Erkin
●    

2014-01-21 17:20
(3719 d 22:42 ago)

Posting: # 12218
Views: 4,506
 

 Fixed effects rather than random effects [Regulatives / Guidelines]

Dear All,

According to EMA BE GL 2010, 4.1.8.Evaluation / Statistical Analysis:
The terms to be used in the ANOVA model are usually sequence, subject within sequence, period and formulation. Fixed effects, rather than random effects, should be used for all terms.

And according to Mr. Helmut Schütz's 23-25 May 2012 Moscow Presentation PK–NCA, PK based Design, Biostatistics Part I Moscow PK-NCA, PK based Design

on page 111 and 112, the default setting of PHNX has to be changed to subject(sequence) to Fixed Effects.

We're using subject (sequence) as Random Effects in our statistical calculations for 2x2 cross-over design. We've done the calculations with two methods of ANOVA, and the results are the same.

The calculations with random effects are acceptable from EMA?

Thank you all...

Erkin Alkan
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-01-21 17:32
(3719 d 22:30 ago)

@ Erkin
Posting: # 12219
Views: 4,092
 

 EMA: fixed

Merhaba Erkin!

❝ We're using subject (sequence) as Random Effects in our statistical calculations for 2x2 cross-over design. We've done the calculations with two methods of ANOVA, and the results are the same.


Results should be identical if you have a balanced design (equal number of subjects in both sequences). Depending on the degree of imbalance you might get increasingly different results from both models.

❝ The calculations with random effects are acceptable from EMA?


IMHO, it is not worth the efforts trying. I received one deficiency letter last year from the Spanish agency.


PS: I switched off the e-mail notification in your post because they bounced. Same story as during your registration. I have no idea what I could do from my side. Maybe you talk to your IT-guys. The error I received in all its glory:

This message was created automatically by mail delivery software.

A message that you sent could not be delivered to one or more of its
recipients. This is a permanent error. The following address(es) failed:

  {your email-address}
    SMTP error from remote mail server after end of data:
    host {your mail-server} {your IP}: 552 Mail with no Date header not accepted here


Of course the timestamp was given in the header. See below:

------ This is a copy of the message, including all the headers. ------

Return-path: <[email protected]>
Received: from {my server} ({my server’s IP})
   by {server gateway} with smtp (Exim 4.69)
   (envelope-from <[email protected]>)
   id 1W5dJz-0007ZV-Cs
   for {your e-mail address}; Tue, 21 Jan 2014 16:34:55 +0100
X-SIL: 0a1b0f9a0001b51829758069000041b5
To: Erkin <{your e-mail address}>
Subject: Reply to your post
From: Bioequivalence and Bioavailability Forum <[email protected]>
MIME-Version: 1.0
X-Mailer: Php/5.2.6-1+lenny6
X-Sender-IP: {my client’s IP}
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: quoted-printable
Message-ID: <[email protected]>
X-Spam-Score: -1.8 (-)
X-Spam-Flag: no
X-Scan-Signature: bcddf3eb0974ec6e728b987b52a48d0a

Hi Erkin,
Helmut has written a reply to this post:

Subject: EMA: fixed


… followed by the text of my post. Strange. :lookaround:

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
ElMaestro
★★★

Denmark,
2014-01-21 18:12
(3719 d 21:51 ago)

@ Erkin
Posting: # 12220
Views: 3,935
 

 Fixed effects rather than random effects

Hello Erkin,

I agree with Helmut would also like to point out that I think the issue is not treated in the same fashion across all Europe. In some countries assessors don't give a rat's fart about it, possibly because they have no idea about what a fixed vs a random effect is, while in other countries the matter is of interest just because the guideline says effects must be fixed. I have not come a across a scientific justification for avoiding random effects in BE (yet keeping them in other analyses btw).

Pass or fail!
ElMaestro
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-01-21 19:04
(3719 d 20:58 ago)

@ ElMaestro
Posting: # 12221
Views: 4,143
 

 The never ending story

Hi ElMaestro,

❝ […] I have not come a across a scientific justification for avoiding random effects in BE.


Me not either – and I don’t expect to see one ever. Do you believe the mighty oracle has got one? The nicest explanation I know is by Gerard E. Dallal (giving an example about nurses measuring blood pressure by three different methods):

Factors can either be fixed or random. A factor is fixed when the levels under study are the only levels of interest. A factor is random when the levels under study are a random sample from a larger population and the goal of the study is to make a statement regarding the larger population.

In this example, method is a fixed factor. The purpose of this study is to examine these three methods of measuring blood pressure. There may be other methods, but they do not concern us here. When we are done, the hope is to make a statement comparing these three methods.

From the description of the study, the status of nurse is less clear. If the investigator cares only about these four nurses, nurse is a fixed factor. This might be the case where the study concerns the staff of a particular research unit and there is no goal of generalizing beyond the unit. Since only these four nurses matter, nurse is a fixed factor. However, it might be that the point of the study is to generalize the results to all nurses. In that case, these four nurses might be viewed as a random sample of the population of all nurses, making nurse a random factor.

One way to decide whether a factor is fixed or random is to ask what would happen if the study were repeated. If the same set of nurses would be used (as in the case of studying a particular research unit) the factor is fixed. If any set of nurses would do equally well, the factor is random.


Read subject instead of nurse and treatment instead of method… That’s BE! Do we want to make inferences about a larger population and/or run similar studies? I always did so.
With all effects fixed EMA should at least be as honest to redefine type I error in BE and replace “patient’s risk” or “consumer’s risk” by “volunteer’s risk”.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Erkin
●    

2014-01-21 21:33
(3719 d 18:30 ago)

@ Helmut
Posting: # 12223
Views: 3,863
 

 The neverending story

Dear Helmut and ElMaestro,

Thank you very much for your comments. Mr. Helmut Schütz, especially, the nurse example is the nicest explanation. Thank you, again.


PS: I've removed our server's e-mail date check on e-mail headers. I hope that the problem was solved.

Best regards,

Erkin Alkan
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,636 registered users;
112 visitors (0 registered, 112 guests [including 9 identified bots]).
Forum time: 16:03 CET (Europe/Vienna)

With four parameters I can fit an elephant,
and with five I can make him wiggle his trunk.    John von Neumann

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5