I have a question. On Feb 26, Sonus had a conference call. I will quote from Dr. Quay, President and CEO of Sonus..."With respect to the reanalysis of data, the FDA for the first time in this letter and for the first time in my experience in 15 years in developing contrast agents, has raised a design issue with respect to the blind degree of our studies." - Because Quay's grammar is so poor, that statement makes little sense. What I think he is trying to say is that Sonus had no indication that their method of analysis was anything less than worthy of a gold star. The problem with the way Sonus analyzed their data is so basic, however, that anyone who has taken even an introductory statistics course would understand it. Sonus had before echogen and after pictures. They could compare the ultrasound images with echogen and without echogen. The readers of the images who evaluated whether echogen provided benefit knew which images were before the administration of the contrast agent and which images were after administration. That creates bias in the analysis. It's so basic that it's absurd that Sonus thought they could get away with it. OK, did Molecular Bio analyze their data the way Sonus did? NO! The analysis was the way it should be - with reviewers not knowing which images were taken with Optison and which were without. In other words, which were the before and which were the after. Later in the conference call, Quay says, " statisticians we are dealing with have requested an unusual, perhaps unique kind of analysis for these studies." Yo, Quay - wake up, have an espresso. Question for all of you on this thread is the following - does the fact that Molecular Bio did their analysis of the data in the way the FDA wants Sonus to do theirs and the fact that Quay is on record in saying that this "method of analysis" is unusual bother anyone? Are his statements in the conference call materially misleading? Why on earth would Sonus not analyze the data in a way that has been shown to be acceptable by the FDA - Molecular Bio got approval last year using the randomized method. My questions to you are just that - questions. I am not judging anyone here, just quoting precisely from the conference call. You folks do your own analysis. If Quay's method of analysis as fine, and if he is correct that the FDA's wanting a randomized analysis of the data is "unusual, perhaps unique", then why didn't Molecular Bio analyze their data that way?
Oh well, what the hell do I know, anyway? I flip burgers at Carl's Junior. |
