ABSTRACT
A primary end point in many medical studies is the time until the occurrence of an event such as disease progression, death, or discharge from a hospital. The statistical analysis of such data, commonly referred to as failure-time data or survival data, requires the use of special methods because the event may not yet have occurred in all subjects by the time the data are analyzed. This paper reviews standard statistical methods for analyzing survival data. We describe the Kaplan–Meier method of estimating a survival distribution and the log-rank test for assessing the equality of two or more survival distributions. We also review Cox’s proportional-hazards regression model, which is a popular methodology for assessing the simultaneous association between multiple baseline factors and survival. These techniques can be very helpful in identifying factors that influence survival time and in describing the survival experience of a population.
