What is the best way to produce survival curves when modeling in proc phreg?

mkit8 — Wed, 22 Apr 2020 05:06:07 GMT

Hello,

I am attempting to do modeling in proc phreg by producing two models (a crude model and a fully adjusted model). I have two questions.

1) Does my code seem correct considering my objective?

2) What is the easiest/best way to produce survival curves for each model? I would like each model to have two lines (representing each of my two age groups).

Exposure = age (20-49 year olds, 50+ year olds)

Outcome = overall survival

Covariates (for the fully adjusted model) = sex, race, stage, primary_site, grade, Yost_quintile

proc phreg data = new2;
class age (ref='50+');
model Survival_months*SEERcausespecificdeathclassifi(0,8) = age / rl;
hazardratio age / DIFF=ALL;
run;

proc phreg data = new2;
class age (ref='50+')sex (ref='Male') race (ref='Non-Hispanic White')
   stage (ref='Localized') primary_site (ref='Proximal Colon') grade (ref='Grade I')
   Yost_quintile (ref='Group 5');
model Survival_months*SEERcausespecificdeathclassifi(0,8) = age sex race stage primary_site grade
   Yost_quintile / rl;
hazardratio age / DIFF=ALL;
run;

topic What is the best way to produce survival curves when modeling in proc phreg? in Statistical Procedures

What is the best way to produce survival curves when modeling in proc phreg?