Fundamentals and applications of survival analysis for health research

Abstract

Survival analysis is a statistical method that focuses on the time it takes for an event of interest to occur. It combines time, which is a continuous variable, with the occurrence of the event, a dichotomous variable; in addition, its distinctive feature is the presence of censored data. The Kaplan-Meier method is a nonparametric test that estimates the probability of survival over time, which is calculated each time an event occurs. The log-rank test is used to compare survival patterns between independent groups. Cox proportional hazards regression is the most widely used multivariate model in survival analysis; it evaluates predictive factors and estimates the Hazard Ratio as a measure of association. The use of traditional models requires assumptions such as proportional hazards and non-informative censoring, and when this criteria is not met, researchers must choose appropriate techniques according to their objectives, population and resources. Options include Bayesian models, stratified, time-dependent covariates or artificial intelligence techniques; the latter offers an alternative for modeling complex scenarios, handling large volumes of data and overcoming the limitations of conventional methods.