Interpretation of mixed model results

#1
Hello all,

I need some assistance with interpreting the results of a mixed model and putting up textual conclusions.

I ran a mixed model with a continuous variable Y, two fixed factors A and B, their interaction A*B, and a random factor ID. Important to say, factor A is the one I am most interested in (it's the new treatment vs. 2 control groups).

The fixed factor and the interaction were all statistically significant (P values of 0.0001,0.0004 and 0.0002 respectively).

The lsmean (least square means as appears in SAS and JMP) of level 1 of factor A (the new treatment) is A1=0.21. In addition, A2=2.6 and A3=3.9.

The lsmeans of the interactions are:

A1*B1=0.37

A1*B2=0.05

A2*B1=1.16

A2*B2=4

A3*B1=2.5

A3*B2=5.2

The covariance parameter estimates are: ID (random effect): 1.1 , residual=1.18

You can see this in the attached figure

Lower values are better, the middle group is the treatment. The blue line is B2 and red line is B1.

Apart from the left red (A2*B1), the treatment differed from A2 and A3 significantly (I have the difference estimated and CI), however, I did the Tukey adjustment, and I think I should have done Dunnett ?

Thank you !