Replies to post #4648 on Biotech Values

[DewDiligence]

Re QLTI’s p-values:

>> when i was in the business, used the combined populations of the 4 subsets. so if there were 25 measures each i needed a .05 p for 100 measures. this is much different than .0125 for each 25. how can we reconcile this? <<

What QLTI needs for a successful trial outcome in the VIO trial is to meet one or more of the four endpoints with p<0.0125. They do not have to meet all of the endpoints.

Using your lab example of four subsets of 25 measurements each, an a priori requirement that one or more subset show p<0.0125 is roughly as demanding as an a priori requirement that the entire set of 100 measurements show p<0.05.