Bioequivalence and Bioavailability Forum • Partial Replicate with Groups

MGR
★

India,
2015-02-03 10:46
(4162 d 02:03 ago)

Posting: # 14357
Views: 5,274

Partial Replicate with Groups [RSABE / ABEL]

Hi All,

We have done a three way partial replicate (TRR|RTR|RRT) with 60 subjects in 3 groups each of 20 subjects with 45 days gap due to the subjects turnover rate for USFDA Submission.

So my question is:

Can we use group effect into the stat analysis?
if so how we can include those to ilat and dlat?
how can we add group effect into the analysis i.e, glm procedure for calculating the limits?

I have searched the USFDA site but i didn't find any article on this.

If there is any USFDA guideline/article, please share.

Thanks in advance.

Edit: Category changed. [Helmut]

—
Regards,
MGR

Helmut
★★★

Vienna, Austria,
2015-02-03 11:54
(4162 d 00:55 ago)

@ MGR
Posting: # 14358
Views: 4,154

Partial Replicate with Groups

Post reply

Hi MGR,

do I understand you correctly: You performed a study without knowing (i.e., stating in the SAP) how to evaluate its results? Bad idea. The partial replicate is a lousy design since FDA’s model in the ABE-code – which you have to use if s_WR <0.294 – is overspecified. It some cases any (!) software will fail to converge. Then you have a lot of data and no (!) outcome at all… In the future consider fully replicated designs (TRT|RTR or TRTR|RTRT) instead.

❝ 1. Can we use group effect into the stat analysis?

In the ABE-part, yes. Please search the forum. It would have been better, to follow FDA’s suggestion in study planning in order to avoid trouble. If all those conditions were fulfilled in the study you could try to argue not to incorporate group effects (also in RSABE).

❝ 2. if so how we can include those to ilat and dlat?

I don’t see how that could be done.

❝ 3. how can we add group effect into the analysis i.e, glm procedure for calculating the limits?

See one experience with the FDA and a suggestion by John.

—
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