Specify non-inferiority margin using mixed model for repeated measurement

[DingTao]

I have a dataset structured as repeated measurement and Mixed model for repeated measurement is used for analysis. The standard code to get average change from baseline (note that here we are comparing baseline with the mean of the last 4 visits) and the associated p-value has been posted below. I wonder based on this, how can I test the non-inferiority using a margin of -1, e.g., change from baseline is greater than -1. The null hypothesis is that this change will be smaller or equal to -1. Can anyone suggest how to write SAS for this test? 

proc mixed data=mydata;
  class id visit sex(ref='F');
  model change = age sex baseline visit*baseline/ ddfm=kr fullx;
  repeated visit / subject = id type = un;
  lsmeans visit ;
  estimate 'Average' Intercept 1 age 1 sex 0.5 0.5 
           visit 0 0 0 0 0 0.25 0.25 0.25 0.25 base & basemean 
           visit*base 0 0 0 0 0 &basemean1 &basemean1 &basemean1 &basemean1 / cl;
run;