Log in
Register|
neethu ☆ 2009-01-21 13:31 (6364 d 03:20 ago) Posting: # 3086 Views: 5,766 |
|
|
Dear All, I am planning for a study with 24 subjects with 3 sequence and 6 periods. I would like to know about the randomization. how will the sequence occur for the last three periods? — Thanks & Regards Nees |
|
bala ☆ Bangalore, India, 2009-01-21 13:35 (6364 d 03:16 ago) @ neethu Posting: # 3087 Views: 5,126 |
|
|
Dear Neethu, Can you tell me how many treatments (how many test and reference)?? — Cheers Bala |
|
neethu ☆ 2009-01-21 13:39 (6364 d 03:12 ago) @ bala Posting: # 3088 Views: 5,064 |
|
|
Dear Bala, Its two test and one refernce? Thanks & Regards Neethu — Thanks & Regards Nees |
|
bala ☆ Bangalore, India, 2009-01-21 13:42 (6364 d 03:09 ago) @ neethu Posting: # 3089 Views: 5,067 |
|
|
Neethu is it cross over or replicate study ??? — Cheers Bala |
|
neethu ☆ 2009-01-21 13:44 (6364 d 03:07 ago) @ bala Posting: # 3090 Views: 5,079 |
|
|
its a cross over study.... — Thanks & Regards Nees |
|
bala ☆ Bangalore, India, 2009-01-21 14:10 (6364 d 02:41 ago) @ neethu Posting: # 3092 Views: 5,265 |
|
|
Dear Neethu... This is the first time I m hearing this kind of study... In FDA and other guidelines I am not able to locate this kind of study, however I will give you one layout for this... Totally 24 subjects
Sequences We can randomize 8 patients with sequences (1,2) 8 patients with sequences (1,3) 8 patients with sequences (2,3) Where subjects who were assigned (1,2) will receive T R T T T R in the 6 periods... Similarly the others... Just check with your protocol whether its allowing this... — Cheers Bala |
|
Helmut ★★★   Vienna, Austria, 2009-01-21 14:17 (6364 d 02:35 ago) @ bala Posting: # 3094 Views: 5,092 |
|
|
Dear bala! Yes, but Neethu stated that it's not a replicate study... — Dif-tor heh smusma 🖖🏼 Довге життя Україна! ![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png) Helmut Schütz ![[image]](https://static.bebac.at/img/CC by.png) The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
Helmut ★★★ Vienna, Austria, 2009-01-21 14:14 (6364 d 02:37 ago) @ neethu Posting: # 3093 Views: 5,053 |
|
|
Dear Neethu, following your conversation, are you sure that you didn't mix-up periods and sequences (6 sequences and 3 periods)? If it's not a replicate study I don't get the idea how you will randomize three treatments in six periods...  — 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 |
|
bala ☆ Bangalore, India, 2009-01-21 14:20 (6364 d 02:31 ago) @ Helmut Posting: # 3095 Views: 5,016 |
|
|
Dear HS Here can we consider like this 1. 1 Sequence will be randomized for first three period and second sequence will be randomly alloted for the next three periods.... 2. If we consider so, then sequence is alocated as cross-over on the first three periods and last three periods... Is this fine? If I am wrong plese correct me...  — Cheers Bala |
|
Helmut ★★★ Vienna, Austria, 2009-01-21 14:27 (6364 d 02:24 ago) @ bala Posting: # 3096 Views: 5,127 |
|
|
Dear bala, not going into the details of your suggestion (limited time & brain capacity), I still don't get the idea why Neethu wants to run such a design. 6 periods will lead to ethical concerns (blood loss), drop-outs, etc. I you have the software/code to evaluate such a design, I haven't.  — 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 |
|
bala ☆ Bangalore, India, 2009-01-21 14:30 (6364 d 02:22 ago) @ Helmut Posting: # 3097 Views: 5,055 |
|
|
Dear HS... Even I agree its going to be a difficult one... I simply tried to design a layout for that study... — Cheers Bala |
|
neethu ☆ 2009-01-22 06:10 (6363 d 10:42 ago) @ bala Posting: # 3102 Views: 5,010 |
|
|
Dear HS/Bala, Thanks for ur replies... Its like 3 sequence and 6 period... It is in both fed and fasting conditions.... ie, 3 periods in fasting and 3 in fed condition.... — Thanks & Regards Nees |
|
bala ☆ Bangalore, India, 2009-01-22 08:04 (6363 d 08:48 ago) @ neethu Posting: # 3105 Views: 5,041 |
|
|
Hi Neethu, Then you can use this logic Three sequences (only for pandomization generation)
P1 P2 P3 P4 P5 P6Randomize sequence 1, 2 and 3 for 8 subjects and so on... — Cheers Bala |
|
Helmut ★★★ Vienna, Austria, 2009-01-23 02:21 (6362 d 14:30 ago) @ bala Posting: # 3112 Views: 5,114 |
|
|
Dear Bala and Neethu, can you please give any reason - besides money! - why you would not run two separate studies (one fasting, one fed)? Just my two cents, you will run into touble for various reasons! With an assumed drop-out rate of 5% / wash-out 3-period studies will end with 21 subjects, but your 6-period study with just 18. Are you really prepared in terms of power to loose 25% of subjects? I'm afraid, designing a study needs a little bit more preparation than writing down some random blocks.  Think about the model first! In a BE study the main effect of interest is 'treatment' (>2 different formulations, but either in fasting or in fed state). In a food effect study it's 'food' (using the same treatment). One of the main assumptions in the usual (nonreplicate) cross-over model is an Independent Identically Distribution (IDD) of effects. This assumption simply may not hold. If e.g., the variability of the reference is higher than the one of the test, we will obtain a high common variance and the test will be penalized for the reference performing badly. That's the reason why some people advocate RSABE. In a food study we see the same story. For most MR formulations we yet would expect different variabilities in fasting and fed state. Even for IR formulations food will change liver blood flow > hepatic clearance > not only the absorption, but also the elimination may be altered. Since the study is of a nonreplicate design with 2 effects (2 levels: fasting/fed, 3 levels: T1/T2/R) the assumption of a common variance IMHO is downright absurd. Metaphorically speaking you are running out of degrees of freedom... and your effects are massively confounded (no unbiased estimates)!Though it's a nice phrase 'to kill two birds with one stone' - I wouldn't call that a scientific concept. So the only chance you have is to modify Bala's last suggestion (@Bala: why didn't you use Latin squares? You randomized no T2 in P1 and no T1 in P3.) ... fasting fed... and evaluate P1-P3 and P4-P6 separately (I would use a 6-sequence Williams' design instead - but that's another story). Can you please give any reason - besides money! - why you would not run two separate studies (one fasting, one fed)? — 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 |
Ing. Helmut Schütz