Dear all,


I am working with multiple imputed data from a RCT with an intervention and a control group. I want to create the table of baseline characteristic. Thus, I would like to calculate percentage, sample size and p of each group for different variables such as sex, age, depression at baseline...


Reading in some post i have found the following code:

mi estimate, post: proportion CIDI_PUNTUA_pre_new, over(CON_INT)

Multiple-imputation estimates Imputations = 50
Proportion estimation Number of obs = 3,062
Average RVI = 0.7613
Largest FMI = 0.5186
Complete DF = 3061
DF adjustment: Small sample DF: min = 165.43
avg = 278.13
Within VCE type: Analytic max = 390.83

_prop_1: CIDI_PUNTUA_pre_new = 0
_prop_2: CIDI_PUNTUA_pre_new = 1

Control: CON_INT = Control
Intervencion: CON_INT = Intervencion

--------------------------------------------------------------
Over | Proportion Std. Err. [95% Conf. Interval]
-------------+------------------------------------------------
_prop_1 |
Control | .9465149 .0081097 .9305029 .9625268
Intervencion | .9434571 .0072748 .9291544 .9577598
-------------+------------------------------------------------
_prop_2 |
Control | .0534851 .0081097 .0374732 .0694971
Intervencion | .0565429 .0072748 .0422402 .0708456
--------------------------------------------------------------

It is really useful, but I would like to know if there is a stata command to calculate n (sample size) and p value with multiple imputed data divided by intervention and control group (t test). I have 50 imputations.

If p values is not possible, I would be happy only with n (%) as well.

Thanks you so much.

Best regards,
Irene