On the C‐statistics for evaluating overall adequacy of risk prediction procedures with censored survival data
Key Points
- To establish a C-statistic for evaluating survival risk prediction models that remains free from the influence of study-specific censoring distributions.
- Formulated a new estimator that consistently assesses a conventional concordance measure independent of censoring patterns.
- Derived large-sample distribution approximations to facilitate formal statistical inference on the concordance metric.
- Assessed finite-sample estimation behavior across simulated numerical studies.
- The proposed C-statistic consistently estimates the target concordance measure without bias from censoring mechanisms.
- Numerical studies demonstrated strong finite-sample performance across evaluated risk prediction models.
Authors
Discussion
No takes yet. Share an insight, caveat, or question.
Implication
Methodological study demonstrates a censoring-free C-statistic in clinical survival cohorts, highlighting improved discrimination accuracy for risk prediction.
Cite This Study
Uno et al. (2011) studied this question.