f*******2
发帖数: 656 | 1 各位大牛，
i have read this phase I study, a basic three plus three design, and
statistical section was written as below. i have problem replicating the
probality calculations (not just a simple binomial distribution, maybe some
bayesian assumptions were involved, but unfortunately my knowledge of
Bayesian Design is zero). if any one can give me an quick lecture to help to
validate the statement, i would be really grateful. Thanks!!!!
Assuming that the probability of DLT lies between 1% and 5% at XXX |
f*******2
发帖数: 656 | 2 自己顶一下.
any comments or suggestions are greatly appreciated! thanks!! |
s*r
发帖数: 2757 | 3 what is 'a basic three plus three design'
what is 'DLT '
those probabilities of early stopping might just be the powers at each stage
of group sequential studies |
f*******2
发帖数: 656 | 4 Thanks for breaking the ice. i can tell you are probably describing phase
III case scenario, and it is not the same as phase I study.
DLT refers to dose limiting toxicity. 3 plus 3 is a common phase I design
trying to find a right dose of a new drug.
looking forward to more inputs!!!
stage
【在 s*r 的大作中提到】
: what is 'a basic three plus three design'
: what is 'DLT '
: those probabilities of early stopping might just be the powers at each stage
: of group sequential studies
|
s*r
发帖数: 2757 | 5 it is good to know.
why do you think it is not usual binomial CIs
there are many ways to get the ci
what is the limit to stop? 2 dlt?
【在 f*******2 的大作中提到】
: Thanks for breaking the ice. i can tell you are probably describing phase
: III case scenario, and it is not the same as phase I study.
: DLT refers to dose limiting toxicity. 3 plus 3 is a common phase I design
: trying to find a right dose of a new drug.
: looking forward to more inputs!!!
:
: stage
|
f*******2
发帖数: 656 | 6 becuase binomial probablity cannot replicate the numbers they provided, and
plus, i don't think that is CIs, it is more like to be point probablities
based on the toxicity indicent rate.
【在 s*r 的大作中提到】
: it is good to know.
: why do you think it is not usual binomial CIs
: there are many ways to get the ci
: what is the limit to stop? 2 dlt?
|
