A Covariance-Based Penalty Estimator for Model Assessment With Censored Data

Area of Science:

• Statistics
• Biostatistics
• Machine Learning

Background:

• Model selection and assessment are key in statistical analysis.
• Covariance-based penalties estimate optimism in prediction errors.
• Existing methods primarily address uncensored data, leaving a gap for time-to-event data.

Purpose of the Study:

• To estimate optimism in survival prediction models using the Brier score.
• To extend covariance-based penalty methods to time-to-event data with right censoring.
• To provide a reliable method for assessing survival prediction models.

Main Methods:

• Analytically derived optimism expression for uncensored data.
• Proposed an algorithm for optimism estimation in Cox regression with covariates and right censoring.
• Utilized a reformulation of optimism for both scenarios.

Main Results:

• Successfully derived an optimism expression for uncensored data.
• Developed and verified an algorithm for optimism estimation in Cox models via simulations.
• Demonstrated the algorithm's applicability to real-world survival prediction.

Conclusions:

• The proposed covariance-based penalty estimator effectively addresses optimism in survival prediction models.
• The new algorithm enhances the assessment of Cox regression models with right-censored time-to-event data.
• This method improves the reliability of predictions in clinical settings, such as predicting dialysis access failure.