Case Study in Cox Regression
Consider the randomized trial of estrogen for treatment of prostate cancer60 described in Chapter 8. Let us now develop a model for time until death (of any cause). There are 354 deaths among the 502 patients. If we only wanted to test for a drug effect on survival time, a simple rank-based analysis would suffice. To be able to test for differential treatment effect or to estimate prognosis or expected absolute treatment benefit for individual patients, however, we need a multivariable survival model.
KeywordsContinuous Predictor High Order Factor Shrinkage Coefficient Differential Treatment Effect Scale Schoenfeld Residual
Unable to display preview. Download preview PDF.