Replies to post #4638 on Biotech Values

[DewDiligence]

Re: QLTI and p-value slices:

>> I dont think the p values associate like that. a p value of .0125 for a particular population is much different from a p value of .05 for a much larger population. it is not just .0125 for each population times 4 <<

brainlessone: It really is as simple as dividing the aggregate 0.05 into four slices of 0.0125 each. QLTI’s CMO, Mohammad Azab, laid out the details during the Q&A session of QLTI’s most recent CC:

http://biz.yahoo.com/cc/2/48162.html

The above does not imply that it’s mathematically rigorous to divide p-values in this manner. However, p-value thresholds are themselves inherently arbitrary. The way the FDA divides an aggregate 0.05 into pieces when there are multiple doses and/or multiple endpoints is as reasonable an approach as any.