Controls were in the Interphone study recruited from general practioners' lists (UK), whereas all controls in the Hardell et al studies were selected from the population register. Recall and ...