ReallyP - you said " If a mild goes from say 3 to 1 in the Leronlimab group, but a moderate goes from 7 to 3 in the placebo group, what can we conclude?
First of all, one cannot conclude anything from that. You use the term "a" above, meaning one particular patient. However, while that certainly can (and likely will be) the case for a given patient, since the trial's statistics will be aggregates, I just cannot believe that the placebo group's score will go from a 7 to a 3. Yes, some of the individual placebo patients' scores will improve, but more are likely to get worse - meaning that for the placebo group as a whole we're more likely to see an aggregate score of something like going from a 7 to an 8 or 9. So, if the Leron group goes from a 7 (i.e., assuming homogeneity of all 84 patients at day 0), to a 2 then we should have a statistically significant difference at the 95% confidence level or higher! Now, that, you can conclude....