0
$\begingroup$

I'm a medical oncologist and had some doubts concerning the stats behind clinical trials design.

My question concerned the use of log-rank vs. univariable Cox-regression to compare the survival curves between two groups. Assumed the proportional hazard assumption is met, why i see clinical trials performing the log rank to derive the p-value, and thereafter the Cox hazard model to derive the HR and CI, couldn't it be done directly using the cox regression model? Moreover, i saw clinical trials performing stratified log-ranks for each covariate, and then using the cox regression analysis only for the multivariable model for significant covariates. Also in this case, assumed the proportional hazard assumption is met, which is the sense to perform a log-rank instead of an univariable Cox-regression?

$\endgroup$
0

0