Actually the website says that the stats apply to those receiving hydroxychloroquine with or without (+-) azithromycin. It is possible the single death occurred with someone who was not receiving the azithromycin.
It’s not randomized, but looks like they had 1577 positive tests. Of those positives, 701 received at least 3 days of hydroxychloroquine±azithromycin and so far had only 1 death. Presumably the other 876 had none, or at most 2 days, of that with 10 deaths in that group. There’s not a shortage of patients so it’s reasonable to try to recruit volunteers to provide formal official proof it works after the fact. For those not wanting to participate in a trial the uncontrolled 8x lower death rate may drive them to take the treatment.
Good catch.