www.cmm.bristol.ac.uk/forum

Hi all,

I am trying to fit a piecewise multilevel model on mlwin where there are four measurements taken during a treatment study. 3 measurements (baseline, mid-treatment and end of treatment) are taken during the treatment phase and one measurement during follow up. I am interested in assessing the effect of treatment (treatment group by time interaction) at the end of treatment and at follow-up. In this case, is it correct to specify the first time piece as 0, 0.5, 1, 1 and the second time piece as 0, 0, 0, 1?

Thanks a lot for your help.

Yes, the way you split the time piece [0, 0.5, 1, 1] for the treatment period and [0, 0, 0, 1] for the follow-up period is correct to assess the treatment effect at end-point and post-treatment in the piecewise hierarchical model.

Yes, your coding looks appropriate for a piecewise multilevel model. The first time piece (0, 0.5, 1, 1) captures the treatment phase with linear change, plateauing after treatment ends, and the second time piece (0, 0, 0, 1) captures change during the follow-up period. This should allow you to assess both the treatment effect and any change post-treatment effectively