I don't work in this field, but the authors seem to be suggesting that hydroxychloroquine makes some things significantly worse. Patients who took hydroxychloroquine were much more likely to "need" supplementary oxygen. However, the two groups of patients were in two different hospitals. The paper doesn't discuss whether each hospital used exactly the same definition of "need" and had staff policies/practices that assessed this need with exactly the same amount of diligence.