Zoloft. You?

It appears that patients are buying millions and billions of dollars worth of medication which barely helps with the ailment it is supposed to target.

A study collating the outcomes of clinical trials of anti-depressants has shown that common kinds of antidepressants don't have adequate effect (in excess of placebo effect) to warrant them being prescribed to most patients.

The drug companies are unhappy with this report. An expert rebuts the study, saying that it cannot be regarded as accurate because of the way the drug-trials are conducted.

The expert says that drug trials are performed on patients who have been carefully selected to be relatively healthy and have only mild cases of depression. This vetting, he says, means that the trial data which has been used in the report is not representative of a true cross section of patients, so cannot be used to make assessments about the (in)effectiveness of the drug in the field.

I have to ask - if the trial data is good enough to cause the FDA to put the drug on the market, surely the trial data is good enough to be used to examine whether the drug works? The expert cannot have it both ways. As I see it, his rebuttal is less of a criticism of the study and more an indictment of the way clinical trial data is used to get pharmaceutical products onto the market.