> I have no explanation for this. One possibility that RCTs are designed to eliminate is "regression to the mean." If the natural course of disease is to wax

> I have no explanation for this. One possibility that RCTs are designed to eliminate is "regression to the mean." If the natural course of disease is to wax and wane and you intervene whenever the disease is waxing it can seem like your intervention is effective even when it has no specific effect. In addition, placebos produce a small effect even when you know you are taking a placebo.