Unfortunately, it's a retrospective cohort study with no word on how the patients were divided into treatment categories by their doctors. Presumably, some of the patients weren't as sick as others. It would have been interesting to know how many of the ventilated patients, alone, did better on HCQ + Z (which was the most effective treatment but of course, I wouldn't think anyone without pneumonia would get Z - so that's already a confounding variable).
Interestingly and of note is that some of the patients also got a steroid, but this was not disaggregated in the study and several reviewers of this article have said that it will be difficult in future to withhold steroids from patient groups who are getting any other regiment, including HCQ.
No one with heart irregularities was put on HCQ + Z (and heart irregularities are a symptom of more severe CoVid). Some severely ill patients with mild cardiac irregularities did get HCQ by itself. No one over 76 was in the study. The median age of the HCQ only patients was younger than the median age of the No HCQ patients, presumably because their doctors worried about giving older patients with mildly abnormal ECG's the drug. There were also more males in the No HCQ group, which is itself associated with greater mortality.
The study itself says that more study is needed (blind). In short, this study shows that nothing bad happened to less sick, slightly younger, non-cardiac patients when they took HCQ - and they got better at a higher rate.
Since we know from other studies that people with cardiac arrhythmias (both prior to CoVid and first appearing during CoVId) don't do so well, that needs to be disentangled in future study.
The double blind studies do not seem to show the same advantage for HCQ, which is of course the kind of study that actually guides its use.