Techniques to predict survival outcomes in nephrology: basics and traditional methods

Predicting survival outcomes is 1 of the most crucial tasks that nephrologists face. On an individual level, predicted survival times affect all aspects of decision-making, including the initiation of dialysis or the consideration of transplantation. On a larger scale, interventions and therapeutics are often based on the perceived probability or survival within particular cohorts. An understanding of the concepts involved in survival analysis helps clinicians appreciate the complexity involved in predicting survival and the assumptions inherent in this process. This is the first of a 2-part series of articles on the topic of survival analysis. This article will outline the basic concepts underlying survival analysis and the traditional methods used to determine time-to-event outcomes. The second article will encompass novel methods for predicting time-to-event outcomes and strategies to evaluate model performance. To facilitate hands-on learning and practical implementation, the R code used to generate these analyses is provided in the Supplementary Code provided in the Supplementary Material, with accompanying instructions, allowing readers to apply these methods to their own datasets.