Interpretation of the Hazard Ratio in a Cox regression model with time varying covariate(s)

data analysis
1

Hello

Let’s consider an exposure and an outcome that is repeatedly assessed at each unit of time (days) after surgery. The association between the current status of exposure is used to predict the occurrence of the event during the next unit of time. The value of the independent variable is update at each unit of time.

Is the HR in this example thus represent the “instantaneous relative risk” conferred by the exposure? (Considering that the proportionality assumption is true)

In this setting, is the HR and the Likelihood Ratio (LR) become one and the same?

Thanks!

William

2

Hello, William.
I have a question about the time-dependent variable, that is how to include the interaction with time for a variable that doesn’t satisfy the proportional hazards assumption in a cox model. In SPSS, there is an auto-generated time variable T_ in the Compute Time-Dependent Covariate menu. But how can we include the interaction with time using R?

Thank you.
Lin.