surv2.logrank {surv2sample}R Documentation

Two-Sample Weighted Logrank Tests and Their Combinations

Description

Compares the distribution of survival times in two samples of censored data using the weighted logrank test with the G(rho,gamma) class of weights, or combinations (maximum or weighted sum) of several weighted logrank statistics.

Usage

surv2.logrank(x, group, rho.gamma = c(0, 0), comb, sum.weights,
              approx = "perm", nsim = 2000, choltol = 1e-07)

Arguments

x a "Surv" object, as returned by the Surv function.
group a vector indicating to which group each observation belongs. May contain values 1 and 2 only.
rho.gamma parameters (rho,gamma) of the weight. This must be a vector of length 2 or a list of vectors of length 2.
comb the method of combining test statistics. Possible values are "max" and "sum". comb is ignored, if there is only one set of parameters in rho.gamma, and defaults to "max" otherwise.
sum.weights weights for the weighted sum in the "sum" combination method. Defaults to equal weights for all statistics.
approx the method of approximating the distribution of the test statistic. Possible values are "perm" for permutations, "boot" for the bootstrap, "asympt" for asymptotics.
nsim the number of simulations. This means the number of permutations or bootstrap samples when approx is "perm" or "boot". When approx is "asympt", nsim is the number of simulations to approximate the asymptotic distribution (only needed for combination methods).
choltol a tolerance parameter for the Cholesky decomposition.

Details

The logrank test uses G(rho,gamma) weights defined as S(t)^rho (1-S(t))^gamma, where S(t) is the Kaplan–Meier estimate computed from the pooled sample.

Combination tests are based on a cluster of several weighted logrank statistics with different parameters (rho,gamma). The maximum statistic uses the maximum of absolute values of standardised statistics. The sum statistic uses the weighted sum of absolute values of standardised statistics. See Chapter 7 of Fleming and Harrington (1991) for details.

Value

A list of class "surv2.logrank" with components:

stat the test statistic.
pval the p-value.
rho.gamma as on input (for a test using a single statistic).
stats a matrix containing parameters of weights, corresponding statistics and p-values (for a combination test).

Further components include approx, nsim, comb, sum.weights from input.

Author(s)

David Kraus (http://www.davidkraus.net/)

References

Andersen, P. K., Borgan, O., Gill, R. D. and Keiding, N. (1993) Statistical Models Based on Counting Processes. Springer, New York.

Fleming, T. R. and Harrington, D. P. (1991) Counting Processes and Survival Analysis. Wiley, New York.

See Also

surv2.neyman, surv2.ks, survdiff, survfit

Examples

## gastric cancer data
data(gastric)

## Prentice--Wilcoxon (G^1) test
surv2.logrank(Surv(gastric$time, gastric$event),
    gastric$treatment, rho.gamma = c(1,0))

## combination of G(0,0), G(1,0), G(1,1) statistics
## maximum test
print(a <- surv2.logrank(Surv(gastric$time, gastric$event),
    gastric$treatment, rho.gamma = list(c(0,0), c(1,0), c(1,1)),
    comb = "max"))
## print individual statistics
a$stats

[Package surv2sample version 0.1-2 Index]