The Cox proportional hazards analysis in words: examples in the renal transplantation field

When renal transplantation was still in its infancy, failures were more prevalent and successes could be directly derived from facts and events. Because results have improved dramatically over the last decades and many factors have seemed to be involved in these continuously improving results, it is difficult to ascertain the individual contribution of each factor. Survival analysis is the appropriate method for evaluation of factors influencing results of renal transplantation. In this overview, two different methods for survival analysis are compared and described. The Kaplan-Meier analysis is the oldest and most frequently used in renal transplantation epidemiology. Important shortcomings of this method are described and substantiated with examples. The Cox proportional hazards (PH) analysis was developed in 1972 by Sir David Cox. With this multivariable analysis it is possible to identify those variables that influence the rate of failure. With this method, the influences of all other variables in the model are taken into consideration, and adjustment for interaction with other variables or with time can be made. In this article, the Cox analysis and the statistical terms that go with it are described in words and examples are given. In a complex, observational study concerning a multifactor-influenced population such as the renal transplant population, the use of the Cox model is mandatory to unravel the influences of the different variables on the failure rate.