I fitted a coxph
model and below is the concordance details from my object named cx
.
I can see that Concordance
is calculated as 4282/(4283+1697) = 0.716
but I'd like to know:
(1) How is se=0.048
calculated?
(2) What does std(c-d)
represent and how is it calculated?
> summary(cx)
Concordance= 0.716 (se = 0.048 )
Rsquare= 0.098 (max possible= 0.852 )
Likelihood ratio test= 20.41 on 2 df, p=3.707e-05
Wald test = 21.07 on 2 df, p=2.665e-05
Score (logrank) test = 21.07 on 2 df, p=2.654e-05
> cx$concordance
concordant discordant tied.risk tied.time std(c-d)
4283.0000 1697.0000 0.0000 3.0000 576.6328