I'm using Cox Proportional Hazards modeling to evaluate the effect of 5-6 baseline variables on overall survival (OS) in a breast cancer dataset. Beyond evaluating the main effects of these covariates, I'd like to assess interactions between baseline variables. I've been using the "Cross" button to add these interactions to my model (e.g. ECOG status * Age). However, I could use some help determining which resulting output is the test for interaction (which p-value should I be referencing)? Is the Wald test giving me a p-value for interaction? I'm more accustomed to seeing a "kappa" value - just want to make sure I' doing this correctly.
Thanks,
KJD
