THE IMPORTANCE OF SURVIVAL ANALYSIS FOR EPIDEMIOLOGICAL AND HEALTH STUDIES

Abstract

Epidemiological and health studies are highly versatile and require methodologies that suit the purposes of each research. Survival analysis is a set of methodological procedures that can be of great utility to assess outcomes and related factors over time. Although more commonly used in oncological studies, we highlight in this study its utilization in other areas such as economics and education. Through parametric and non-parametric techniques, it becomes possible to measure event occurrence and the impact of explanatory variables on the outcome. Kaplan-Meier curves are easily understandable, and Log-rank, Tarone-Ware, and Gehan-Breslow tests are important to confirm if there is a statistically significant difference between analyzed groups. We also address risk analysis measures such as hazard rate and hazard rate ratio, which are important for quantifying the probability of event occurrence in different groups. We discuss Cox and Weibull models, useful tools for evaluating the role of a set of explanatory variables in the outcome over time. The article is a theoretical essay that clearly and objectively explains the theoretical foundations for the use of this methodology and suggests greater dissemination of the subject in undergraduate and graduate education.