I prescribe medications off-label all the time. But do this when I've seen research to indicate it likely will produce the desired benefit with minimal or at least tolerable side-effects. Many drugs have plenty of good data for off-label use, but they simply don't have the official indication because no drug company has ever bothered to undertake the considerable expense and effort for the approval.
But with hydroxychloriquine, we have such overwhelming evidence from multicenter randomized, double-blinded studies that it does not improve the studied outcomes in COVID-19. I would seriously question the intentions (or clinical competence) of any medical provider who still hypes it. In this case, I'm trusting the openly-published and freely available outcome data on hydroxychloroquine rather than a vague suspicion that there is some kind of behind-the-scenes malfeasance.
I and probably every other medical provider would be thrilled if we were proven wrong on hydroxychloroquine if some newer, and larger studies actually showed some benefit, although we might need to make a little red-faced mea culpa. Nothing wrong with that, no medical provider is going to make perfect predictions every time. One of the things we learn in training is how to admit and grow from our mistakes.