pre post zero inflated model?

[telligent]

I understand that nlmixed is required for longitudinal zero inflated models. But given that I only have 2 time periods, will I be fine in genmod? Can I still put time in the repeated statement?

[StatDave]

GENMOD cannot fit a zero-inflated model with repeated measures. 

[telligent]

I realize that glimmix and nlmixed don't do it either.  Is it because it is impossible or non-sensical or just because it is hard to difficult program? 

[StatDave]

You can fit zero-inflated models using PROC NLMIXED as shown in this note. You could add a RANDOM statement in that procedure. 

[telligent]

Is it like the random residual in glimmix or a g-side random effect? or can it be used both ways?