Bioequivalence and Bioavailability Forum • What software can change alpha to evaluate BE?

wxp
☆

China,
2013-03-29 10:31
(4825 d 01:22 ago)

Posting: # 10307
Views: 7,229

What software can change alpha to evaluate BE? [Software]

Dear all,

We did a typical 2×2 cross-over study for BE study in 23 subjecct. The results (by WinNonlin 6.3) showed the power<0.8 and also not BE (in the "Average Bioequivalence" table). We want to try to evaluate the data by alpha =0.0294 (based on Potvin's Method C). But I did not find the option in WinNonlin. I wonder if anyone could tell me which software can do this? I think it sound a bit stupid, but I really hope someone could give me a favor.
What I have and know: pharmacy background and have WinNonlin (learn on my own by the Guide), R (basic skill, but I hear the R package bear and PowerTost in the forum :-D

) and NONMEM.

Thanks in advance!

—
Best regards,

Xipei
China

ElMaestro
★★★

Denmark,
2013-03-29 11:38
(4825 d 00:14 ago)

@ wxp
Posting: # 10308
Views: 6,379

What software can change alpha to evaluate BE?

Post reply

Hello Xipei,

❝ We did a typical 2×2 cross-over study for BE study in 23 subjecct. The results (by WinNonlin 6.3) showed the power<0.8 and also not BE (in the "Average Bioequivalence" table). We want to try to evaluate the data by alpha =0.0294 (based on Potvin's Method C).

Erm... is that really called for? I think alpha = 0.0294 will fail if alpha=0.05 failed.

❝ But I did not find the option in WinNonlin. I wonder if anyone could tell me which software can do this? I think it sound a bit stupid, but I really hope someone could give me a favor.

❝ What I have and know: pharmacy background and have WinNonlin (learn on my own by the Guide), R (basic skill, but I hear the R package bear and PowerTost in the forum :-D ) and NONMEM.

I would use R, but this is just because that's all I have :-D

. Key is to use the qt function with the alpha of your choice after you have the treatment effect and residual SS and df from an lm.
Good luck.

—
Pass or fail!
ElMaestro

wxp
☆

China,
2013-04-02 12:06
(4821 d 00:47 ago)

@ ElMaestro
Posting: # 10325
Views: 5,788

What software can change alpha to evaluate BE?

Post reply

❝ Erm... is that really called for? I think alpha = 0.0294 will fail if alpha=0.05 failed.

Thank you, ElMaestro. Yes, it fails, :no:

. And we are considering to do a second stage of the BE study.

❝ I would use R, but this is just because that's all I have :-D . Key is to use the qt function with the alpha of your choice after you have the treatment effect and residual SS and df from an lm.

I am trying to use R to handle the problem. It is a good experience. :-D

—
Best regards,

Xipei
China

krishna
☆

India,
2013-03-29 12:32
(4824 d 23:21 ago)

@ wxp
Posting: # 10309
Views: 6,146

What software can change alpha to evaluate BE?

Post reply

Hi Xipei,

In WinNonlin 5.2 version, you can put directly 94.12 in Confidence level option box. Check the same in your version.

Regards,

Krishna.

wxp ☆ China, 2013-04-02 12:03 (4821 d 00:50 ago) @ krishna Posting: # 10324 Views: 5,780	What software can change alpha to evaluate BE? Post reply
	❝ In WinNonlin 5.2 version, you can put directly 94.12 in Confidence level option box. Check the same in your version. Thanks a lot, Krishna. It works in version 6.3. — Best regards, Xipei China

Helmut
★★★

Vienna, Austria,
2013-03-29 12:46
(4824 d 23:07 ago)

@ wxp
Posting: # 10310
Views: 6,161

Phoenix settings for Potvin B/C

Post reply

Hi Xipei!

In the Bioequivalence Options set Confidence Level 94.12. You will get (Potvin’s Example 2 after the second stage 2):

Output Data >

Average Bioequivalence

| CI_User_Lower | CI_User_Upper |

|     88.445196 |     116.37701 |

and (if Core Output selected in General Options):

Bioequivalence Statistics

User-Specified Confidence Level for CI's and Power = 94.1200

…

    Ratio(%Ref) =   101.4544



                      Classical                Westlake

    CI  80% = (   92.6916,  111.0456)  (   90.6629,  109.3371)

    CI  90% = (   90.1755,  114.1440)  (   87.7333,  112.2667)

    CI  95% = (   87.9420,  117.0430)  (   84.9996,  115.0004)

    CI User = (   88.4452,  116.3770)  (   85.6277,  114.3723)

    Average bioequivalence shown for confidence=94.12 and percent=20.0.

BTW, ElMaestro is right. If you already failed with the 90% CI you will fail with the 94.12% CI as well (smaller α → wider 1–2×α CI).

—
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

wxp
☆

China,
2013-04-02 12:20
(4821 d 00:32 ago)

@ Helmut
Posting: # 10326
Views: 5,784

Phoenix settings for Potvin B/C

Post reply

❝ In the Bioequivalence Options set Confidence Level 94.12. You will get (Potvin’s Example 2 after the second stage 2):

Thank you very much for step-by-step explanation. Through this, I also know the core information in the core output file. :-)

❝ BTW, ElMaestro is right. If you already failed with the 90% CI you will fail with the 94.12% CI as well (smaller α → wider 1–2×α CI).

The result is the same as you said.

—
Best regards,

Xipei
China