Please Note: Blog posts are not selected, edited or screened by Seeking Alpha editors.

Statistical Power and GNBT

|Includes: Generex Biotechnology Corp. (GNBT)

I thought maybe I would take the time to comment on Generex Biotechnology Corp's (OTCQB:GNBT) announcement today that their enrollment total is now over 400 subjects for their Phase III global clinical trial of Oral-lyn (NYSE:TM) before Adam's (with Derek's rehash) comments reached the newswire.

As many already know, prior to conducting a clinical trial, a hypothesis is proposed,  Often this hypothesis is conducted in the 'null' approach.  That is no difference will be detected between two methods or approaches to the same problem.  In the case for Oral-Lyn the primary outcome is:

"To compare the efficacy of Generex Oral-lyn™ RapidMist™ System and standard regular human insulin therapy as measured by HbA1c, in type-1 diabetes mellitus subjects on BID NPH intermediate acting insulin therapy [ Time Frame: change in HbA1c from the Day 0 (V3) to Day 180 (V9) of Treatment Phase. ] [ Designated as safety issue: No ]"

which effectively translates to a null hypothesis that Oral-lyn will have no significant impact on the HbA1c of a subject over the course of a 6-month period, when compared to subjects who on BID NPH intermediate acting insulin therapy.

Of course there needs to be a statistical parameter or significance that is assigned.  Usually the criteria is a probability which typically ranges from 0.05 (1 in 20) to 0.001 (1 in 1000).  Unfortunately the information at does not indicate which probability GNBT is working towards, but my guess would be 0.01 (1 in 100).

Based on the GNBT statement released on March 10, 2009, it should come as no surprise that the number of subjects has decreased, given that the first 60 subjects did not show inferiority to the standard BID NPH therapy at a rate that warrants a larger data set.  Of course the assumption here is that the statisticians did it right in March and that the Company talked with the FDA regarding their power analysis. 

So where do we stand now:

1.  The original clinical trial called for an estimated 750 subjects
2.  Early/initial study data indicate that fewer subjects may be necessary
3.  Currently 400 subjects are enrolled
4.  An interim analysis will be conducted in July (per
5.  Many sites have already completed enrollment and/or studies (per
6.  Some sites have not started recruiting, but may start if results in July are not as expected.

I have to admit that I am looking forward to July as I think this will make or break the company.  I think that Adam will have a piece coming out in about two weeks, just before Seaside 88 makes their next purchase of stock with Derek summarizing the article the next day.

Then of course, I could be completely wrong, which wouldn't be the first time, and hopefully the last time, for me.


Disclosure: Long GNBT

Disclosure: Long GNBT