In the trials registry [1] this was registered as a three arm trial with 100 patients each, so 300 overall. Yet the paper mentions ony two groups with 31 patients each. That alone looks very dubious...
It's not necessarily dubious to change a protocol, but it is concerning that they didn't report on the changes along with their reasoning.
It would make sense, if they couldn't get 300 research subjects, that there would be a change of protocol. Also, I think it could have been a good idea to switch from viral load and t cell recovery time to something more meaningful and quick to measure like TTCR and pulmonary condition, but again, they don't explain the change so for all we know the change was to get their names on a more exciting paper.
I see no reason why they would drop the placebo either.
We really do have to wait for the bigger and better trials.
In case you are interested, here is a larger study from in Wuhan that showed lower time to recovery on hydroxychloroquine: https://www.medrxiv.org/content/10.1101/2020.03.22.20040758v...