Survival analysis techniques in clinical research.

In clinical research the outcome of interest has increasingly become the duration until an event occurs. That is, an analysis of the time until an event occurs. Such events include the time to respond to treatment, relapse-free survival time, time to death, time to device failure, and time to regain mobility. Since traditional statistical methods cannot be used to examine such outcomes, survival analysis techniques were developed and have proven to be very useful for studying these outcomes. This paper aims to introduce and explain the basic concepts of survival analysis. Most survival analyses seen in the literature utilize Kaplan-Meier plots, logrank tests, and Cox (proportional hazards) regression. We will discuss the background and interpretation of these methods. In addition, we demonstrate how to produce survival curves, and how to quantify and test survival differences between two or more groups of patients. An increased understanding of results from survival analyses will aid clinicians and clinical researchers. Furthermore, utilizing survival analysis techniques may allow researchers to perform more powerful studies.