CLASSIFICATION TREES FOR SURVIVAL DATA WITH COMPETING RISKS

摘要: Classification trees are the most popular tool for categorizing individuals into groups and subgroups based on particular outcomes of interest. To date, have not been developed competing risk situation where survival times recorded more than one outcome is possible. In this work we propose three classification to analyze data with multiple outcomes, using both univariate multivariate techniques, respectively. After describe method used in growing pruning risks, demonstrate performance simulations a variety model configurations, compare currently available tree-based methods. We also illustrate their use by analyzing concerning patients who had end-stage liver disease were waiting list receive transplant.Public Health Significance: Competing risks common longitudinal studies. The tree will provide accurate estimates distinct subpopulations current techniques can provide.