Cytodyn Inc (CYDY)

[iddrisw]

Hi iddrisw, Boreal and I also shared some thoughts about the statistics, which you can read. His approach is different but also useful.
Borel is absolutely correct on using the TSS and its even better because the change in TSS is treated as a continuous outcome in meta analysis. By adding a baseline day0 score i.e. i used RANDBETWEEN(10, 12) both populations and same ideas. as long as the average change in TSS improvement is 1.2'ish or better, the stats are encouraging and dramatically better ie pvalue <0.0001 with improvement of 2 or better.

suspect that both methods lead to similar answers. One think to ask you is if you calculated endpoint differences the same way he did, which is based on the trial protocol and he described yesterday. In other words, did your random number generator provide ranges from negative to positive as Boreal stated or can it be modified if it didn’t? This would be helpful to know if both methods are consistent.

Borels chime in is the way to go. i did check the remdesivir study and they were using odd ratios which is another way to it but that is more of a logit model. try running the numbers yourself. With the TSS you need only use the t-test since TSS is treated as a continuous outcome. the whole exercise was just to get an idea of threshold improvement versus my expectations for LL. Even remdesivir was statistically significant at day14 outcome score for the secondary endpoint

Thanks