I have included a minimal example file of the problem I’m having fitting a PKPD dataset with resets.
I fit a two compartment population PK model with random effects on V and CL. I copied this model to run sequential population PKPD. I linked the Eta from my PK model to the Random Effects in my PKPD model. Initial estimates are functioning normally and look good, but when I run the fit, my resets are not recognized and each animal is given a sum of all the doses.
Am I doing something wrong, or is this a bug? If so does anyone have a workaround? Thanks!
For some reason your Concentration was not mapped in your workflow (not sure why this happened). I recreated the model (see under Edited workflow) and its working fine.
Thanks so much for checking this out! The example you recreated is not employing sequential PKPD though. Once you click the “sequential” box concentration is not mapped and each animal receives all doses.
I think that your PKPD object is fine and that the Reset function works correctly. You can check the Reset function by incorporating your Dose_Group as covariate. When you add a table to the output with DoseGroup,C,E you will see that each DoseGroup receives their dose after Reset:
Thank you very much for your attention Bernd. I still don’t fully understand how, but it looks like adding the DoseGroup covariate changed the model behavior and fixed the output plots somehow. For instance IPRED vs IVAR looks like this with the covariate:
Can you confirm that there was some oddity when not using the covariate? I am on v8.1 for forseeable future so just want to check as I cannot seem to run your object:
You might think this is kind of odd, but what you look at are just general plots from general output. The basis for most diagnostic plots is the Residuals worksheet (under Results). Since there is no information given about the different dose groups, the plots cannot display any differences. If you include a differentiator like the dose group it can display the differences. If you don’t trust the plots, I would recommend to generate tables, e.g. Add Table under Run Options and extract the crucial information. You can analyze the table output using plots or stats as you prefer.
Thanks, I can confirm that the Dosegroup covariate is mapped properly.
A couple other people running 8.1 received the same error when running your example file.
I tried recreating your solution: I added the covariate DoseGroup, mapped it in Main and output a table including the DoseGroup, and I still received same error. Is there something I’m missing, perhaps when defining the covariate? It is defined as continuous and is not assigned to any model parameter.
My Random Effects look like this if this could be the issue: