Speaking of getting things right, read this quote from Mr. Castello:
<<Nonetheless, we believe that we are now close to reaching the specified number, allowing us to conclude accrual with the statistical power called for in our trial design.>>
Does anyone else remember last Fall when Xoma announced the DSMB had continued the trial (hard to keep track, there have been so many, but I think that was the second time) and Xoma dropped a little statement about the DSMB planning to meet again in December, among other things to consider a "redesign" of the P-3 trial?
There were some who asked at the time what on earth could that mean? Now I see Mr. Castello or his ghost writer (our friend RobertK1, perhaps) seems to be using the same phrase, "trial design" to refer to a specific "statistical power." Do you suppose that in addition to the total body count Xoma agreed in advance to some sort of minimum statistical ratio for Neuprex to be determined a success? Would such a thing be common? Ever used? Sane? If so, what could such a ratio be??
The reason I ask is that I, for one, have been assuming that ANY ratio -- whatever it might be -- which was enough to exclude chance as a probable explanation for favorable treatment results would be adequate to declare victory and withdraw the troops, so to speak. But what if Xoma somehow bound itself to a specific ratio SUBSTANTIALLY GREATER than "enough"? That, indeed, would have provided a powerful incentive to talk about "redesigning" the P-3 trial at such a late stage as last December -- especially if the company suspected Neuprex was showing statistically "adequate" efficacy but not "enough."
In other words, did Xoma agree to a tougher-than-normal statistical significance in order to have a lower-than-usual patient enrollment? |
