare ulcer incidence in the two groups of patients, and the corresponding confidence intervals are stated. On the other hand, if the registered parameter is prevalence rather than incidence rate, for instance the occurrence of prepyloric erosions at a certain time, the effect measures employed are "risk difference" and "risk ratio". An alternative to the latter is "odds ratio".