iwfal:
The CEGE CEO also staed at their CC, that Vital 1 also used the conservative O'Brien Fleming rules to allocate its interim alpha p value. A nonstatistician googling that would still be somewhat confused, except to confirm that it is considered to be one of the most conservative approaches;
“O'Brien-Fleming's method, corresponding delta = 0, is very conservative in early rejection of the null hypothesis. Pocock's method, corresponding delta = 0.5, spends Type I error uniformly across different stages. Generally speaking, a large delta (e.g. 0.8) will lead to a design that spends Type I error more at earlier stages than later stages”
If you assume that the number of events required for the interim was 200 (out of a total planned 600, and total actual enrollees of 626), what would your guess be as to the actual p value that CEGE allocated for the interim. TIA.
