Relative risk for genetic associations: the case-parent triad as a variant of case-cohort design.

The contribution of this paper is to conceptualize the case-parent triad within an epidemiological framework. We propose that the case-parent triad design is a variant of the case-cohort design. The affected offspring of case-parent triads come from a source cohort of all offspring of parents in a population. We first demonstrate that if the source cohort is restricted to offspring of a certain parental mating type then the relative risk in relation to genetic exposure can be estimated simply from the ratio of the number of exposed to the number of unexposed affected offspring. We then extend the logic to studies including offspring of all parental mating types; provided that the allele frequencies and possible parental mating types are specified, a valid relative risk can still be estimated. Compared to prior descriptions of the case-parent triad design, the proposed approach is readily understandable, epidemiologically meaningful and provides a relatively simple perspective for estimating valid measure of effect. Also, by allowing the potential sources of selection bias to be revealed more easily the design is made more accessible both conceptually and practically to epidemiologists.