- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi
I have some difficulties in calculating differences in time dependent AUC using Proc Phreg (SAS 14.3). I have only been able to calculate the estimates and the confidence intervals and not the p-values using this equation:
proc phreg data=survival_ph plots(overlay=individual)=roc(tick)
rocoptions(method=ipcw(cl iter=50 seed=20918) at=12 auc aucdiff);
model ITTtime*ITTdeath(0)=diabetic age Albumin0 nPNA0 GFR0
/roclabel='Full Model';
roc "diabetic age Albumin0 nPNA0" diabetic age Albumin0 nPNA0;
roc "diabetic age Albumin0" diabetic age Albumin0;
roc "diabetic Albumin0 nPNA0" diabetic Albumin0 nPNA0;
roc "diabetic Albumin0 nPNA0 GFR0" diabetic Albumin0 nPNA0 GFR0;
run;
I sincerely hope that someone can help med finding a way to show the p-values.
Thank you in advance
Best regards
Alexander
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
p-values are for a statistical test. What test are you trying to perform?
If you want to compare the submodels listed on the ROC statements, then look at the example in the PHREG doc for using the CONCORDANCE= option. The syntax is PROC PHREG CONCORDANCE=UNO(DIFF), which tests the pairwise differences between the specified submodels.