Nisha
☆    

Israel,
2015-09-17 15:28
(3115 d 03:29 ago)

Posting: # 15419
Views: 6,302
 

 Different power of AUC vs. Cmax at stage I - continuation rules [Two-Stage / GS Designs]

Hi all,
I would like to raise a scenario following stage I results with Potvin C:

If, hypothetically, our results at stage I would have had good ratio, and failed to meet 94.12% CI BE criteria at stage I for both Cmax and AUC.
The power analysis would have shown results higher than 80% for AUC, and lower than 80% for Cmax. Based on Cmax, we could continue to stage II. In that case - what would be the practice for AUC? I would assume it would have been calculated also for stage II. How should this be covered in the protocol?

Would appreciate your input.

Many thanks,
Nisha
ElMaestro
★★★

Denmark,
2015-09-17 15:56
(3115 d 03:01 ago)

@ Nisha
Posting: # 15420
Views: 5,337
 

 Different power of AUC vs. Cmax at stage I - continuation rules

Hi Nisha,

good to hear the 2-stage approaches have not drawn their last breath.

❝ If, hypothetically, our results at stage I would have had good ratio, and failed to meet 94.12% CI BE criteria at stage I for both Cmax and AUC.

❝ The power analysis would have shown results higher than 80% for AUC, and lower than 80% for Cmax. Based on Cmax, we could continue to stage II.


Typically you do not need to make a distinction between AUCt and Cmax. You can go by the maximum observed CV - at the end of the day this can be argued to be both the most ethical and most scientifically relevant metric for the purpose of deciding if a stage II is needed and what the sample size should be.
In the (vast) majority of your studies you will see that CV for Cmax is higher than CV for AUCt; therefore it is generally also ok to write in the protocol that Potvin C is followed on basis of the Cmax data.

You would still be analysing AUCt and Cmax at stage II if one such is conducted, even if AUCt has been shown BE at stage I. And don't worry too heavily about the latter; if AUCt was ok at stage I then in all likelihood it will be as well at stage II.

Good luck.

Pass or fail!
ElMaestro
Nisha
☆    

Israel,
2015-09-17 22:27
(3114 d 20:30 ago)

@ ElMaestro
Posting: # 15424
Views: 5,303
 

 Different power of AUC vs. Cmax at stage I - continuation rules

Hi ElMaestro,
Many thanks for your prompt response.

❝ good to hear the 2-stage approaches have not drawn their last breath.


May I ask why it is not so commonly used?

❝ You would still be analysing AUCt and Cmax at stage II if one such is conducted, even if AUCt has been shown BE at stage I.


What about a case in which AUC doesn't meet the BE criteria at stage I, but its power is >80% due to the low ISCV, while Cmax power is <80%? Could we still continue to stage II?

Thanks again,
Nisha
ElMaestro
★★★

Denmark,
2015-09-17 23:53
(3114 d 19:04 ago)

@ Nisha
Posting: # 15425
Views: 5,486
 

 AUC vs. Cmax

Hi Nisha,

❝ May I ask why it is not so commonly used?


The Potvin methods and their derivatives all suffer from one drawback: It is extremely difficult to argue they are usefulk when there is undcertainty about the point estimate. They are almost only useful if the uncertainty is just on the CV.
The reason is that a fixed GMR is used for quantification of the second stage; if you plug in the observed GMR then the methods just outright fail.

❝ What about a case in which AUC doesn't meet the BE criteria at stage I, but its power is >80% due to the low ISCV, while Cmax power is <80%? Could we still continue to stage II?


Yes, I would write in the protocol that you go by the highest CV or the CV observed for Cmax.

Pass or fail!
ElMaestro
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-09-18 02:15
(3114 d 16:42 ago)

@ Nisha
Posting: # 15426
Views: 5,278
 

 Assumption

Hi Nisha,

❝ May I ask why it is not so commonly used?


Maybe because people belong to the group described in the last paragraph of this post?
If they tell me “I don’t like TSDs since I have to assume a T/R-ratio!” I use to reply “And what do you do in your beloved fixed-sample design?“

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
92 visitors (1 registered, 91 guests [including 10 identified bots]).
Forum time: 17:57 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5