Bioequivalence and Bioavailability Forum

hiren379 ★ India, 2013-02-19 15:40 (4448 d 21:54 ago) Posting: # 10057 Views: 9,808	Sample Size Case [Power / Sample Size] Post reply
	Hello All, Can someone help me in calculating appropriate sample size for a drug (relatively safe drug with very few drop outs expected due to AE) having T/R = 0.90 and % CVintra = 60 for a Reference Replicated Pivotal BE study (3 period)? We opt for 80% power...!!!

Helmut
★★★

Vienna, Austria,
2013-02-19 16:22
(4448 d 21:12 ago)

@ hiren379
Posting: # 10058
Views: 8,618

R/PowerTOST or Lászlós’ tables

Post reply

Hi Hiren!

❝ […] appropriate sample size for a drug […] having T/R = 0.90 and % CVintra = 60 for a Reference Replicated Pivotal BE study (3 period)? We opt for 80% power...!!!

Which regulation: FDA or EMA?

FDA
sampleN.RSABE(theta0=0.9, CV=0.6, design="2x3x3", details=FALSE) ++++++++ Reference scaled ABE crit. +++++++++ Sample size estimation --------------------------------------------- Study design: 2x3x3 log-transformed data (multiplicative model) 1e+05 studies simulated. alpha = 0.05, target power = 0.8 CVw(T) = 0.6; CVw(R) = 0.6 Null (true) ratio = 0.9 ABE limits / PE constraints = 0.8 ... 1.25 Sample size n power 33 0.817090
EMA
require(PowerTOST) sampleN.scABEL(theta0=0.9, CV=0.6, design="2x3x3", details=FALSE) +++++++++++ scaled (widened) ABEL +++++++++++ Sample size estimation --------------------------------------------- Study design: 2x3x3 log-transformed data (multiplicative model) 1e+05 studies simulated. alpha = 0.05, target power = 0.8 CVw(T) = 0.6; CVw(R) = 0.6 Null (true) ratio = 0.9 ABE limits / PE constraints = 0.8 ... 1.25 Regulatory body: EMA - CVswitch = 0.3, cap on ABEL if CV > 0.5 - Regulatory constant = 0.76 Sample size search n power 48 0.814780

Note that FDA’s method (function sampleN.RSABE) is not available in the current official release of PowerTOST (v1.1-00). I used a private copy of 1.1-01 provided by Detlew. I guess he will upload it to CRAN returning from his vacation in two weeks. In the meantime see the approximation in v1.1-00:

require(PowerTOST) sampleN.scABEL(theta0=0.9, CV=0.6, design="2x3x3", regulator="FDA", details=FALSE) +++++++++++ scaled (widened) ABEL +++++++++++ Sample size estimation --------------------------------------------- Study design: 2x3x3 log-transformed data (multiplicative model) 1e+05 studies simulated. alpha = 0.05, target power = 0.8 CVw(T) = 0.6; CVw(R) = 0.6 Null (true) ratio = 0.9 ABE limits / PE constraints = 0.8 ... 1.25 Regulatory body: FDA Sample size n power 30 0.800130

See also Tóthfalusi and Endrényi (2011)* – Tables A1 and A3 for GMR 0.9, CV 60%, 80% power:
EMA: 46 (round up to the next multiple of three: 48)
FDA: 31 (round up to the next multiple of three: 33)

Tóthfalusi L and L Endrényi
Sample Sizes for Designing Bioequivalence Studies for Highly Variable Drugs
J Pharm Pharmaceut Sci 15(1), 73–84 (2011)
online

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Ken Peh ★ Malaysia, 2013-02-20 13:45 (4447 d 23:49 ago) @ Helmut Posting: # 10076 Views: 8,372	R/PowerTOST or Lászlós’ tables Post reply
	Dear Helmut, Is PowerTost a commercial software ? Do we need to subscribe to gain access to this wonderful software ? I would very much love to have it for my work. Thank you. Regards, Ken

Helmut
★★★

Vienna, Austria,
2013-02-20 15:04
(4447 d 22:30 ago)

@ Ken Peh
Posting: # 10077
Views: 9,966

R/PowerTOST and other packages

Post reply

Dear Ken!

❝ Is PowerTost a commercial software ? Do we need to subscribe to gain access to this wonderful software ?

