survMisc (version 0.5.3)

rsq: r^2 measures for a a coxph or survfit model

Description

r^2 measures for a a coxph or survfit model

Usage

rsq(x, ...)
"rsq"(x, ..., sigD = 2)
"rsq"(x, ..., sigD = 2)

Arguments

x

A survfit or coxph object.

sigD

significant digits (for ease of display). If sigD=NULL, will return the original numbers.

...

Additional arguments (not implemented).

Value

A list with the following elements:

cod

The coefficient of determination, which is $$R^2=1-\exp(\frac{2}{n}L_0-L_1)$$ where $L[0]$ and $L[1]$ are the log partial likelihoods for the null and full models respectively and $n$ is the number of observations in the data set.

mer

The measure of explained randomness, which is: $$R^2_{mer}=1-\exp(\frac{2}{m}L_0-L_1)$$ where $m$ is the number of observed events.

mev

The measure of explained variation (similar to that for linear regression), which is: $$R^2=\frac{R^2_{mer}}{R^2_{mer} + \frac{\pi}{6}(1-R^2_{mer})}$$

References

Nagelkerke NJD, 1991. A Note on a General Definition of the Coefficient of Determination. Biometrika 78(3):691--92. JSTOR

O'Quigley J, Xu R, Stare J, 2005. Explained randomness in proportional hazards models. Stat Med 24(3):479--89. Wiley (paywall) Available at UCSD

Royston P, 2006. Explained variation for survival models. The Stata Journal 6(1):83--96. The Stata Journal

Examples

data("kidney", package="KMsurv")
c1 <- coxph(Surv(time=time, event=delta) ~ type, data=kidney)
cbind(rsq(c1), rsq(c1, sigD=NULL))