Log in
RegisterEvaluation according to design [General Statistics]
posted by Helmut 
– Vienna, Austria, 2012-01-10 19:22 (5270 d 03:24 ago) – Posting: # 7894
Views: 4,535
Dear Luv!
Yes. Bad luck.
What? You had four periods and two sequences (RTRT|TRTR).
I strongly oppose Chiku's opinion. Evaluate the study according to its design (SAS, Phoenix/WinNonlin, and R could do the job). No way to mean responses of T and R. How would you code the (pseudo-)periods? Example from EMA’s Q&A document, data set I (I removed all incomplete data); 69 subjects, 276 observations, EMA’s Method A (all effects fixed):
Note two more degrees of freedom if the observations are stacked and recoded (two periods instead of four). The PE is not affected but the CI is anticonservative. Taking the means is IMHO not acceptable at all.
❝ We did a full replicate BE study for EU submission a few days back with 40 volunteers. %ISCV observed for reference product was […] 28% […], thus we have to comply with normal 80-125% CI limit criteria, …
Yes. Bad luck.
❝ … considering the study as a two way study.
What? You had four periods and two sequences (RTRT|TRTR).
❝ Now which is the right way Statistically, Practically and Regulatory approach wise.
I strongly oppose Chiku's opinion. Evaluate the study according to its design (SAS, Phoenix/WinNonlin, and R could do the job). No way to mean responses of T and R. How would you code the (pseudo-)periods? Example from EMA’s Q&A document, data set I (I removed all incomplete data); 69 subjects, 276 observations, EMA’s Method A (all effects fixed):
PE CL (lo – hi) df CVintra
Full model 115.46 106.49 125.19 203 42.37%
Stacked and recoded to RT|TR 115.46 106.51 125.17 205 42.26%
Means of RR and TT, recoded to RT|TR 113.50 103.07 124.98 67 34.90%Note two more degrees of freedom if the observations are stacked and recoded (two periods instead of four). The PE is not affected but the CI is anticonservative. Taking the means is IMHO not acceptable at all.
—
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
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
Complete thread:
- Full Replicate BE study! Calculation Issues luvblooms 2011-12-16 16:46 [General Statistics]
![[Mix]](https://static.bebac.at/img/mix.png "open in Mix view")
- Full Replicate BE study! Calculation Issues Chiku 2012-01-10 11:19
- Evaluation according to designHelmut 2012-01-10 18:22
Ing. Helmut Schütz