One option open to researchers who would like to increase power in the context of limited case series is just to increase the control collection. This strategy might include using cases for one disease as extra controls for another (assuming suitably different disease aetiologies and similar population history). We investigated the utility of such an approach by performing simulations with 1000 cases and an increasing number of control (Figure S8). Although the gains are not as strong as increasing both the case and control sample sizes (Figure 2), the ability to reject the null hypothesis of no association increase considerably with the size of the control panel. For example, adding an extra 2000 controls to a case-control study with sample size 1000–1000 increases power to detect an effect of 1.5 typically by 20%. Subject to care in their use, the growing availability of genotyped sets of controls promises to make this a possibility worth investigating for many studies.
