If anyone has the data, I propose we use the % of deaths in eIND as the worst case % of deaths in LL arm. These were the sickest of sick and could be used to make an educated conservative guess. The balance of deaths can be attributed to placebo. What p value would these assumptions make?