Clarification on purportedly “blinded” samples in Duke case
I wrote in my previous post about Joseph Nevins’ responsibility as primary investigator in the Duke-led work that has now been shown to include falsified data. In that post, I noted that Dr. Nevins said he continued to believe in the integrity of the work in large part because it had been confirmed using an independent and supposedly blinded set of samples. I also pointed out that one IOM committee member mentioned during the meeting this week that not only were they not blinded but that the all-important patient response data was included with the sample tubes.
I mentioned that last point to highlight the extent of Dr. Nevins’ lack of oversight; if he had looked even cursorily, he would have seen the samples were not blinded at all.
One of my readers subsequently pointed out to me, in a private email, that the phrasing made it sound like this was the first time it became clear that the samples were not blinded. That is not the case, and I apologize for any confusion. In fact, The Cancer Letter reported as early as October 2009 that the validation set was not blinded.
That doesn’t make Dr. Nevins argument any more defensible.