**From dsmws6.dnet.sandoz.com!jia_x@sandoz.com Wed May 3 14:53:51 1995
**

Hi, my problem with NONMEM is as follows:

I want to simulate concentration data by first simulating for each subject several values for the time variable T, and then simulating the concentration CP.

For example, I assume that the time variable, T, follows a Normal distribution with mean=6 hours, and std=2 hours, say. Then I worry about the order of the simulated values of T within each subject. For, they may not be in the increasing order. Is there a way in a PRED block to direct the simulated values of T to be in the increasing order?

Thanks for help.

****

**From pascal Wed May 3 17:06:45 1995
**

I do not see very well the interest of simulating time as a normal variable, however we may imagine that the differences between some nominal time as given in the protocol, and the actual times

are normally distributed.

Anyway, I think it is not possible in one step within NONMEM, since NONMEM does not offer any sort facilities. So you will have to do this in 3 steps: 1/ simulate times within NONMEM, or using another software, 2/ Sort the simulated times outside NONMEM, 3/ Simulate the DV using the sorted simulated times.