Schering-Plough data manipulation?
Letter from reader:
I thought you might be interested in two potentially major problems with the recently released Schering Plough boceprevir data for the Sprint 1 trial.
1. Slide 12 of Dr. Kwo’s presentation of the boceprevir results at EASL 2008 includes some significant detail on the discontinuations in the two 28 week arms.
The two 28 week boceprevir arms average 27% total dropouts. (Unlike the Schering press release of August 4 that only identifies dropouts from adverse events; Dr. Kwo lists dropouts from adverse events, viral breakthrough and other or lost to follow-up, non-compliant, etc.) It is entirely reasonable to expect the 48 week arms to have a higher or at least equivalent dropout rates. Yet the 48 week arms showed 74% and 66% SVR12. If you simply added the 28 week drop out average of 27% you would get 101% and 93%. Beyond the 101% being mathematically impossible, you have no or very little room for the patients who finished treatment but did not SVR. And one would expect 15 – 20% finish but don’t have a SVR. I can only come up with three explanations. One, the 27% average dropout rate Dr. Kwo presented is wrong. This does though come from a fairly detailed slide presented at a major medical conference. Two, the 48 week boceprevir have a far, far lower dropout rate than the 28 week arms. This is not very likely. Three, there is something seriously wrong in the data presented in the Schering press release from August 4.
2. The Schering press release of August 4 mysteriously also leaves out 16 patients from the no lead-in treatment arms. Both the abstract from EASL and Dr. Kwo’s slideshow show 226 patients in the no lead-in arms, but the press release only shows 210 in the no lead-in arms. Since this was an intention to treat analysis those 16 should have been left in and I have to assume that if they had a SVR they would be in. Just supposing those 16 were equally distributed between the two arms and they were not SVR, they would reduce the effective percentages from 55% to 51% in the 28 week arm and 66% to 61% in the 48 week arm. This is certainly speculation but those 16 patients should be accounted for and until they are the data is suspect.
Is Schering using this press release to manufacture an advantage here to boost their stock price?
For more on this story, read, Standing on the Shoulders of Giants -- the "Missing" 16 Patients in Schering's Boceprevir Press Releases. . . .