I'm performing a survival analysis using a large, complex health survey linked to mortality outcomes. I have appended several survey years to create an unweighted sample size of approximately 200,000 observations. With weights, the number of observations is over 2.2 million. I have used PROC SURVEYPHREG to estimate adjusted hazard ratios for several categorical variables. Now I would like to check the proportionality assumption using a time-dependent covariate, as illustrated on page 9200 (page 38 of the chapter 113 pdf) of the SAS/STAT 14.1 User's Guide https://support.sas.com/documentation/onlinedoc/stat/141/surveyphreg.pdf
Here's the example from the user's guide:
proc surveyphreg data=Test;
model T*Status(0)=A X1 X2;
I have modified the code using the variables from my data and tried running it, but SAS does not generate any output. Due to the large size of the dataset, I let it run for over and hour, but SAS just said "running" and never shows anything in the results viewer.
Is there a better way to statistically test the proportionality assumption?
Try running it with a smaller subset and see if it runs or see if the example runs on your computer. What version of SAS do you have, SAS/STAT latest is 15.1? The following code will let you know what version you're running.
Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.
If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website.