Log in
Register|
yjlee168 ★★★   Kaohsiung, Taiwan, 2011-04-20 11:47 (5536 d 14:39 ago) Posting: # 6926 Views: 5,372 |
|
|
Dear all, Recently I was asked by one of reviewers here about the criterion to applying two-one sided test (TOST) to a 2×2×2 crossover BE study. The reviewer said that TOST should not be applied to the BE study when the intra-subject CVs were less than inter-subject CVs. I cannot find any ref. for this. Is this true? If yes, then what will you suggest? Thanks in advance. — All the best, -- Yung-jin Lee bear v2.9.6:- created by Hsin-ya Lee & Yung-jin Lee Kaohsiung, Taiwan https://www.pkpd168.com/bear Download link (updated) -> here |
|
ElMaestro ★★★ Denmark, 2011-04-20 12:22 (5536 d 14:04 ago) @ yjlee168 Posting: # 6927 Views: 4,517 |
|
|
Dear yjlee168, never heard of such a thing. — Pass or fail! ElMaestro |
|
yjlee168 ★★★ Kaohsiung, Taiwan, 2011-04-21 15:06 (5535 d 11:20 ago) @ ElMaestro Posting: # 6932 Views: 4,499 |
|
|
Dear ElMaestro & Helmut, Thank you both for quick response. When I just got the question, I tried to read the question several times to make sure what I interpreted the meaning of the reviewer correctly. And now I am pretty sure that's exactly what the reviewer meant. YJ — All the best, -- Yung-jin Lee bear v2.9.6:- created by Hsin-ya Lee & Yung-jin Lee Kaohsiung, Taiwan https://www.pkpd168.com/bear Download link (updated) -> here |
|
ElMaestro ★★★ Denmark, 2011-04-21 15:18 (5535 d 11:08 ago) @ yjlee168 Posting: # 6933 Views: 4,428 |
|
|
Dear yjlee168, I think that TOST will always be OK as long as the residuals are following the normal distribution. This is one of the essential assumptions (we generally don't even test it because we are all sheep), and it has nuffin to do with the magnitudes of between- and within-errors. — Pass or fail! ElMaestro |
|
Helmut ★★★  Vienna, Austria, 2011-04-20 14:39 (5536 d 11:47 ago) @ yjlee168 Posting: # 6928 Views: 4,511 |
|
|
Dear Yung-jin, that’s weird. We perform a cross-over because generally CVintra < CVinter ⇒ less subjects are needed. There are (very few!) cases were CVintra > CVinter; in these cases a parallel design would be the better choice from an ethical point of view. TOST is independent from the design. Maybe the reviewer mixed up intra and inter? However, the applicability of TOST is independent from both the size of the CV(s) and the order (i.e., which one is larger). — 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 |
Ing. Helmut Schütz