>> Looking just at deaths, 2.8% of the patients who took hydroxychloroquine died, and 4.6% of the patients who did not take it died. That difference was also not found to be statistically significant. <<
So, you design a study with an impossibly small sample size, and when the HCQ cuts the death rate by 40%, you claim that the reduction is “statistically insignificant.”
Difference between 3 and 4 deaths.