p value tells you whether the clinical results are most likely true or not. So they are important and the FDA of course is interested in whether results are true or by chance. What would counter that to some extent is the clinical results backing up the MOA. If test results (ie. IL-6, TFNa, viral load) correlate with the MOA and the clinical results the FDA will take that into consideration.

There is a lack of solid MOA in many of the drugs being tested for COVID-19. For instance in RLF-100 they've hypothesized a very broad MOA. They have shown no test results to support it which is why I've withheld judgement on it. Others like the ones from Sorrento are in vitro only which is fairly useless.

The one thing that has befuddled the FDA from the time that we filed an IND for COVID-19 is the complexity of interactions that lead to our MOAs. Blocking CCR5 for HIV is a straightforward proposition. Once you get into the cascading effects of blocking chemokines it gets very complicated very quickly. So complicated that no one so far has been able to tell me whether my "leronlimab downregulating the mTOR/4E-BP1/elF4E pathway acts as an antiviral" is most likely correct..