How to adjust data of confounding variables adjusting diabetes to BMI and age

[FSNWRM]

I am doing a study of pregnant diabetes women. There are 3 types Type 1 , 2 and gestational. Then various ages and bmis. Outcomes study include type of delivery , gestational weeks for delivery, diet or insulin use, etc. Baby stats are birth weight, admission to NICU, apgar score , congenatal anomolies etc. How to go about analyzing these data. I have compared the groups using T test, anova .
How would you compare the 3 diabetes group across 3 outcomes miscarriage , c section or natural labour.
How do you adjust diabetes to the variables BMI and age?
Any help would be useful
thanks in advance

[Karabiner]

You could use a multinomial regression (which is like logistic
regression but with more than 2 outcome categories), with
diabetes type, bmi and age as predictors.

Kind regards

K.

[FSNWRM]

But how do i adjust diabetes to age and BMI. What can we use in SPSS?

[FSNWRM]

Thanks for your reply . I was told to adjust diabetes to age and BMI.Does that apply only to odds ratios or even to t tests? how do i go about doing this adjustment? Any help

[Karabiner]

You could use a multinomial regression with diabetes type,
bmi and age as predictors. This will do the adjustment. If
you have no idea what adjustment, is you might perhaps
want to ask the persons who gave you the order to supply
you with more information.

Kind regards

K.

[FSNWRM]

thanks i got it now.
Another question i got about another study i have a 2 by 3 table of data. continous data . Would we use 2 way anova to calculate significance across the group. 3 divisions like improved , died, went against medical advice and then 2 rows with low anion gap and the next group is high anion group( continous variable) Is 2 way anova correct for this

[Karabiner]

What do the cells contain, number of subjects? Or did you measure
some variable and the cells contain means of that variable?

Kind regards

K.