Please check my interaction test code after multiple imputation.

[Byounghwi Ko]

Dear statalist,

What I want to do is test the interaction with CKD status(factor) and IGF10(continuous) with a multiple imputed dataset.
In complete case analysis, I do that like this:

stset time_mortal, f(ev_mortal==1) scale(365.25) id(n_eid)
stcox IGF10 $C c.IGF10##i.CKD
testparm c.IGF10##i.CKD


In the multiple imputation dataset, there is no exact 'testparm' for the interaction,
so I did this code. :

mi stset time_mortal, f(ev_mortal==1) scale(365.25) id(n_eid)

mi xeq: stcox IGF10 $C c.IGF10##i.CKD
mi xeq: testparm c.IGF10##i.CKD

After that, Prob>chi2 were 0.000 for every imputed dataset, like this.

(previous this code, there were 5 sets of stcox results from m=0 to m=5.)

. mi xeq: testparm c.IGF10##i.CKD

m=0 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000

m=1 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000

m=2 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000

m=3 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000

m=4 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000

m=5 data:
-> testparm c.IGF10##i.CKD

( 1) IGF10 = 0
( 2) 1.CKD = 0
( 3) 1.CKD#c.IGF10 = 0

chi2( 3) = 351.01
Prob > chi2 = 0.0000



I wonder... can I translate it as P for interaction <0.001?
Could anyone know how to estimate the 'real' P for interaction using Rubin's rule?