This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Introduction to Cox Regression for censored survival data Intercept is different from previous model, because here it is function of time, when all independent variables are 0. Example • Treatment is bad for you • Hazard of death is higher in treatment group • When hazard function is low in baseline, it is also low in treated group • It is always eβ times higher in treated group as it is in baseline • At any time, the risk is twice as high • They are proportion to each other, not parallel to each other • If we put in log scale, the curves are parallel to each other • Log of hazard function in treatment group exceeds log of hazard function of baseline by β Relationship to survival function Picture Cox regression does not require any particular type of hazard function, though in this example, the hazard function is constant over time. Estimates and Confidence Interval Hypothesis test In a large sample, score test equals log rank test In STATA Fit cox regression model • Previously have stset command • Hazard ratio has treatment code higher value in numerator and lower value in denominator Second Example Continuous Linear X High value of Karnofsy performance status will have low hazard function and β <0 In STATA The person with 1 unit higher Karnofsy score has 0.97 of the hazard compared to score has 0....
View Full Document
This note was uploaded on 02/09/2012 for the course STAT 513 taught by Professor Barbaramc.knight during the Spring '11 term at University of Washington.
- Spring '11