Methods
Concordance index (C-index)
Also called: Harrell's C, C-statistic
The C-index measures how well a model ranks patients by risk: the chance that, of two patients, the one predicted at higher risk has the event first. 0.5 is chance, 1.0 is perfect.
The concordance index (C-index) is the standard discrimination measure for survival models. Take every usable pair of patients: the model is concordant for a pair if the patient it predicts at higher risk is the one who has the event sooner. The C-index is the proportion of concordant pairs.
A value of 0.5 means the model ranks no better than chance, and 1.0 means perfect ranking. In clinical prediction, values around 0.7 are often considered acceptable and above 0.8 strong, though what is useful depends on the decision being made. For a yes or no outcome, the C-index equals the area under the ROC curve (AUC).
The C-index measures ranking only. A model can rank well but still predict risks that are systematically too high or too low, so it should be reported together with calibration.
Tools that use this
- Aneurysm Recanalization StudioIndividualized risk of major recanalization after endovascular coiling over 12 to 48 months, with packing density what-ifs. C-index 0.83 on 3,530 aneurysms.
- Open Spina Bifida Frailty and SurvivalFrom findings at birth, predict frailty and the chance of survival at ages 5, 25, 35 and 40 for children born with open spina bifida.
Related terms
Educational reference. Not medical advice and not for clinical decision-making.