Identifying Important Risk Factors for Survival in Kidney Graft Failure Patients Using Random Survival Forests

摘要: Background: Kidney transplantation is the best alternative treatment for end-stage renal disease. Several studies have been devoted to investigate predisposing factors of graft rejection. However, there inconsistency between results. The objective present study was utilize an intuitive and robust approach variable selection, random survival forests (RSF), identify important risk in kidney patients. Methods: data set included 378 patients with obtained through a historical cohort Hamadan, western Iran, from 1994 2011. event interest chronic nonreversible rejection duration considered as time. RSF method used among potential predictors Results: mean time 7.35±4.62 yr. Thirty-seven episodes were occurred. most cold ischemic time, recipient's age, creatinine level at discharge, donors’ age hospitalization. predicted better than conventional Cox-proportional hazards model (out-of-bag C-index 0.965 vs. 0.766 Cox integrated Brier score 0.081 0.088 model). Conclusion: A outperformed traditional hazard model. promising that may serve more