While doing a simulation using Pheonix v1.3, there is an option to check “Freeze etas to Pop. estimates”. I assume the output for this simulation to have the same “IPRED” for each individual across replicates and its only CObs which will change based on the sampling from the epsilon distribution - basically , an individual prediction conditional on that subjects etas. I wanted to check this and hence used the “Add Sim Table” option to output variables “C” and “CObs” and requested 2 replicates of the dataset. The ALLPCdata.csv output gives out only CObs and the requested additional table, “Rawsimtbl01.csv” very strangely gives only one replicate for all individuals. I take it that this may be a bug, or I actually have the wrong understanding of what “Freeze etas to Pop.estimate does”. Any insights are appreciated. Best, Vijay
Dear Vijay Can you share the project. The lack of replicates may be a bug that was in the released version and fixed for the upcoming version. It I difficult to know without seeing what you did. best Serge
Hi Serge, Hear is an example project file. In fact, the pre-release v1.4 does not even have this feature in the sim/pred check. Its been removed? Vijay [file name=freeze_eta.phxproj size=1080440]Certara | Drug Development Solutions (1.03 MB)
Hi Vijay The only replicate was a bug from the current release but ahs been fixed in the upcoming version. I the new version, you do not have anymore this option of freeze eta. If you want to freeze eta to the pop, just change the random effects to 0. It will work when doing simulations. Best Serge