And where is the clinical evidence to support the claim? What you've posted is a multiple choice opinion poll and we don't even know what the other choices were. Or how the questions were asked.
So far there have been 2 tiny medical studies on the effectiveness of hydroxychloroquine against covid-19. The first study indicated that it was sometimes useful, but the study was highly flawed in multiple ways. The second study indicated that it was no more useful than nothing. Both studies were so small that they're not statistically significant. The sample size is simply too small. They're both very hasty initial looks at it to see if maybe it's worth looking at it properly.
EDIT: I found the complete report of the opinion poll you linked to. It's not how it's being portrayed. Doctors were asked their belief (not knowledge, not evidence, just belief) about which from a list of 15 possible treatments was the most effective therapy for covid-19. I stress again this was an opinion poll about beliefs. Any claim that it's evidence of efficiacy of treatment is wrong. But even the results of the opinion poll are misleadingly presented.
The poll offered 15 possible treatments to doctors and asked them to choose what they believed to be the most effective. Just belief. Not evidence.
Two-thirds of the doctors refused to answer that question.
I think that's important. They knew the question was wrong.
From the minority of doctors who answered the question, the number of answers was 2.18 times the number of doctors. So most, maybe even all, of them were suggesting a combination of options.
So the opinion poll returns the result that 37% (of the 35% of doctors who answered the question) chose hydroxychloroquine as what they believed to be either the most effective option from the list of 15 or (far more likely) one of the most effective options, used in combination with something else.
37%? Not a bad score in a poll. But not anything like what the headline you quoted claimed. Which is probably another one of the reasons why 65% of the doctors who agreed to fill out this opinion poll refused to answer that question - they knew it would be used in a misleading way.
But anyway, 37% is more than 0% It's something. It's not nothing.
But take a look at what came in 2nd in this poll. It is nothing. 32% of the doctors polled expressed the belief that nothing was the most effective therapy. So no therapy at all was rated almost as highly as hydroxychloroquine.
31% of them expressed the belief that the most effective therapy from that list was acetaminophen aka paracetamol.
So you've put forward an opinion poll as if it was proper evidence, you chosen a question that was so obviously bad that 2/3rds of the doctors who answered the other questions refused to answer that one and you've misrepresented the results anyway. Poor show.