So why would such inappropriate patients not be excluded? It would be the job of the IDMC to suggest that such patients should not be retained in this trial and that replacements should be enrolled.
With a midpoint enrollment of 63 months for the 795 patients, a consequent fatality rate of about 60% and a completely ineffective multikine treatment, we would have had about 480 fatalities before data lock. I grant you that with about a 10% dropout (LTFU) at various times during the trial, the number of fatalities would have been somewhat lower than 480. It is However difficult to account for about (give and take) 182 patients (480-298=182) who beat the odds. They could not all had the less dangerous base of the tongue cancer.
If so many inappropriate patients were enrolled and with other potential irregularities, where was the IDMC during the last 5+ years? Also does CVM have a Scientific Advisory Board (SAB) or any scientific advisors?
Of the 135 enrollments from year 2011 thru 2013, many of us place a bulk of that 135 into year 2013.
It's not really possible. CVM 2016 10K (pg 5) has a chart for monthly enrollments, Mar - Dec 2013 = 19.
To have 60 to 80 of the 135 enrolled in 2013 would mean the ousted CRO would have enrolled 25 a month in Jan and Feb. Couldn't cancel them for non performance, if that were the case.
Therefore there are 60-80 people who have been enrolled a year longer than we are figuring.
So something is quite out of whack.
It's my surmise that Geert took the action to enroll more than original design, the additional 48 - to up the number of trial participants who are within design protocol.
Which leads to another issue - increase the participants by 48 but the end point of 298 deaths remained unchanged.
OS of around 55% at 3 years and ADD 48 more participants in the study!
298 would of necessity be achieved quicker with more participants.
Under original statistical design the trial could still be waiting for the 298th event to start the Data Gathering in order to achieve Data Lock.