08-31-2016 11:58 PM
I am using SAS 9.4 and I would like to use a Cox model to model an outcome cluster (multiple, unordered events per subject of different types) in survey data. I have located SAS documentation on how to model multiple events using PROC PHREG, but I have not found anything on how to implement this in SURVEYPHREG. My primary concern is whether I can obtain a corrected variance estimate in SURVEYPHREG. Is this possible in SURVEYPHREG? It appears that covsandwich is only an option in PHREG.