The Trade-Off Between Cluster Crossover and Randomized Trials

A major strength of a cluster crossover trial is that essentially it eliminates selection bias. For oxygen, a ubiquitous therapy for mechanically ventilated patients, you're studying all the patients who receive the intervention in the real world. It's still possible that there are non-random imbalances in patient characteristics between periods that are unmeasured and therefore not possible to account for. So I don't think that's a major problem, but it is a thing to bear in mind and consider when you look at these data in this trial.