The Disadvantages of Randomised Trials

The criticism that people will quite often raise is that you lose generalizability for the result. Even without a run-in, trial participants tend to be healthier. So what are the disadvantages of this approach? If this increases our ability to preserve our power in the trial and get a reliable or robust answer, why wouldn't you always do this? Well, it's really mainly relevant to longer-term trials. Trials where you need to get your treatment to the patient very rapidly, acute disease, for example, isn't relevant.