0
$\begingroup$

I have a question, I am using survival analysis to study the relationship between smoking and lung cancer and the covariates are BP, BMI, and gender. Exposure to smoking satisfies the ph assumption, and the covariates BP, BMI satisfy the proportional risk assumption, but gender does not. Now I've used the time function tt to get the covariate gender to meet the ph assumption.

So my question is: at this point, isn't the coefficient corresponding to exposure to smoking directly the ln(HR) I need? There is no need to account for the time effect of the covariate gender?

$\endgroup$
2
  • 2
    $\begingroup$ What is the function "tt"? What does it do? Is in an R function (if so, which package) or a Python function, or something else? $\endgroup$
    – Peter Flom
    Commented Apr 25 at 9:10
  • $\begingroup$ Thanks for the reply, I am using the survival package inside r for survival analysis, tt is a parameter of coxph to set the function of the time dependence coefficient $\endgroup$
    – just one
    Commented Apr 25 at 12:32

1 Answer 1

2
$\begingroup$

You may have a multiplicity problem hen assessing PH for many predictors. Use the R survival package coxph function and then run the fit through cox.zph which will give you a global assessment of satisfaction of the PH assumption. This accounts for multiplicities.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.