☆

2008-06-27 15:24
(5460 d 05:38 ago)

Posting: # 1978
Views: 6,878

## Three way crossover study comparing two T with one R [General Sta­tis­tics]

Dear all,

Kindly let me know whether we can have separate Variance in three way crossover study comparing two Test with one Reference (e.g T1 and T2 with R).

When I used WinNonlin software for analysis I got combined variance for T1 T2 and R. I want Varince for T1 R and T2 R separatly is it possible. Kindly let me know.

Regards,

Atish
Helmut
★★★

Vienna, Austria,
2008-06-27 20:56
(5460 d 00:06 ago)

Posting: # 1979
Views: 6,232

## Variance 'Extraction'

Dear Atish!

❝ When I used WinNonlin software for analysis I got combined variance for T1

❝ T2 and R.

Not only in WinNonlin. You are getting only the residual variance - which includes information from all sequences/periods/treatments/subjects...

❝ I want Varince for T1 R and T2 R separatly is it possible.

Not from a three way study. If you have used a variance-balanced design (6 sequences) it should be possible to extract two 'pseudo-two-way studies' (see this post for an example), but I don't see the reason behind in your case.

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
☆

2008-06-28 09:01
(5459 d 12:01 ago)

@ Helmut
Posting: # 1980
Views: 5,738

## Variance 'Extraction'

Dear HS,

This is a pilot three squence, three period crossover study.

1. Ratio for T2/R is around 98.63% and for T1/R is around 50.00%. I guess Variance is too high Which has affected 90% CI for T2/R to go out of Bioequivalence limit.
2. Calculated sample size for ratio T2/R and Using common variance is too high (more than 100).
3. When I did the analysis by blocking T1 from pharmacokinetic output of WinNonlin ratio of T2/R is around 105.00% and CI fall within bioequivalence limit and intraCV is around 27.99% (calculated from variance).
4. Calculated sample size for ratio T2/R and Using variance (intra CV of 27.99%) is around 33.
5. By blocking T1, ANOVA table shows two degree of freedom for Sequence and Period, since there are 3 Sequence and 3 Period.
6. Is my analysis by blocking T1 is correct?
7. Can I use sample size of 36 if a pivotal study is planned for T2 and R?

Regards,

Atish
Ohlbe
★★★

France,
2008-06-28 16:32
(5459 d 04:29 ago)

Posting: # 1982
Views: 6,093

## Variance 'Extraction'

Dear Atish,

❝ intraCV is around 27.99%

How many subjects did you have in your study to be able to determine intraCV with such a precision ?

Have a look at this message. From a pilot study, poorly designed (3-sequence, not 6), "around 28 %" or even "around 30%" would be sufficient...

Regards
Ohlbe
☆

2008-06-29 12:01
(5458 d 09:00 ago)

@ Ohlbe
Posting: # 1985
Views: 6,149

## Variance 'Extraction'

Dear Ohlbe,
The study was done on 15 subject and one was dropped from satistical analysis.

Regards
Atish
Jaime_R
★★

Barcelona,
2008-07-11 23:41
(5445 d 21:20 ago)

Posting: # 2017
Views: 5,563

## Variance 'Extraction'

Dear Atish!

❝ [...] one was dropped from satistical analysis.

SCNR!                    ^^^^^^^
Sometimes a typo tells more than a thousand words...

Regards, Jaime
☆

2008-07-12 20:37
(5445 d 00:24 ago)

@ Jaime_R
Posting: # 2019
Views: 6,101

## Variance 'Extraction'

Dear Jaime_R/all,

I am sorry for the spelling mistake.

Regards,

Atish
Jaime_R
★★

Barcelona,
2008-07-14 18:05
(5443 d 02:56 ago)

Posting: # 2027
Views: 5,992

## Variance 'Extraction'

Dear Atish!

❝ I am sorry for the spelling mistake.

No need for excuses; it only sounded that funny - and it's a typing error I'm used to make myself (unconciously?!).

Regards, Jaime