Log in
Register|
pash413 ★ India, 2011-01-27 12:10 (5621 d 21:10 ago) Posting: # 6499 Views: 7,282 |
|
|
Dear all can any one suggest about sample size calculation for parallel study planing on basis of pilot study results or available literature Edit: Category changed. [Helmut] |
|
pash413 ★ India, 2011-02-08 07:08 (5610 d 02:11 ago) @ pash413 Posting: # 6576 Views: 6,015 |
|
|
Weidson ☆ Brazil, 2011-02-08 19:47 (5609 d 13:32 ago) (edited on 2011-02-08 23:22) @ pash413 Posting: # 6597 Views: 5,960 |
|
|
Well, before you try to help I'll ask another question: What goal are you trying to achieve with the implementation of the pilot study ?  Edit: Full quote removed. Please delete anything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Ohlbe] |
|
Helmut ★★★   Vienna, Austria, 2011-02-09 17:27 (5608 d 15:53 ago) @ Weidson Posting: # 6603 Views: 5,969 |
|
|
Dear Weidson! ❝ What goal are you trying to achieve with the implementation of the pilot study? Me – nothing. The question came from Pash… Pilot studies are good scientific practice – especially if your information are limited and based on literature only. A nice justification is given in FDA’s guidance ( ![[image]](img/uploaded/IAsmall.png) Section III.A.2).— 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 |
|
Weidson ☆ Brazil, 2011-02-09 18:21 (5608 d 14:59 ago) @ Helmut Posting: # 6605 Views: 5,892 |
|
|
❝ Pilot studies are good scientific practice – especially if your information are limited and based on literature only. A nice justification is given in FDA’s guidance ( Hello Mr. Helmut Well, this is what many researchers still believe to be correctly. I understand that it is necessary to develop predictive ways to know if the main study will be approved or not without spending much money  . Unfortunately, "there is no other way to made a cake without breaking eggs"  . That is, if we do a pilot study and the high variability is found with a low N (intra-subject in the case of a crossover trial), then we have a study with little power to draw valid conclusions in respect of bioequivalence. Recently I did a data mining in my database (I've done many studies over 10 years experience) and the pattern found was that the ratio's of Cmax and AUC is not stable for studies with underpowered. Please also do this analysis in your bank you will be surprised ... If you want to discuss more about it I'll be happy ... |
|
Helmut ★★★ Vienna, Austria, 2011-02-09 20:29 (5608 d 12:51 ago) @ Weidson Posting: # 6608 Views: 5,903 |
|
|
Dear Weidson! ❝ ❝ Pilot studies are good scientific practice ❝ ❝ Well, this is what many researchers still believe to be correctly. A mathematician is a blind man Substitute 'A mathematician' by 'Any scientist performing a main study without prior knowledge'.  ❝ […] if we do a pilot study and the high variability is found with a low N (intra-subject in the case of a crossover trial), then we have a study with little power to draw valid conclusions in respect of bioequivalence. Well, it’s not the objective of a pilot study to demonstrate BE. I always prefer to take the uncertainty of the estimated CV into account. The smaller the pilot, the larger the penalty (i.e., the larger the main study). ❝ Recently I did a data mining in my database [...] and the pattern found was that the ratio's of Cmax and AUC is not stable for studies with underpowered. ❝ Please also do this analysis in your bank you will be surprised ... It’s rather difficult to surprise me – I can confirm your observations.  Another reason why not to perform small pilot studies. I don’t know of any statistical method which would help us in dealing with the uncertainty of the PE besides the confidence interval itself – but that’s trivial. Therefore it's always a good idea to perform a sensitivity analysis as part of the sample size estimation (i.e., treat the PE from the pilot not as “fixed” as well). Nothing shocks me. — 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 |
|
Helmut ★★★ Vienna, Austria, 2011-02-09 18:40 (5608 d 14:40 ago) @ pash413 Posting: # 6607 Views: 5,991 |
|
|
Dear Pash! ❝ can any one suggest about sample size calculation for parallel study planing on basis of pilot study results or available literature No big difference to methods for cross-over designs. Don’t forget that in parallel designs only the total variability (between + within subjects) is accessible. Essentially you have three options:
— 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