I apologize, I looked at the wrong rows when analyzing the data, please disregard that post. Correcting:
early treatment shows consistent 50% improvement but with very small sample sizes: 3 out of 7 studies that had >100 people (100 to 300) in each of the arms have 12%, 16%, 37% improvement.
So, it is some modest improvement with not a lot of data. Understand that for early treatments, we are talking low mortality, so 100 patients = 4 deaths, the numbers are really low.
Prophylaxis has 5 studies with 400+ people in each arm with 30% improvement
late treatment: much more data, 17% improvement on average, but all big trials negatives
so, overall seems that HCQ is somewhat helpful and should be used early, maybe as prophylaxis upon exposure