concord {coxphw} R Documentation

## Compute Generalized Concordance Probabilities for Objects of Class coxphw or coxph

### Description

Compute generalized concordance probabilities with accompanying confidence intervalls for objects of class coxphw or coxph.

### Usage

 concord(fit, digits = 4) 

### Arguments

 fit an object of class coxphw. digits integer indicating the number of decimal places to be used. Default is 4.

### Details

The generalized concordance probability is defined as P(T_i < T_j | x_i = x_j + 1) with T_i and T_j as survival times of randomly chosen subjects with covariate values x_i and x_j, respectively. Assuming that x_i and x_j are 1 and 0, respectively, this definition includes a two-group comparison.

If proportional hazards can be assumed, the generalized concordance probability can also be derived from Cox proportional hazards regression (coxphw with template = "PH" or coxph) or weighted Cox regression as suggested by Xu and O'Quigley (2000) (coxphw with template = "ARE").

If in a fit to coxphw the betafix argument was used, then for the fixed parameters only the point estimates are given.

### Value

A matrix with estimates of the generalized concordance probability with accompanying confidence intervalls for each explanatory variable in the model.

Daniela Dunkler

### References

Dunkler D, Schemper M, Heinze G. (2010) Gene Selection in Microarray Survival Studies Under Possibly Non-Proportional Hazards. Bioinformatics 26:784-90.

Xu R and O'Quigley J (2000). Estimating Average Regression Effect Under Non-Proportional Hazards. Biostatistics 1, 423-439.

coxphw
data("gastric")