Sir Richard Peto takes umbrage at the idea that he might skew a statistical analysis to benefit a drugmaker. The eminent Oxford statistician, as you know, recently weighed in on a study that seemed to link the Merck/Schering-Plough cholesterol med Vytorin with an increase in cancer cases and deaths. Peto concluded that the apparent risk was a statistical anomaly, a freak of chance.
Now, Peto is defending that conclusion against the opinions of other experts and questions from the U.S. Congress. In a letter to Reps. John Dingell and Bart Stupak, who are investigating the Vytorin controversy, Peto asserted that "any competent trial statistician" would agree with his analysis. Suggestions that he might bias his work in favor of Merck and Schering-Plough are absolutely wrong: They don't pay his salary and had no impact on his analysis of the trial data.
Apparently, Peto takes a dim view of the competency of the New England Journal of Medicine and its experts. Two NEJM editorials recently critiqued Peto's analysis, saying that an increased risk of cancer death in Vytorin patients can't be ruled out. Peto hasn't specifically addressed their conclusions--yet.