Patients in this study are unlikely to be switched for efficacy reasons. ... You have to understand that no physician would take a patient who was doing poorly on atorvastatin and switch them to simvastatin in the hopes that they would respond better to simvastatin.
On this I significantly disagree. Even with my small data set I have personally seen it happen several times in different classes of drugs - when there were no other classes anywhere near as potent.
I'd be willing to bet that the *vast* majority were likely to have been switched because of a cost issue--even before it went generic, simva was significantly less expensive than Lipitor.
Not substantially/uniformly true - at least in the US over the last 5 years.
In any case, there is no way to design a randomized trial to test this. You have to rely on real-world data. This is the best you're going to get
Sometimes the best isn't good enough - e.g. that difference in drop out rate speaks volumes.