Competing and Noncompeting Risk Models for Predicting Kidney Allograft Failure.

Area of Science:

• Nephrology
• Biostatistics
• Clinical Trials

Background:

• Prognostic models are vital in clinical trials and patient management.
• The impact of competing risks on model performance is understudied.
• Kidney transplantation presents competing risks: allograft failure and death with a functioning graft.

Purpose of the Study:

• To assess the performance of competing risk and noncompeting risk models for kidney graft failure prediction.
• To investigate the influence of competing risks (death with a functioning graft) on prognostic model accuracy.
• To provide a comprehensive evaluation of model performance in kidney transplant recipients.

Main Methods:

• Included 11,046 kidney transplant recipients from ten countries.
• Developed prediction models using Cox, Fine–Gray, and cause-specific Cox regression.
• Assessed model development, stability, discrimination, calibration, fit, utility, and generalizability using over 15 metrics.

Main Results:

• Both competing risk and noncompeting risk models demonstrated similar and stable risk estimates for graft failure.
• Model discrimination and overall fit were comparable across validation cohorts (concordance index 0.76–0.87).
• A trend toward improved calibration was observed with competing risk models in specific high-risk groups.

Conclusions:

• Competing risk and noncompeting risk models exhibit similar performance in predicting long-term kidney graft failure.
• The choice of model may have minimal impact on overall prediction accuracy in this context.
• Further investigation may be warranted for specific high-risk subpopulations.