The Limits of Generalizability in a Target Trial

Matt, Doug Russell has congratulated you on a beautiful trial and commented on a nice clear separation between groups. But I suppose one of the questions that has popped up is you didn't quite get the separation between the low, medium and high oxygen levels. Matt: We should be designing trials to emulate the way that those interventions are delivered in clinical care rather than figure out statistical pieces to match that not the other way around. And then the last thing I'll say is that unlike a trial in which you have the treatment A versus treatment B that patients just receive once and are done with, understanding what you should expect about separation between groups in a target trial is a really interesting challenge. It's true