Longitudinal Data (Continuous Outcome) - Unequal follow-up intervals.

[ntilton]

I have a dataset where the (continuous) outcome is a weight-for-length z-score (WLZ) for children presenting at a growth & nutrition clinic. The problem is that there was no set follow-up period; appointments were scheduled at the discretion of the clinician. Consequently, while some children may have been measured at two-month intervals, others might have been measured at one or three month intervals, or even 6.

I know that PROC TRAJ can handle variable follow-up measurements, but I dont want to treat WLZ as a latent factor. From what I know of PROC MIXED, the follow-up times must be fixed.

I'm not entirely familiar with GLM or GENMOD. Is one of these procedures appropriate?

Thanks