Log in
Register|
Dr Andrew Leary ★ Ireland, 2012-05-02 19:48 (5153 d 03:52 ago) Posting: # 8492 Views: 8,875 |
|
|
Dear Helmut & Co. - Greetings from sunny Ireland! [I promise - it REALLY is sunny!]  A regulatory authority has asked some questions about a BE study my client performed back in the 1990s. In particular, they have asked that: “From the ANOVA tables, influence of treatment, period and subjects (subjects within sequence & sequence) effects should be submitted and discussed by the applicant”. This was a standard two-way crossover BE study (properly designed with an adequate washout). All that is available of the statistical analysis are the summary statistics and ANOVA tables. I have had a look at these and see that for all primary parameters, the P values for treatment, period and sequence are non-significant (>0.05). However, again for all primary parameters, the P values for subject (sequence) are highly significant. Can you tell me what, if anything, this might mean? In fashioning your answer, please take into account that I am a physician and not a statistician (i.e. you will need to confine yourself to words of one or two syllables only!).  |
|
Helmut ★★★   Vienna, Austria, 2012-05-02 20:31 (5153 d 03:09 ago) @ Dr Andrew Leary Posting: # 8493 Views: 7,995 |
|
|
Dear Andrew! ❝ A regulatory authority has asked some questions about a BE study my client performed back in the 1990s. Which one? Answer might depend… Sometimes it’s useful to have a look at guidelines being in force at the time of performing the study. ❝ In particular, they have asked that: “From the ANOVA tables, influence of treatment, period and subjects (subjects within sequence & sequence) effects should be submitted and discussed by the applicant”. Oh no! They were asking for treatment and period effects as well!  ❝ I have had a look at these and see that for all primary parameters, the P values for treatment, period and sequence are non-significant (>0.05). However, again for all primary parameters, the P values for subject (sequence) are highly significant. Hhm. Sequence? I would expect to see only subject(sequence) in the tables. ❝ Can you tell me what, if anything, this might mean? Chance? In a large metastudy* significant effects were found in about ~10% of 2×2 studies (n=324), and ~5% of 6×3 studies (n=96). In other words you can expect a significant effect at the level of the test – which makes a true sequence effect (= unequal carry-over) unlikely in properly performed cross-over studies. ❝ [… ] you will need to confine yourself to words of one or two syllables only! Well, in your post you used multisyllabic words as well. The last time we met you were quite fluent…
— 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 |
|
ElMaestro ★★★ Denmark, 2012-05-02 22:10 (5153 d 01:30 ago) @ Dr Andrew Leary Posting: # 8494 Views: 8,031 |
|
|
Hi Andrew, ❝ However, again for all primary parameters, the P values for subject (sequence) are highly significant. ❝ ❝ Can you tell me what, if anything, this might mean? What you see is that Subjects contribute to "very much" to the variability of your data. This is very often so, because you have different subjects who absorb, distribute, metabolise and eliminate the drug individually and thereby differently. A big person with lots of plasma will -all other factors equal- have low AUC's and Cmax's and vice versa. And so on and so forth. Your observation is a confirmation that all men are not born equal. It does not imply that your study is invalid and it does not affect the conclusion re. bioequivalence. — Pass or fail! ElMaestro |
|
Dr Andrew Leary ★ Ireland, 2012-05-03 12:20 (5152 d 11:20 ago) @ ElMaestro Posting: # 8497 Views: 7,991 |
|
|
Dear Helmut & El Maestro Many thanks to both of you for your words of wisdom. The authority concerned is Belgium. Perhaps they have a statistician who gets excited about this sort of thing! Kind regards from Ireland (where it is back to the usual rain!)  |
|
Helmut ★★★ Vienna, Austria, 2012-05-03 13:46 (5152 d 09:54 ago) @ Dr Andrew Leary Posting: # 8498 Views: 7,973 |
|
|
Dear Andrew! ❝ The authority concerned is Belgium. Oops! ❝ Perhaps they have a statistician who gets excited about this sort of thing! Perhaps the assessor is a physician understanding statistical terms “of one or two syllables only”. Maybe he didn’t bother reading the GL and had still the 2008’s draft on his desk (lines 512–7): […] tests for difference and the respective confidence intervals for the treatment effect, the period effect, and the sequence effect should be reported for descriptive assessment. A test for carry-over should not be performed and no decisions regarding the analysis (e.g. analysis of the first period, only) should be made on the basis of such a test. The potential for carry-over can be directly addressed by examination of the pre-treatment plasma concentrations in period 2 (and beyond if applicable). Based on stakeholder's comments (pp.141–3) suggested tests were removed and are not required according to the current GL.Reading ElMaestro’s post I went back to the original question. I overlooked that the assessor asked for the subject effect. This is ridiculous. If you don’t have a significant subject effect maybe you have set BMI-limits to 21.25–21.75 kg/m², have included quadruplets+, or even clones in the study… Statistics – A subject which most statisticians find difficult — 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