Proc logistic with two separate dependent variables depending on the status of the independent var?

[BWDAWG]

Hi everyone,

I am very new with using sas so part of the problem I am having is using the correct terminology to ask this question or search for solutions. Here it is. I am trying to compare the association between meeting national exercise recommendations (Dependent) based on exposure status of having CKD using proc logistic. However, the national exercise recommendations are different for each group and therefore are two different variables that I am trying to use as one, if that makes sense? Does anyone understand and know how to do this? The varaibles are CKD (independent binary 0=non CKD 1= CKD) and CKDmetEX and noCKDmetEX (dependent binary 0,1). Any help would be greatly appreciated

[PaigeMiller]

However, the national exercise recommendations are different for each group and therefore are two different variables that I am trying to use as one, if that makes sense?

No, that does not really make sense to me, two different variables generally should not be combined into one, unless you have some subject matter expertise that explains how to do this.

What makes sense to me is for each group, you predict the response based upon the appropriate x-variable for that group.

[Reeza]

You need to combine the variables into one, how requires subject matter knowledge, which I dion't have.

You could model it as meeting recommendations regardless of group and add group in as a different variable in the model, or you could use ordinal/nominal logistic regression as appropriate.

[ballardw]

Can you share some of the actual data? Only the variables involved in this logistic regression need be shown This sounds like the transformation might be simple.

Instructions here: https://communities.sas.com/t5/SAS-Communities-Library/How-to-create-a-data-step-version-of-your-dat... will show how to turn an existing SAS data set into data step code that can be pasted into a forum code box using the {i} icon or attached as text to show exactly what you have and that we can test code against.