proportional hazard assumption was not satisfied for a covariate

Question

[akutasame]

Hello,

I am doing multivariate Cox model for my data. I wonder what if proportional hazard assumption was not met for a covariate in the model, and stratifying that covariate did not solve the problem. Does it make sense to remove that covariate ? If I remove, then assumption was satisfied. Can I remove a covariate if it violates PH assumption?

Thanks in advance

[Kevin Blighe]

Hey, why did you include the covariate in the first instance? Is it an independent statistically significant predictor of your outcome?

[akutasame]

I fit a full model, so I include all covariates. Let say covariate x is a categorical variable and it is not significant but there are some other significant covariates. Then I wanted to check PH assumption for this full model with cox.zph() function in R. Only covariate x seems to violate PH assumption.