The ICPHREG procedure fits proportional hazards regression models to interval-censored data, including left-censored data and right-censored data as special cases. You can select a piecewise constant function as the baseline hazard, you can model the log of the cumulative baseline hazard function by using a cubic spline (Royston and Parmar 2002), or you can assume a discrete cumulative hazard function in which jumps are identified by the Turnbull method (Finkelstein 1986). PROC ICPHREG estimates the regression coefficients and the hazard parameters by maximizing the full likelihood function. For more information about PROC ICPHREG, see Chapter 70, The ICPHREG Procedure.