No. The FDA does not require anything other than studies to show the drug is safe when used as directed, and that it meets markers showing its primary target.
For example, an Alzheimer drug has been declared effective because it changes a metabolite thought to be a marker of disease severity, and the clinical trials showed that it does not have significant side-effects.
One of the problems we have had with the FDA is that our focus has been on recovery - cure - an end-point which is not in the FDA play-book. They would prefer us to be claiming some change in an inflammatory marker - like CRP, for example...
Even though there is a newer department for that detail, the FDA still doesn't do much more than put molecular models in the file. That's likely a big reason why only 20% of drugs have PDB structures available, aside from prehistoric timelines a particular drug was introduced.
A lot of regulatory changes are expected with the Critical Path initiative, but I think it is still fair to guestimate government regulatory agencies are going to be years behind the ever growing numbers of researchers and the technology explosion curve on which they operate.
I appreciated his comment that PubMed abstracts were a bit difficult to capture the details of any particular study that could apply to literature search. Sad, since literature search has been the backbone of a lot of science for a very long while. On the upside, at least we have abstracts and keyword search now.