Another good comment on clinical trials:
It's not necessarily true that "efficacy from clinical trials is a worst case scenario." Study participants can differ from patients in the general population in ways that favor research subjects with regard to outcomes. Some of these differences are medical, such as patients having coexisting medical conditions that were exclusion criteria in the study, thereby making treatment outside that setting potentially more dangerous. Others have to do with the fact that study participants are selected for factors other than their specific medical situation that enhance the likelihood they will complete the study. This includes everything from literacy to having the means to get to and from study visits.
The preious commenter was speaking of cancer trials, where there may be fewer exclusion criteria. But it's still a good point.