Concordance Index

The concordance index (C-index) is a statistical measure of how well a prediction model ranks patients by their risk or expected survival. A value of 0.5 means the model predicts no better than chance, while 1.0 means perfect prediction; most clinically useful models fall between 0.6 and 0.8. It is the survival-analysis equivalent of the area under the ROC curve.

Origin

The concordance index was formalized by biostatistician Frank Harrell and colleagues in 1982 as a way to evaluate prognostic models when patient outcomes are censored (some patients are still alive at the end of the study). It built on earlier rank-correlation work by Kendall and Goodman-Kruskal and has since become a standard benchmark for survival models in oncology.

Significance in Asbestos Context

For mesothelioma and asbestos-related lung cancer, doctors and experts often use prognostic models (like EORTC, CALGB, or nomograms incorporating histology, stage, blood markers, and performance status) to estimate survival. The C-index tells you how trustworthy those predictions are for an individual patient. A model with a C-index of 0.55 is barely better than guessing, while 0.75 is meaningfully informative. In legal proceedings, life-expectancy estimates drive damages for future medical care, lost earnings, and loss of consortium; defense and plaintiff experts may rely on different prognostic models, and the C-index is one way to compare which model is better validated. Patients should understand that even a strong C-index reflects average performance across a population, not certainty about their own course.

Example Sentence

The oncologist explained that the nomogram she used to estimate survival had a concordance index of about 0.72, meaning it was reasonably accurate but could not guarantee how long any single patient would live.

Related Terms

Prognostic Model, Kaplan-Meier Curve, Hazard Ratio, Life Expectancy, Nomogram, Survival Analysis

Notes

• If a doctor or expert cites a survival prediction, ask which model was used and what its published C-index is. Models validated only in small studies may not apply well to your situation.
• Request that prognostic estimates in your medical record include the source model (EORTC, CALGB, etc.) so your legal team can have an expert evaluate its accuracy for damages calculations.
• Be aware that defense experts in asbestos cases sometimes use older or poorly validated models that systematically underestimate survival; ask your attorney whether your expert has addressed model validation and C-index.
• A high C-index does not mean the prediction for you personally is correct. It means the model ranks patients well on average. Individual outcomes still vary widely.