Cox proportional hazard model is one of the most common methods used in analysis of time to event data. The idea of the model is to define hazard level as a dependent variable which is being explained by the time-related component (so called baseline hazard) and covariates-related component. Model is based on several restrictive assumptions which need to be carefully verified before interpretation of parameters estimates. One of them is the assumption of proportional hazard which results directly from the model formula and means that hazard ratio needs to be constant over time. However, if this assumption is violated, it does not necessarily prevent analyst from using Cox model. The current paper presents two ways of model modification in case of non-proportional hazards: introducing interactions of selected covariates with function of time and stratification model. Both of them are easily applicable with the use of PHREG� procedure� in� SAS®. The paper consists of introduction, two sections dedicated to the methods of non-proportional hazards handling and conclusions. References, acknowledgements and contact information are included at the end of this article.