I am running cox-regression in SPSS and have 'age-quartiles' with values 1,2,3,4 as a variable. How do I calculate risk for disease 'D' for a said predictor 'P' based on these quartiles keeping the first age quartile '1' as reference group? Age in this case is not a predictor in the analysis.
I think this is the same question as this one.
If you specify your age quartiles variable as a strata or stratification variable, then the relative risk estimates for the age quartiles as a function of your predictors are all the same. All that will differ will be the baseline functions.