Why are we not comparing infection and death rates from people who take hydroxychloroquine long-term (e.g. lupus patients) against non-takers of this drug with the same overall demographics to determine efficacy?