Outcome Risk Modeling for Disability-Free Longevity: Comparison of Random Forest and Random Survival Forest Methods

Key Points

• Random Forest and Random Survival Forest show comparable discrimination and calibration in modeling outcomes.
• Both models provide insights into risk modeling but RSF does not consistently outperform RF in accuracy.
• Analysis focuses on understanding when time can improve existing outcomes risk modeling techniques.
• Further studies in various clinical trial cohorts are necessary to validate these findings and enhance prediction models.