Estimation of exposure prevalence in a population at risk using data from cases and an external estimate of the...

An estimate of exposure prevalence is often of interest, for example when targeting public health interventions. Unfortunately, this information may not be available from field studies. In this paper, I propose a method to estimate exposure prevalence when data are available only from a series of cases and an estimate of the relative risk for the exposure is available from published studies. The method is most useful when (1) the prevalence of exposure in the case series is a valid estimate of exposure prevalence among cases in the study base, and (2) the external estimate of relative risk is unbiased and can be generalized to the study base. I recommend that a sensitivity analysis be done when either condition 1 or 2 above is not met with certainty. I illustrate the method to estimate the prevalence of aspirin use in a population at risk for myocardial infarction using data from 1,228 cases and an estimate of the relative risk for the effect of aspirin use from a published meta-analysis.