Gravar-mail: Selecting controls in a case–control study