# How to model multinomail logistic regression to get desired outcome?

I want to run multinomial logistic regression in SPSS.

Dependent variable: Code=0 Cognitively normal(CN) Code=1 Mild cognitive normal(MCI) Code=3 Alzheimer’s(AD)

Independent variable: Genotype.1=0 Genotype.2=1 Genotype.3=2

What I want to know is:

How much is OR (odds ratio) for CN to get MCI and how much is the OR for CN to get AD?

If I consider MCI as a reference group, then I can see the OR for CN to get MCI and AD to get MCI (I do not need to report the latter OR) so by this model, I just have OR for CN to get MCI and do not have OR for CN to get AD.

However, If I consider AD as a reference group then the model reports the OR for CN to get AD (desired) and the OR for MCI to get AD (I do not want the latter one).

See the result in the following image.2

So, how I can run the multinomial logistic regression to report the OR for CN to get MCI and CN to get AD?