h******e
发帖数: 1791 | 1 请问如果一些time-to-event的值是负数该如何处理？谢谢。 | z*******n
发帖数: 15481 | | D*********2
发帖数: 535 | | h******e
发帖数: 1791 | 4 看一种ae，但少数发生在第一次吃药之前。
【在 D*********2 的大作中提到】
: re.
: 为啥会有负数。
| s*r
发帖数: 2757 | 5 data error. it always happens.
for ae, that is pre-existing condition, not teae | n*****3
发帖数: 1584 | 6 what is the purpose of the this trials/study?
if you want to see the effects of that drug, these subjects are not
eligible to be considered in the analysis, I guess.
But if I see these negative results, I will say these are very very
likely programming errors, instead of data errors. because time to events
are derived variable, not raw variable.
these negative time subjects should be considered censored! even though they
have some kind of changes from pre-dose to date of random/dosing, these
should not count, the time direction is not right.
I will say, check the program/macro first.
【在 h******e 的大作中提到】
: 看一种ae，但少数发生在第一次吃药之前。
|
|
