Anàlisi de la supervivència per dades censurades

Abstract

This work tries to introduce us in the survival analysis for censured information. Survival analysis born with the intention of analyzing information that represent the duration between two events. With him we will want to explain the evolution throughout the time of an event. We will define the time zero, from which the durations measure up, and the time of the event of interest, that is to say, the moment in the one that produces to himself the event that we are studying. Examples where this type of study is applied are: demographic studies, industrial applications to prove the electronic reliability of components, epidemiology or clinical tests. In this work we will centre on the latter area, since there will be applied the concepts introduced to two real true databases of clinical tests. We will begin the work defining the basic concepts that will be the functions that are in use for the survival analysis. In the chapter 2 we will introduce the concept of censoring and truncation, two characteristics of the information in this type of studies. In the chapter 3 we will present the non-parametric estimators for the survival analysis: the Kaplan-Meier’s estimator for the survival function and the Nelson-Aalen’s estimator for the hazard function, principally. Also, the properties of these. Finally, in the chapter 4, they will present the tests of hypothesis for the comparison of two or populations.