I have a dataset of 16 patients with 10 variables. The dependent variable is "bmizpre". It is a longitudinal study with a total of 3 time-points (variable time is "point"). T. The independent variables are "gender", "drug", "bmicategory " and "diseasetype". The variable defining different patients is "ptid". I would like to get result as the mean difference (95% CI) in bmizpre for different covariates

I am interested to analyze bmizpre over time by gender, drug, "bmicategory " and "diseasetype" with random effects for subject.

My question is do I have to run different univariate model included time, the covariate, and the interaction between time and the covariate ?

This is the command that I used:


Code:
mixed bmizpre gender##c.point || ptid: point
Do I have to repeat for all covariates ?


Following is the output

Code:
 mixed bmizpre gender##c.point || ptid: point

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log likelihood = -46.530546  
Iteration 1:   log likelihood = -45.773561  
Iteration 2:   log likelihood = -45.698352  
Iteration 3:   log likelihood = -45.698038  
Iteration 4:   log likelihood = -45.698038  

Computing standard errors:

Mixed-effects ML regression                     Number of obs     =         32
Group variable: ptid                            Number of groups  =         16

                                                Obs per group:
                                                              min =          2
                                                              avg =        2.0
                                                              max =          2

                                                Wald chi2(3)      =       1.46
Log likelihood = -45.698038                     Prob > chi2       =     0.6918

--------------------------------------------------------------------------------
       bmizpre |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------+----------------------------------------------------------------
      1.gender |      .1375   .5429129     0.25   0.800    -.9265897     1.20159
         point |    .225625   .1912063     1.18   0.238    -.1491324    .6003824
               |
gender#c.point |
            1  |   -.179375   .2704065    -0.66   0.507     -.709362     .350612
               |
         _cons |    -.21125   .3838974    -0.55   0.582    -.9636751     .541175
--------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
ptid: Independent            |
                  var(point) |   1.06e-18   1.40e-17      6.07e-30    1.86e-07
                  var(_cons) |   .5940602   .3300578      .1999428    1.765042
-----------------------------+------------------------------------------------
               var(Residual) |   .5849574   .2068139      .2925356    1.169687
------------------------------------------------------------------------------
LR test vs. linear model: chi2(2) = 4.69                  Prob > chi2 = 0.0960

Note: LR test is conservative and provided only for reference.