On the calibration of survival models with competing risks

2026-01-30

AI Summary

This study addresses the challenge of probability calibration in competing risks survival analysis, where existing calibration methods are often inapplicable, leading to unreliable predicted probabilities. To tackle this issue, the authors develop a dedicated calibration framework tailored to the multi-event and time-to-event nature of competing risks. They introduce two novel calibration metrics grounded in proper scoring rules, along with corresponding algorithms for calibration assessment, statistical testing, and recalibration. The proposed approach significantly enhances the reliability of predicted probabilities while preserving the model’s original discriminative performance. This work thus provides both theoretical foundations and practical tools for generating trustworthy survival predictions in competing risks settings.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Probabilistic Programming
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Survival analysis deals with modeling the time until an event occurs, and accurate probability estimates are crucial for decision-making, particularly in the competing-risks setting where multiple events are possible. While recent work has addressed calibration in standard survival analysis, the competing-risks setting remains under-explored as it is harder (the calibration applies to both probabilities across classes and time horizon). We show that existing calibration measures are not suited to the competing-risk setting and that recent models do not give well-behaved probabilities. To address this, we introduce a dedicated framework with two novel calibration measures that are minimized for oracle estimators (i.e., both measures are proper). We also introduce some methods to estimate, test, and correct the calibration. Our recalibration methods yield good probabilities while preserving discrimination.

Problem

• survival analysis
• competing risks
• probability calibration
• calibration measures
• risk prediction

Innovation

• competing risks
• calibration
• survival analysis
• proper scoring rules
• recalibration