Bioequivalence and Bioavailability Forum • Calculation of sample size for a 3 period cross over design

MGR
★

India,
2006-12-10 08:26
(7127 d 00:04 ago)

Posting: # 390
Views: 11,015

Calculation of sample size for a 3 period cross over design [Power / Sample Size]

hi to all,
can anybody suggest me the formula or method to calculate a sample size for three period three treatment cross over design by using SAS software. Also please tell me how to get intrasubject variability for a three way cross over design.

(Note: We have already done a three way cross-over pilot study under fed conditions for a particular molecule where we did not get intra subject variability and we want to know how much sample size is required for a Pivotal study for the same molecule.)

So please give me suggestion or any reference / formula, how to proceed in SAS / WinNonlin software under this situation.

Thanks and Regards,
MGR

—
Regards,
MGR

Helmut
★★★

Vienna, Austria,
2006-12-10 13:51
(7126 d 18:39 ago)

@ MGR
Posting: # 391
Views: 10,106

CV-intra from MSE

Post reply

Dear Mamidi!

❝ can anybody suggest me the formula or method to calculate a sample size for three period three treatment cross over design by using SAS software. Also please tell me how to get intrasubject variability for a three way cross over design.

I'm not using SAS, but the calculation is not different form a 2×2 design... (for detail have a look at Chow & Liu, 2000, pp 303)

❝ So please give me suggestion or any reference / formula, how to proceed in SAS / WinNonlin software under this situation.

Concerning WinNonlin, I just checked that versions 5.0.1 and 5.1.1 do not give the intra-subject CV for a 3×6 Williams' design (as it does for a 2×2 design in the ASCII output just below the residual-variance line).
Therefore you have to go back to the pre-5.0 method, namely calculating the CV manually! ;-)

CV = sqrt(exp(MSE) - 1), where
MSE is the residual variance

You also may use this formula for your SAS-output…

WinNonlin procedure:
In the workbook derived from the BE Wizard
File > Dependencies... > click your workbook > Detach > Yes > Exit
In the workbook navigate to 'Final Variance Parameters' sheet
In the row with parameter 'Var(Residual)' move to the cell right to the 'Estimate' column (generally cell E2), enter
=SQRT(EXP(D2)-1)

Maybe this paper helps?

—
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

mamidi.ganesh
●

2008-02-12 15:32
(6697 d 16:58 ago)

@ Helmut
Posting: # 1611
Views: 8,662

CV-intra from MSE

Post reply

Dear HS,

❝ WinNonlin procedure:

❝ In the workbook derived from the BE Wizzard

❝ File > Dependencies... > click your workbook > Detach > Yes > Exit

❝ In the workbook navigate to 'Final Variance Parameters' sheet

❝ In the row with parameter 'Var(Residual)' move to the cell right to the

❝ 'Estimate' column (generally cell E2), enter

❝ =SQRT(EXP(D2)-1)

Can you please explain the same as above for a 2 treatment 4 period 2 sequence replicate design? I am using WinNonlin 5.1.1, in this the Intrasubject CV is not shown in the output. Kindly guide me in this. Thanks in advance.

Regards,
Raghavendra