Nagelkerke’s R2 and C-statistics in using proc lifereg

[redspring]

Happy Tuesday Everyone!

I have built a Weibull survival model using proc lifereg and I now need to validate it using the bootstrap approach.

As per the study protocol I need to calculate the Nagelkerke’s R2 and C-statistics.

In validation I continue needing to use proc lifereg. I did some search for the option for c statistics for proc lifereg and it seems proc lifereg does not have this option.

First of all is it right proc lifereg does not support the option for c statistics? If it is true what alternative I may use to fulfill the same purpose?

Can anyone please advise?

Thank you in advance.

[sbxkoenk]

Hello,

Look in below paper for 'modified' concordance statistic (or C-statistic) in survival models.

In the context of survival analysis, various C-statistics have been formulated to deal with right-censored data (Harrell 1986; Uno et al. 2011).

Paper SAS462-2017
Evaluating Predictive Accuracy of Survival Models with PROC PHREG
Changbin Guo, Ying So, and Woosung Jang, SAS Institute Inc.

https://support.sas.com/resources/papers/proceedings17/SAS0462-2017.pdf

BR,

Koen

[redspring]

Thank you for sharing that paper.

I use proc lifereg. It seems it does not support the c statistics.

Any further comments?