Hello SAS Support community,  I'm trying to model change vs baseline in visual acuity, assessed on both eyes on multiple timepoints, at 12, 18 and 24 months.    Variables involved:  treatment = random treatment assigned in a parallel design study 1:1 active drug vs placebo PT_ID = subject ID side = Can be right or left, it identifies the eye within the subject VISIT = ordered visit ID chg_VA = Change from baseline in Visual Acuity bsl_VA = Visual Acuity at baseline     proc mixed data=ds; class side PT_ID treatment (ref=first) VISIT sex ; model chg_VA = treatment VISIT sex age bsl_VA treatment*VISIT BL_VA*VISIT; repeated VISIT / subject=side(PT_ID) type=un r rcorr; lsmeans treatment treatment*VISIT / diff; random PT_ID ; run;   My main concerns is about the random statment. Should I specify the relation with side here? Should I use a subject option?  While for the repeated statment, with the subject=side(PT_ID) I'm specifing the hierarchical structure. But I'm not 100% sure if it's correct or should be viceversa.   Are there any other comments? What are the diagnostics you would suggest producing to  check validity of the model?    Many thanks for your help.  Dani     ... View more