martin
★★  

Austria,
2018-07-23 20:17

Posting: # 19086
Views: 1,507
 

 visualization of power-law model [General Sta­tis­tics]

Dear colleagues,

I would be happy to get the forum members thoughts regarding graphical display of the outcome of a dose-proportionality assessment based on the power-law model

In case that the power law model used to assess dose proportionality consists only of fixed factors intercept and slope I would think a figure like Figure 1 and 3 in the Smith et al paper (Smith et al. , 2000. Confidence interval criteria for assessment of dose proportionality. Pharm Res. 17:1278-1283) is good way to go.

However, frequently the corresponding model includes period and sequence effect as additional fixed effects (e.g. Williams design) and I am looking for some thoughts how to adequately visualize the corresponding results (other than showing the observed data and corresponding predictions grouped by sequence and period in a lattice plot or residual plots).

Best regards

Martin
d_labes
★★★

Berlin, Germany,
2018-07-26 14:49

@ martin
Posting: # 19105
Views: 1,236
 

 visualization of power-law model

Dear Martin,

» In case that the power law model used to assess dose proportionality consists only of fixed factors intercept and slope I would think a figure like Figure 1 and 3 in the Smith et al paper (Smith et al. , 2000. Confidence interval criteria for assessment of dose proportionality. Pharm Res. 17:1278-1283) is good way to go.

I think so.

» However, frequently the corresponding model includes period and sequence effect as additional fixed effects (e.g. Williams design) and I am looking for some thoughts how to adequately visualize the corresponding results (other than showing the observed data and corresponding predictions grouped by sequence and period in a lattice plot or residual plots).

For sake of simplicity I would go with the same plots as in the first case.
Only if period effects and/or sequence effects have a greater magnitude than the power law you will have a scatter in the data which seems not fit. I would expect such an behavior in only very rare cases, if any. If you have such a case, 'correct' your data for the period and/or sequence effects and plot then.

Hope you are well.

Regards,

Detlew
Activity
 Admin contact
20,346 posts in 4,274 threads, 1,405 registered users;
online 20 (0 registered, 20 guests [including 8 identified bots]).
Forum time (Europe/Vienna): 04:41 CET

The belief that there is only one truth and
that oneself is in possession of it,
seems to me the deepest root of all
that is evil in the world.    Max Born

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5