Hi there,
I am building a Cox model. I have two variables that are different measures of the same thing and they are therefore correlated with each other. Both variables are strongly associated with survival.

My understanding is that it is not ideal to place correlated variables in the same model. Nevertheless, when they are in the same model, one remains highly significantly associated with survival, while the other losses significance. Does this tell me anything about these two variables? In particular, does it suggest that the variable that remains significant is a better predictor of survival?

Many thanks for any suggestions.