Risk estimation in retrospective studies

Abstract

The analysis of data arising from retrospective studies traditionally followed methods<br>involving estimation of relative risks and/or odds ratios. Little attention was given to<br>estimation of risks, presumably due to constraint of such experimental designs, as well<br>as the ease of modelling of odds ratio by logistic regression. Here we present some<br>results for the estimation of risks in a general 2 by k contingency table setting, for a<br>dichotomous outcome variable and under some reasonable assumption of prevalence.<br>We also examine the properties of proposed estimators, and apply them to a largescale<br>genome-wide association (GWA) study data to demonstrate some relevance of the<br>methods.