running model developed in NLME 1.1 using NLME 1.0

Hi, I recently finalized a pop pk model using NLME 1.1. However, the company I developed the model for has not updated to NLME 1.1 and is still using 1.0. When they attempted to run the model, they get very different parameter estimates. I don’t have access to 1.0 version of NLME and is very hard for me to diagnose what’s going on. I think it may have to do with mapping in the “main” mapping panel. In NLME 1.1, I map ID to the SubjID (while leaving the sort option unchecked). However, NLME 1.0 does not have this option (ID mapping option). Instead of mapping ID to SubjID, they are sorting by SubjID. Can you tell me if this is the problem? And how can we fix it? Thank you, Dora

Dear Dora, Phoenix projects are compatible if they are created with an older version of PHX. This means that a project created in PHX NLME 1.0 will automatically set the correct settings (if they have changed) in PHX NLME 1.1 unless stated in the release notes. However, the opposite is not always necessarily true an a project in PHX NLME 1.1 might not load correctly in PHX NLME 1.0. It is difficult to diagnosed the issue you mentioned but you are correct that in PHX NLME 1.0, the subject was mapped to a ‘Sort’ . In PHX NLME 1.1 you can now map ID variables under the column ID and Sort is used for subpopulations. This would be the easiest thing to try to re-map the ID variable to the Sort. The structure of the model should also be reviewed as several new options were added in PHX NLME 1.1. Hope this helps, Ana

Dear Ana, Thank you for your reply. I did run the model with sorted by SUBJID, but the parameter estimates are still different. In addition, the run time in NLME 1.0 is signifficantly longer than in NLME1.1: 1.5 h vs 4 min, respectively. This model is not a typical model and I had to run it in the command line mode. Is this mode compatible between the two NLME versions? Could you tell me if the FOCE-ELS (which is the estimation method I used) is the same in NLME 1.0 and 1.1? I know that in 1.1, this method is with interactions. What about 1.0? Could you send me the NLME 1.1 release notes? I looked them up online and on Pharsight website and I couldn’t find them. Thanks again!! Dora

Dear Dora, The command mode is compatible with the 2 versions. You need what we call a PML license to run command mode. If you were not given this license it can be requested at no cost to the Pharsight sales group. In reference to the method FOCE-ELS the method is fundmamentally the same in 1.0 and 1.1 In particular, both versions are with interaction. The main differences are in some numerical tolerance adjutments in how numerical gradients are computed, and perhap some minor changes to improve efficiency. Please write me at ana.henry@certara.com with your email and I will sent the release notes document you requested. Kind regards, Ana