No, it is free software (GNU General Public License 2/3). The author is a member of the forum, Detlew Labes. :ok:

First you have to install the statistical software R (mirror closest to Malaysia is maintained at the Department of Statistics and Applied Probability of the National University of Singapore; available for Linux, MacOS X, and Windows).

Once installed start the R-console > Packages > Install package(s)… > CRAN mirror > OK. You need two packages: mvtnorm and PowerTOST (select them in the drop-down list).
To start type ~~load(PowerTOST)~~* library(package) or require(PowerTOST). Type help(package=PowerTOST) in the R-console to display the documentation in your bowser.

There are some other packages which might be of interest. Incomplete list:

bear: Average bioequivalence and bioavailability data analysis tool
ivivc: A data analysis tool for in vitro-in vivo correlation (IVIVC)
PK: Basic Non-Compartmental Pharmacokinetics
PKfit: A Data Analysis Tool for Pharmacokinetics
randomizeBE: Function to create a random list for crossover studies
stab: Data analysis of drug stability for shelf life estimation

See Detlew’s post.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Ken Peh
★

Malaysia,
2013-02-24 10:38
(4444 d 02:55 ago)

@ Helmut
Posting: # 10114
Views: 8,289

R/PowerTOST and other packages

Post reply

Dear Helmut,

❝ There are some other packages which might be of interest. […]

Thank you very much for providing the link to the software packages above. Very useful to our work. As I am not good in computer, will need time to understand how to operate the software. I presume the installation and instruction to run the software is the same as R/PowerTOST.

Regards,
Ken

Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Helmut]

Helmut
★★★

Vienna, Austria,
2013-02-24 18:00
(4443 d 19:34 ago)

@ Ken Peh
Posting: # 10115
Views: 8,309

R’s packages

Post reply

Dear Ken!

❝ I presume the installation and instruction to run the software is the same as R/PowerTOST.

Yes. Once you installed a package and loaded it – by ~~load(package)~~* library(package) or require(package) – you can always get instructions by help(package="…").
Write your scripts in any text-editor and save it with the extension .R (not .txt!). Once in the R-console > File > Source R code… (navigate to the *.R-file > Open) runs the script.
I prefer require(package) over ~~load(package)~~ library(package) because the former loads the package only if not already loaded (saves some fractions of a second). ;-)

P.S.: In the future please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post!

See Detlew’s post.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

d_labes
★★★

Berlin, Germany,
2013-02-28 22:20
(4439 d 15:14 ago)

@ Helmut
Posting: # 10136
Views: 8,400

load() ?

Post reply

Dear Helmut!

❝ I prefer require(package) over load(package) because the former loads the package only if not already loaded (saves some fractions of a second). ;-)

> load(PowerTOST)

Error in load(PowerTOST) : object 'PowerTOST' not found

From the help file of

load():

>? load

load {base}                                       R Documentation

Reload Saved Datasets



Description



Reload datasets written with the function save.

May be you meant library(PowerTOST) instead of load(PowerTOST) or missed I somefink here?
I must confess that I never understood really the difference between require(...) and library(...)

BTW:
@Ken Peh and hiren379:
PowerTOST version 1.1-02 is on the way now which makes accessible the function sampleN.RSABE() and also the complementary function power.RSABE() not only to Helmut, my favorite nitpicking power tester, but to the general audience. But be patient ... Maybe it takes a couple of days until it arrives on you.

—
Regards,

Detlew

Helmut
★★★

Vienna, Austria,
2013-03-01 17:22
(4438 d 20:12 ago)

@ d_labes
Posting: # 10141
Views: 8,148

load() ???

Post reply

Dear Detlew,

welcome back; hope you had a nice vacation. :smoke:

❝ ❝ I prefer require(package) over load(package) because the former loads the package only if not already loaded (saves some fractions of a second). ;-)

❝ May be you meant library(PowerTOST) instead of load(PowerTOST) or missed I somefink here?

Oops; sure.

❝ I must confess that I never understood really the difference between require(...) and library(..)

require(…) loads the package only if not already loaded. library(…) loads in all cases. Therefore the former might be a little bit faster.

Edit: PowerTOST 1.1-02 (2013-02-28) available at CRAN.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Sample Size Case [Power / Sample Size]

R/PowerTOST or Lászlós’ tables

R/PowerTOST or Lászlós’ tables

R/PowerTOST and other packages

R/PowerTOST and other packages

R’s packages

load() ?

load() ???