Hazard Ratio's for different categories

[Sophie Belt]

Dear all,
In my survival model, I want to calculate hazard ratio's of a grouped variable (change in proteinuria, divided into three groups), but then comparing two other groups (effect on blood pressure above or below median). So I want to find 6 Hazard ratio's; for change in proteinuria (in all three categories) stratified (or how you would call this) for effect on blood pressure. I have calculated the Hazard ratios without the effect on blood pressure using dummy variables: stcox dum2 dum3 (as I wanted the first category to be the reference category)
I hope anyone can help me, it is propably not too difficult, but I wouldn't know how to do this,
thank you!

[Carlo Lazzaro]

Sophie:
welcome to the list.
Please read and act on the FAQ on how to post more effectively (-dataex-). Thanks.
As far as I can get it (I'm not clear which is the dependent variable in your Cox regression model), it sounds as you were serarching for an interaction between -change in proteinuria- and -blood pressure- (I assume that both those variables are predictors).
something along the following lines might be what you're striving for:

stcox i.proteinuria##i.blood_pressure

As an aside, why wasting your time in creating categorical variables and interactions when -fvvarlist- can do it for you?

[Phil Bromiley]

There are some issues with interactions in stcox that require a little different consideration than regression or similar analyses. You might find these posts helpful.

http://www.statalist.org/forums/foru...ns-after-stcox
http://www.stata.com/statalist/archi.../msg00968.html
http://maartenbuis.nl/publications/interactions.html
http://www.statalist.org/forums/foru...interpretation