You can fit stratified proportional hazards models by using a STRATA statement. Supported models include the piecewise constant hazards model (default) and the cubic spline model.

You can fit models to data that might be left-truncated by specifying the ENTRY= option in the MODEL statement. The variable that you specify is treated as the left-truncation time for each subject of the data set.

You can produce hazard plots by specifying the PLOT=HAZARD option in the PROC ICPHREG statement. By default, a plot of the baseline hazard function is created. If you specify an input data set by using the COVARIATES= option in the BASELINE statement, the procedure plots the estimated hazard functions conditional on the input covariate patterns.

The Lagakos and deviance residuals (Farrington 2000) are available for assessing the fitted proportional hazards model. You can display these residuals against their observational index in the input data by default (INDEX option) or against the estimated linear predictor by specifying the XBETA option.