Bad luck [Study Assessment]
Hi Mikalai,
Bad luck. If you planned the study for assumed CV, T/R-ratio, and a dropout-rate for a desired power π the chance that you fail to demonstrate (if all assumptions turn out to be correct) is β = 1 – π. In other words, if you aimed at 80% power, 1∕5 of studies of products which are BE will fail by pure chance. That’s part of the game.
Although post-hoc power is irrelevant, if you planned for 80% and fail to show BE it will be <80%. As expected.
I don’t know how your agency deals with that. For the EMA and the FDA chances are close to nil.
Nice, but you have to demonstrate BE for all.
You can’t change the protocol once the study is done, right?
I don’t see how you could do that.
Example: Study planned for CV 25%, T/R-ratio 0.95, 80% power ⇒ n 28. All as expected, except a worse T/R 0.88 ⇒ 90% 78.66–98.45%. If you play around with α you will see that the patients’ risk will be 7.97% instead of ≤5%. I don’t think that any agency will accept that.
❝ […] Unfortunately, results one of analytes, a parent drug, fell sightly outside of the 80-125% range, below 80% for Cmax.
Bad luck. If you planned the study for assumed CV, T/R-ratio, and a dropout-rate for a desired power π the chance that you fail to demonstrate (if all assumptions turn out to be correct) is β = 1 – π. In other words, if you aimed at 80% power, 1∕5 of studies of products which are BE will fail by pure chance. That’s part of the game.
❝ […] and post-hoc power was less than 80% for this analyte.
Although post-hoc power is irrelevant, if you planned for 80% and fail to show BE it will be <80%. As expected.
❝ Are there any ways we can defend bioequivalence without conducting another BE trail?
I don’t know how your agency deals with that. For the EMA and the FDA chances are close to nil.
❝ All other analytes fell within 80-125%.
Nice, but you have to demonstrate BE for all.
❝ Specifically, what should we put in the protocol …
You can’t change the protocol once the study is done, right?
❝ … or in a supporting letter to justify our lower border of CI around 79% for Cmax?
I don’t see how you could do that.
Example: Study planned for CV 25%, T/R-ratio 0.95, 80% power ⇒ n 28. All as expected, except a worse T/R 0.88 ⇒ 90% 78.66–98.45%. If you play around with α you will see that the patients’ risk will be 7.97% instead of ≤5%. I don’t think that any agency will accept that.
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png)
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png)
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Complete thread:
- Cmax is out of the range Mikalai 2018-08-01 12:05 [Study Assessment]
- Bad luckHelmut 2018-08-01 13:37
- Bad luck Mikalai 2018-08-01 13:50
- Cherry-picking Helmut 2018-08-01 15:01
- Bad luck Mikalai 2018-08-01 13:50
- Cmax is out of the range jag009 2018-08-02 17:56
- Bad luckHelmut 2018-08-01 13:37