Modelo de Cox de riesgos proporcionales

Abstract

[en] The objective of this paper is to study the Cox model of proportional risks and apply it to real data to build a model adjusted to them. First, the basic functions necessary for the analysis of survival, as well as their properties, are introduced. The censorship and truncation of our data is defined below, phenomena very common in this type of studies given the nature of the data studied. The next chapter is about the Cox model, how it is defined and its main characteristics. It is explained which parameters compose it and the method of best estimate. The chapter ends with a study of Cox residuals, as well as its importance when it comes to seeing the validity of the adjustment. Finally, we apply the Cox model to some data on patients with multiple myeloma. From the data of 48 subjects, an adjusted model is elaborated, the residuals are analyzed and it is checked if our adjustment is good. To conclude, some small conclusions are drawn about how each variable influences the risk of death.