survcomp (version 1.22.0)

D.index: Function to compute the D index

Description

Function to compute the D index for a risk prediction, i.e. an estimate of the log hazard ratio comparing two equal-sized prognostic groups. This is a natural measure of separation between two independent survival distributions under the proportional hazards assumption.

Usage

D.index(x, surv.time, surv.event, weights, strat, alpha = 0.05, method.test = c("logrank", "likelihood.ratio", "wald"), na.rm = FALSE, ...)

Arguments

x

a vector of risk predictions.

surv.time

a vector of event times.

surv.event

a vector of event occurrence indicators.

weights

weight of each sample.

strat

stratification indicator.

alpha

apha level to compute confidence interval.

method.test

Statistical test to use in order to compute the p-values related to a D. index, see summary.coxph for more details.

na.rm

TRUE if missing values should be removed.

...

additional parameters to be passed to the coxph function.

Value

d.index

D index estimate.

se

standard error of the estimate.

lower

lower bound for the confidence interval.

upper

upper bound for the confidence interval.

p.value

p-value for the statistical test if the estimate if different from 0.5.

n

number of samples used for the estimation.

coxm

coxph.object fitted on the survival data and z (see below).

data

list of data used to compute the index (x, z, surv.time and surv.event). The item z contains the scaled rankits which are the expected standard Normal order statistics scaled by kappa.

Details

The D index is computed using the Cox model fitted on the scaled rankits of the risk scores instead of the risk scores themselves. The scaled rankits are the expected standard Normal order statistics scaled by kappa = sqrt(8/pi). See (Royston and Sauerbrei, 2004) for details.

Note that the value D reported in (Royston and Sauerbrei, 2004) is given

References

Royston, P. and Sauerbrei, W. (2004) "A new measure of prognostic separation in survival data", Statistics in Medicine, 23, pages 723--748.

See Also

coxph, coxph.object, normOrder

Examples

set.seed(12345)
age <- rnorm(100, 50, 10)
stime <- rexp(100)
cens   <- runif(100,.5,2)
sevent  <- as.numeric(stime <= cens)
stime <- pmin(stime, cens)
strat <- sample(1:3, 100, replace=TRUE)
weight <- runif(100, min=0, max=1)
D.index(x=age, surv.time=stime, surv.event=sevent, weights=weight, strat=strat)