2 ms·

But this would "just" reduce the power of the study and not increase bias, right? It's effectively turning it into an intention-to-treat design.