Bioequivalence and Bioavailability Forum • Clarification on model statement in EMEA submission

RK
☆

India,
2013-10-08 15:39
(4236 d 09:51 ago)

Posting: # 11623
Views: 6,356

Clarification on model statement in EMEA submission [RSABE / ABEL]

Hi all,

I would like to clarify two things from below given program for reference replicate design,

what is the use of keeping G matrix in the model. if we ignore, wether it will affect the final results.
if the study is for EMEA submission, random effect will not be considered, but where as in the below given program treatment effect given as both fixed and random. what to do???

can we considered this as Subject within TRT as random effect for EMEA?

Calculation of unscaled 90% bioequivalence confidence intervals:

PROC MIXED

data=pk;

CLASSES SEQ SUBJ PER TRT;

MODEL LAUCT = SEQ PER TRT/ DDFM=SATTERTH;

RANDOM TRT/TYPE=FA0(2) SUB=SUBJ G;

REPEATED/GRP=TRT SUB=SUJ;

B ESTIMATE 'T vs. R' TRT 1 -1/CL ALPHA=0.1;

run;

Regards
RK.

Helmut
★★★

Vienna, Austria,
2013-10-08 16:08
(4236 d 09:21 ago)

@ RK
Posting: # 11625
Views: 5,380

EMA: all effects fixed

Post reply

Hi RK,

❝ […] if the study is for EMEA submission, random effect will not be considered, but where as in the below given program treatment effect given as both fixed and random. what to do???

For unscaled use Proc GLM dropping the RANDOM line (I guess – don’t speak SAS). For scaling use EMA’s code in the Q&A-document. Seems that “Method A” is preferred over “Method B”.

—
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

RK ☆ India, 2013-10-09 08:32 (4235 d 16:57 ago) @ Helmut Posting: # 11628 Views: 5,306	EMA: all effects fixed Post reply
	❝ For unscaled use `Proc GLM` dropping the `RANDOM` line (I guess – don’t speak SAS). For scaling use EMA’s code in the Q&A-document. Seems that “Method A” is preferred over “Method B”. Dear Helmut, I have compared the methods B and C, The results are, Ratio and 90% CI, Method B:92.21(80.81,105.21) Method C:92.41(82.57,103.42) which one is correct??? Regards RK

d_labes
★★★

Berlin, Germany,
2013-10-09 10:30
(4235 d 15:00 ago)

@ RK
Posting: # 11629
Views: 5,379

EMA: all effects fixed - Method A!

Post reply

Dear RK!

❝ I have compared the methods B and C,

❝

❝ The results are,

❝

❝ Ratio and 90% CI,

❝

❝ Method B:92.21(80.81,105.21)

❝ Method C:92.41(82.57,103.42)

❝ which one is correct???

Since we are talking about EMA submission: none of them.
As Helmut already pointed out the EMA prefers what they call 'Method A' i.e. the Proc GLM code given in the EMA Q&A-document. The code behind 'Method A' is the same as is usually employed for a classical 2x2x2 crossover.

Talking about the scientific rationale behind this preference is impossible since it seems a pure 'political' decision :crying:

.

BTW: In case of no missings in your dataset 'Method A' and 'Method B' will give practically the same results.

—
Regards,

Detlew

Helmut
★★★

Vienna, Austria,
2013-10-09 16:50
(4235 d 08:39 ago)

@ RK
Posting: # 11634
Views: 5,203

EMA: Method C “incompatible with CHMP GL”

Post reply

Hi RK,

❝ I have compared the methods B and C,

I hope you mean EMA’s Methods A and B?
Method C is FDA’s – which is unlikely to be accepted according to the Q&A-document (only A and B are mentioned in the respective headings as “compatible with CHMP guideline”).

—
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

RK ☆ India, 2013-10-10 08:01 (4234 d 17:28 ago) @ Helmut Posting: # 11637 Views: 5,200	EMA: Method C “incompatible with CHMP GL” Post reply
	yes i got it Thank you 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]