Power and replication in case-control studies

Abstract The case-control study design, a common staple of epidemiology, is increasingly used to test for genetic association. The simplicity of the design accounts for both its appeal and its limitations. Too often, however, apparent controversy arises for lack of appreciation of basic tenets underlying statistical testing. Power and replication are two concepts most commonly ignored in evaluating such studies. We review the basic principles of statistical testing, recall simple means to calculate power, and provide numerical examples pertaining to the association between angiotensinogen and essential hypertension.