Proc mianalyze

[Jep]

Hello,

I am trying to compare two independent group means using a t-test. The variable I am trying to compare between the two groups is difference in weight change six months after an intervention between two groups who underwent different interventions. Each individual has a baseline weight measurement but some individuals are missing the weight variable for the sixth month follow up period, so I performed  multiple imputation for the missing values. How do I use proc mianalyze to pool the results of the multiple  t-tests from the imputed data sets?

Here is what I have so far.

Thank you for your help.

/* imputation phase*/

proc mi data =b; nimpute=10 out = mi_b;

run;

/*Analysis phase*/

proc ttest data =mi_b;

var v6wdiff;

class group;

by _imputation_;

ods output parameterestimates =a_b;

run;

quit;

/* pooling phase*/

Proc mianalyze data =a_b;  (This is where I need help with.)

run;

[SAS_Rob]

Picking up from the Proc TTEST step, the code below should combine the means and differences.

proc ttest data =mi_b;

var v6wdiff;

class group;

by _imputation_;

ods output statistics=ttest_ds;
run;

proc sort data=ttest_ds;
by class _imputation_;
run;
proc mianalyze data=ttest_ds;
by class;
modeleffects mean;
stderr stderr;
run;