You give the control arm the current best standard of care until and unless it becomes statistically obvious that your test arm is better and then ethically you're supposed to switch everyone to the drug at that point. It's not like the control group is denied the best care we know how to provide, just to accommodate experimentation.