>more likely to be prescribed to patients with more severe disease
Yes, this is the money line in this study. The main thing the study "shows" is that their doctors were giving HCQ as a last resort treatment.
This is also a retrospective study, meaning that they reviewed patient files to find who did and who didn't get HCQ and saw what happened. It is not a controlled trial and conclusions about HCQ are impossible.
There were no inclusion/exclusion criteria beyond having the virus and certain clinical assessments.
What it shows is what they did didn't work, and that they were using HCQ wrong in light of the other evidence being collected.