the small numbers, when you look at it using different statistical analyses you get
different answers, then you feel even less secure, so. I think that it’s well taken.
statistician” and then they make good statistical points. some of the examples that you
are talking about involve monitoring over time and that’s the kind of situation where if
you look many times you are going to find sometimes it goes the wrong direction, and
that’s why we have to use monitoring boundaries to take that into consideration. Having
said that though, you are correct that there have been some trials where they tried to
repeat the trial, they thought the patients that they repeated the trial in were very similar,
should have gotten the same results and they got different results. It does happen, and so,
there is a limit to how confident that you can be, you know, how confident you could be,
that’s for sure.
the examples that you use, number one, were in the course of monitoring with multiple
looks being taken. Some of them were examples of subgroups; subgroup results that
were not born out when the same question was asked in a unified prospective way. So
those are all factors that you have to think about. I think we have to remember what is
this question about, and sensitivity for what, for better or for worse, I think the bar that
has been set for approvability with the understanding that we may request and mandate
additional studies for greater degrees of confidence.
I guess I am still reassured because sometimes in a situation like that with
DR. BURMAN: Thank you. Dr. Konstam.
DR. KONSTAM: So, you know, I agree with Mike and John. Some of
The bar that’s been set is an upper confidence level of 1.8. So, I guess in
thinking about that, I mean that’s really the issue at hand, at least for the first part of our
DR. PROSCHAN: I hate when these people start out with “I’m not a
Scribes, LLC Toll Free 1-800-675-8846 www.scribesllc.com