- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi. i'm using 'proc phreg' to perform time-dependent survival analysis.
I'm trying to get harrell's c index, using 'concordance' option. like this.
proc phreg data=test concordance plots=roc;
class variable1(ref='1') variable2(ref='1') variable3(ref='1')
model (start, stop)*censored(0) = variable1 variable2 variable3/TIES=
EFRON RL;
RUN;
However, 'Neither concordance nor ROC analysis is performed with the START/STOP specification' appears in log screen.
Results of survival analysis came out normally, but 'harrell's c index' didn't came out.
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi @km0927,
It seems that there are limitations in PROC PHREG regarding ROC analyses (and hence the c statistic) under certain circumstances, including time-dependent analyses (of some sort): see this thread:
and perhaps this unanswered question: https://communities.sas.com/t5/SAS-Statistical-Procedures/How-to-create-ROC-curves-using-PROC-PHREG/....
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi @km0927,
It seems that there are limitations in PROC PHREG regarding ROC analyses (and hence the c statistic) under certain circumstances, including time-dependent analyses (of some sort): see this thread:
and perhaps this unanswered question: https://communities.sas.com/t5/SAS-Statistical-Procedures/How-to-create-ROC-curves-using-PROC-PHREG/....
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
thanks. maybe SAS doesn't support 'Harrell's c index concordance option' in time-dependent survival analysis....