1 reply to this topic

[Segarra]

Hi all,

 

I am running some data using a 1 compartment or a 2 compartment models un PopPK module with a FOCE ELS algorithm. However, when I get the theta values they lack the CV%, the CI and the Stderr values while they are returned with the 1 compartment model. If I choose to run the QRPEM algorithm, then all those parameters and calculated for both models.

1. Is there a reason why I cannot get the CV, CI, etc with the FOCE ELS for 2 compartment model?
2. In order to choose between both models, would AIC be affected whether FOCE ELS or QRPEM algorithms are used (for both)?

Thanks

 

Ignacio

[Simon Davis]

Hi Ignacio, we had a similar discussion here,

https://support.cert...m-theta/?p=4494

 

you can also try to accept all your final estimates and then re-execute FOCE-ELS, sometimes you will get SE's then.

Regarding your AIC question, the maximum value of the likelihood function for the model will probably be different between the two algorithms, so I although this may guide you, i don't think you can use it as quantitative proof when comparing across different model engines.

 

   Simon

Edited by Simon Davis, 19 April 2021 - 10:39 AM.