6 replies to this topic

[Simon Davis]

Nathan, can you attach the project or email it to support if the content is sensitive so we can investigate what may cause this error message?

 

  Simon.

[Nathan Teuscher]

Simon,

 

Perhaps you could help me with my problem anyway. I developed a population PK model with a single covariate (body weight) from dataset A. Now I want to predict Cmax values for a different group of patients (dataset . Because the model includes a unique absorption model (2 different absorption components), I need to simulate a 24-hr profile for each subject to then extract a Cmax. Additionally, dataset B includes multiple doses and weights (e.g. Week 1, Week 2, Week 4, Week 8, etc.) for which I need to input a dose and weight and output a Cmax (doses can change each week).

 

1. Can I directly output a Cmax as a secondary variable from my simulation? If so, how is that done when it is not calculated?

2. What is needed in my input dataset besides subject ID, Day, Dose, and weight? Are rows with no concentration measurements needed to perform the simulation? Or with the "seq(0,24,1)" be enough to allow the simulated concentrations to be output with only dosing rows?

 

Right now I am spending a lot of time building a NONMEM-like dataset with no concentration measurements in order to do the simulation. I tried to build a dataset with just the dose information, but then I received the error noted above ... and now I'm stuck.

 

Thanks in advance,

Nathan

[serge guzy]

Dear Nathan

In order to still preserve confidentiality, can you copy your model to the workflow, go to edit as textual, send to us the model code as a first step.

If you simulate, you can also send us all the columns of your data set except the response columns and we can try to see if we can simulate.

Once we simulate we can se how to fit too.

Let us know. It is very difficult without getting a template to see where is the problem.

best

Serge

[Nathan Teuscher]

Serge,

 

I sent the workflow to pharsight support, and I think Ann has it for review.

 

Nathan

[serge guzy]

Thanks Nathan

Serge

[Ana Henry]

All, In order to close the loop on this issue, I am publishing below some of the conclusions that took place outside the forum and that could be of interest to some:

The reason for the message is from a known issue in Phoenix 1.3:

When using Simulation run option with more than one table, the table names are identical and this causes error "already an object named Raw" (QC 11892): The error message is: "There is already an object named Raw under Results" Users should try to put all information in one table. If a second table is required, users can copy and paste the model in the same workflow, and then change the table in the copy.


In your project, you only have one table but you have selected a variable as a sort. That is why some of the run modes are grayed out. The program probably should have not tried to run Pred Check in this situation but it tries anyway and thus, you get multiple simulation tables (per sort level) and the error above appears.

If you do wish to do the subpopulation analysis the only current workaround is to split the data by sort level (can use split worksheet data tool) and run a model for each level.

Please note that the defect listed above will be corrected in the upcoming version of Phoenix 1.4. In addition the tool will not allow the execution of Sim/Pred check option when sorts are selected.


Ana