Bioequivalence and Bioavailability Forum • Not positive about that

Not positive about that [Nonparametrics]

posted by Helmut – Vienna, Austria, 2010-09-16 21:52 (4963 d 07:38 ago) – Posting: # 5921
Views: 19,948

Dear Martin!

❝ I spend some time on thinking how to compare tmax and have some doubts that using the difference between average values (e.g. medians, Hodges-Lehmann or Harrell-Davis estimate) is appropriate as the effect size in tmax between test and reference depends on the sampling times used.

Well, you know better than I do that we are comparing not medians, but the median of ranked differences - that's the HL-estimator. Of course the comparison depends on the sampling schedule. But: keep in mind that t_max (and <nitpicking>C_max</nitpicking> as well) are surrogates of the rate of absorption k_a - which we cannot access by NCA. For an example see this post.

❝ [...] I would suggest using relative effects...

❝ Definition (taken from this presentation http://www.biopharmnet.com/doc/2010_05_20_webinar.pdf )...:

Heavy stuff to swallow for an amateur. :cool:

❝ "The relative effect is a a measure of how often a random subject receiving treatment X will outperform a random subject receiving treatment Y. It can be interpreted as a probability that a randomly selected patient in the control reveals a smaller response value than a randomly selected patient in the treatment group"

I'm not happy with this definition.

❝ $Analysis.of.relative.effects

❝ Comparison rel.effect confidence.interval t.value p.value p.perm

❝ 1 p(RT,TR) 0.633 [ 0.354 ; 0.844 ] 0.7804535 0.435 0.41

❝

❝ $Wilcoxon.Test

❝ Comparison rel.effect p.value

❝ 1 p(RT,TR) 0.633 0.4345742

❝

❝ what do you think about using relative effects for comparing tmax values between test and reference?

I'm not familiar with the coding. The problem IMHO comes from the results given as ratios. Commonly for t_max a linear model is postulated (besides PK theory think about the way we speak: 'The maximum concentration of treatment A was observed one hour earlier than the one of treatment B' – not 'The time point of the maximum concentration of treatment A was 63.3% of treatment B's.'). Now if we get a percentage, to which location parameter should we apply it? To the median of the reference – or what?

Interesting to read the presentation, page 90:

The choice of the difference-based relevance threshold delta is nasty, use ratio-to-control instead.

Nasty or not, I don't think that regulators are happy with a multiplicative model. And so am I.

—
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:

tmax in case of ties: R vs. R vs. SAS d_labes 2010-09-10 12:05
- tmax in case of ties: StatXact, Phoenix, and... Helmut 2010-09-10 18:28
  - Ties, no ties, ties, no ties ... d_labes 2010-09-13 13:21
    - simulation martin 2010-09-14 08:33
      - simulants of the world unite d_labes 2010-09-15 10:14
        
        simulants of the world unite martin 2010-09-15 14:53
        
        simulants of the world unite Helmut 2010-09-15 16:17
        
        simul ants with no ties d_labes 2010-09-16 11:19
        
        simul ants with no ties Helmut 2010-09-16 14:01
        
        two different approaches martin 2010-09-16 15:35
- evaluation of tmax: use of relative effects? martin 2010-09-16 18:04
  - Not positive about thatHelmut 2010-09-16 19:52
    - Not positive about that martin 2010-09-16 21:00
      - Stupidity Helmut 2010-09-17 13:10
- tmax in case of ties: R vs. R vs. SAS Jack 2010-09-20 14:03
  - R packages Helmut 2010-09-20 14:20
    - R packages code d_labes 2010-09-27 11:27
    - R packages Jack 2010-09-27 16:14
  - Is Exact exact? d_labes 2010-09-27 09:53