The point is that you don't need to show that the therapeutic effect is the same, you need to show that you expect the therapeutic effect to be the same. The trials required to prove the latter are orders of magnitude simpler, faster and less expensive.