Categorical analysis

Hello,
I've some analysis problems.
Could U help me please?

So my problem is
I 've:
a binary variable reponse D= affected or not affected.
a categorical variable E= 0,1,2
a variable for adjustement age.
a categorical variable group= 0,1,2,3.
I want to estimate the risk of D for E adjusted for the age and test
the homogeneity of the odds ratio (association in each group or not).
I use SAS 9.0 for the analysis
I ve got a small sapmle.
For exple:
group 0
aff unaff
E=0 2 3
E=1 4 8
E=2 3 0

group 1
aff unaff
E=0 8 7
E=1 10 12
E=2 21 33

group 2
aff unaff
E=0 2 0
E=1 24 8
E=2 12 0


group 3
aff unaff
E=0 22 18
E=1 5 4
E=2 2 2

I do:
proc logistic descending ;
class E (ref=first) group (ref=first);
model D= E group E*group age;
run;
and i Do
proc logistic descending ;
class E (ref=first) group (ref=first);
model D= E group age;
run;

And I compare -2Log L for the test of homogeneity.

Could I do this or are there other way to do the analysis with small
sample.
Could I estimate the Mantel Haenzel odds ratio ajusted for the age and
do the Breslow day test knowing that E has 3 categories.

Thanks U very much for yours help.
Sincerely.

.

Loading