# Bioequivalence and Bioavailability Forum 00:31 CET

mahmoud-teaima
Regular

2017-10-24 19:18
(edited by mahmoud-teaima on 2017-10-24 20:44)

Posting: # 17916
Views: 3,763

## T/R ratio for estimation of sample size in BE study of NTI drugs [Power / Sample Size]

Salam to all,
Pease, advise me about what is the best T/R ratio to be used for estimating sample size for a BE study of NTI drugs? knowing that the GL to be followed set the 90% CI acceptance range to (90-111.11) for Cmax if a "2x2x2" design be followed.
I'm confused to use T/R ratio of 0.95 or 0.975?!

```+++++++++++ Equivalence test - TOST +++++++++++             Sample size estimation ----------------------------------------------- Study design:  2x2 crossover log-transformed data (multiplicative model) alpha = 0.05, target power = 0.8 BE margins = 0.9 ... 1.1111 True ratio = 0.95,  CV = 0.0945 Sample size (total)  n     power 40   0.808788 +++++++++++ Equivalence test - TOST +++++++++++             Sample size estimation ----------------------------------------------- Study design:  2x2 crossover log-transformed data (multiplicative model) alpha = 0.05, target power = 0.8 BE margins = 0.9 ... 1.1111 True ratio = 0.975,  CV = 0.0945 Sample size (total)  n     power 20   0.820356```

Greetings.

Mahmoud Teaima, PhD.
Helmut
Hero

Vienna, Austria,
2017-10-24 23:15

@ mahmoud-teaima
Posting: # 17917
Views: 3,344

## T/R ratio for estimation of sample size in BE study of NTI drugs

Salam Mahmoud,

» Pease, advise me about what is the best T/R ratio to be used for estimating sample size for a BE study of NTI drugs? knowing that the GL to be followed set the 90% CI acceptance range to (90-111.11) for Cmax if a "2x2x2" design be followed.
» I'm confused to use T/R ratio of 0.95 or 0.975?!

Both are fine. Note that the FDA requires tighter batch release spec’s for NTIDs. Hence, in function `sampleN.NTIDFDA()` of `PowerTOST` the default is 0.975 – contrary to the usual 0.95.

```library(PowerTOST) sampleN.NTIDFDA(CV=0.0945) +++++++++++ FDA method for NTID's +++++++++++            Sample size estimation --------------------------------------------- Study design:  2x2x4 log-transformed data (multiplicative model) 1e+05 studies for each step simulated. alpha  = 0.05, target power = 0.8 CVw(T) = 0.0945, CVw(R) = 0.0945 True ratio     = 0.975 ABE limits     = 0.8 ... 1.25 Implied scABEL = 0.9054 ... 1.1044 Regulatory settings: FDA - Regulatory const. = 1.053605 - 'CVcap'           = 0.2142 Sample size search  n     power 16   0.777210 18   0.831050```

Cheers,
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. ☼
Science Quotes
ElMaestro
Hero

Denmark,
2017-10-25 12:34

@ Helmut
Posting: # 17920
Views: 3,294

## T/R ratio for estimation of sample size in BE study of NTI drugs

Hi Hötzi,

» Both are fine. Note that the FDA requires tighter batch release spec’s for NTIDs. Hence, in function `sampleN.NTIDFDA()` of `PowerTOST` the default is 0.975 – contrary to the usual 0.95.

This strays away from the usually stringent scientific principles that this forum usually promotes, doesn't it?
We plug in the GMR that we think/predict/guess is a realistic and relevant indicator of relative product performance, possibly with a buffer to factor in worst cases. Habits, experience and gut feeling.
There is no divine aspect of NTID's, as far as I know, that makes them easier to copy than non-NTID's from a strict formulation perspective so I don't know if the default is much science based.
Although I am sure 0.975 is condoned by the guy in the Armani suit. He would definitely also like it for non-NTID's.

Perhaps all this boils down to the use of anything "default" in a power calculation. It is extremely handy. And it is a bit dangerous. Perhaps the world would be a better place altogether if such functions did not assume defaults for CV and GMR and designs - this would force users to actually *** badum tssss *** think just a little bit about the values they plug in before deriving a power.

OK, rant over, I'll take my medicine now. Stage III syphilis...

` if (3) 4 `

Best regards,
ElMaestro

"(...) targeted cancer therapies will benefit fewer than 2 percent of the cancer patients they’re aimed at. That reality is often lost on consumers, who are being fed a steady diet of winning anecdotes about miracle cures." New York Times (ed.), June 9, 2018.
nobody
Senior

2017-10-26 14:03

@ ElMaestro
Posting: # 17921
Views: 3,220

## T/R ratio for estimation of sample size in BE study of NTI drugs

...to be true: I would support discouraging the use of defaults without any knowledge on the actual products (dissolution, pharmaceutical quality etc.). Planning a BE-study should involve that you know that the "correct" GMR / sample size is hindsight knowledge and one way or the other there is a risk to loose money (too many volunteers, too little power to conclude BE), one way ore the other...

Kindest regards, nobody
d_labes
Hero

Berlin, Germany,
2017-10-26 14:49

@ nobody
Posting: # 17922
Views: 3,235

## Defaults in PowerTOST

Dear Öberster Größter Meister, Nobody!

» ... Planning a BE-study should involve that you know that the "correct" GMR / sample size is hindsight knowledge ...

You never know the population values of CV and/or GMR you need to have a reliable sample size estimation. You always rely on assumptions / guesses, educated or not / gut feeling or whatever else for these values. If this is scientific is left to you. Default or not .

And remember: At the end of the day the sponsor (with or without Armani suite) is always right.
I never came across with cases where the sponsor desired sample size was not "scientifically" estimated and justified.
"Power Calculation - A guess masquerading as mathematics." Stephen Senn
Remember the BOSS button in Dave's Fartssie.xls! More than once I was asked to introduce such tool in PowerTOST.

Nevertheless you may contact the Maintainer of PowerTOST and suggest changes or create an issue on http://github.com/Detlew/PowerTOST/issues. Best to suggest to abstain from all defaults including alpha=0.05 or the BE acceptance limits 0.8 ... 1.25 to force the user to think about those settings.
But be warned. It's best not to tangle that Maintainer guy without an Armani suite. May be he is biting back .

Regards,

Detlew
Bioequivalence and Bioavailability Forum |  Admin contact
18,923 posts in 4,039 threads, 1,284 registered users;
online 6 (0 registered, 6 guests [including 6 identified bots]).

If you think it’s simple,
then you have misunderstood the problem.    Bjarne Stroustrup

The BIOEQUIVALENCE / BIOAVAILABILITY FORUM is hosted by
Ing. Helmut Schütz