"Randomization" is ONLY necessary if you are "measuring" clinical symptoms in order to avoid the "placebo error". The French study directly measured viral loads by PCR (pollymerase-chain-reaction), not clinical symptoms. All that is necessary in such a case is that a control group be included, as was done. PCR is both exquisitely sensitive and exquisitely specific.
I'll take the French data over the Chinese anytime.
Hear here! Wisdon is around the place!