BookmarkSubscribeRSS Feed
cindyf
Calcite | Level 5

Hi all, 

I'm testing 4 different types of treatments regarding a same disease outcome (binary variable 1/0) 

I am able to plot 4 roc curves and see relative stats (i.e. p-value, c stat, etc). 

Is there a way I could compare all these 4 roc curves and get a p value, so I could say all of them are same or not. 

Model I used is:

model outcome=bmi age gender . I used "by treatment" to generate 4 individual plots. I also used https://support.sas.com/kb/52/973.html method to overlay all 4 plots and do the pairwise comparison. I'd like to have a global stat, maybe a chi-quare test to show the p. 

Thanks so much 

1 REPLY 1
PaigeMiller
Diamond | Level 26

This allows you to compute p-values for differences of each possible pair of ROC curves. https://support.sas.com/kb/45/339.html

 

I am not aware of a way to get an overall p-value comparing all ROC curves.

--
Paige Miller

hackathon24-white-horiz.png

The 2025 SAS Hackathon has begun!

It's finally time to hack! Remember to visit the SAS Hacker's Hub regularly for news and updates.

Latest Updates

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

SAS Training: Just a Click Away

 Ready to level-up your skills? Choose your own adventure.

Browse our catalog!

Discussion stats
  • 1 reply
  • 630 views
  • 1 like
  • 2 in conversation