It's too soon. Legitimate, blind randomized tests take years. Don't forget, they have to also be peer reviewed by an independent tester repeating the test and getting the same results before the medical community accepts it as "scientifically proven".
That's not to say there isn't documented evidence (above and beyond anecdotal evidence). For instance, a study in France (by IHU-Méditerranée Infection in Marseille) screened 31,000~ patients for SARS-CoV-2 (COVID-19) in a clinical setting. 3,737 patients tested positive for COVID-19, 3,119 (83.5%) of which were treated with hydroxychloroquine. The results:
"The French team reported that overall the treatment regimen of HCQ-AZ led to a decreased risk of transfer to the intensive care unit (ICU) as well as death (Hazard ratio 0.18 0.11-0.27), a decreased risk of hospitalization ≥ 10 days (odds rations 95% CI 0.38 0.27-0.54 (and shorted duration of viral shedding (time to negative PCR: HR 1.29 1.17-1.42). They did observe OTc prolongation (>60 ms ) in 25 patients (0.67%) which led to a halting of that treatment in 12 cases including 3 cases with QTc> 500 ms. The team observed absolutely no cases of sudden death."
Also:
"The team can make the claim based on the observations, or retrospective analysis that It would appear that if patients diagnosed with SARS-CoV-2 are diagnosed, isolated and treated early and administered at least 3 days of HCQ-AZ that this led to significantly improved clinical outcomes as well as expedited viral load reduction when compared to other treatments."
The study was performed in a clinical setting treating with real patients, not in a testing laboratory. The link is provided below: