Hazard regression is the procedure for modeling the hazard function which does not depend on the suppositions made in Cox's proportional hazards model, namely that the log-hazard function is the additive function of time and the vector of covariates both. In this approach, spline functions are frequently used to model the log-hazard function.