An example of what I did for a project was that we know patients with AF need to be anticoagulated and the best thing at the time was warfarin so you compare warfarin to dabigatron instead of comparing dabigatron to nothing. Your control changes to the best current treatment