Cox Regression for Survey Data: The SURVEYPHREG Procedure

The SURVEYPHREG procedure fits the Cox proportional hazards model to sample survey data. The procedure is similar to the PHREG procedure, except that it incorporates complex sample design information in the analysis. The proportional hazards regression coefficients are estimated by maximizing a partial pseudo-log-likelihood function that incorporates the sampling weights. PROC SURVEYPHREG provides design-based variance estimates, confidence intervals, and tests for the estimated regression coefficients. For more information about PROC SURVEYPHREG, see Chapter 97: The SURVEYPHREG Procedure.