surv1samp: Sample Size Calculation for a One-Sample Exponential

In raredd/desmon: Functions for design and monitoring of clinical trials

Sample Size Calculation for a One-Sample Exponential

Description

Computes the sample size for a given size and power for the test for a specified improvement in the exponential rate parameter.

Usage

surv1samp(
  control.rate = NULL,
  pct.imp = NULL,
  acc.per,
  add.fu,
  alpha = 0.1,
  power = 0.9,
  control.med = NULL,
  control.survp = NULL,
  tp = NULL,
  pct.reduc = NULL,
  nonpar = FALSE,
  alt.rate = NULL,
  alt.med = NULL,
  alt.survp = NULL
)

Arguments

control.rate	The hazard rate under the null hypothesis
pct.imp	The percent improvement in the median failure time under the alternative
acc.per	The length of the time period during which patients are accrued
add.fu	The length of time after the end of accrual during which patients continue to be followed prior to analysis
alpha	The one-sided type I error rate
power	The power of the test
control.med	The median under the null
control.survp	The survival probability at tp under the null
tp	Time point for survival probability values and/or the nonparametric test
pct.reduc	Percent reduction in the hazard rate under the alternative
nonpar	If TRUE, the test based on the nonparametric cumulative hazard estimator at tp is used
alt.rate	The hazard rate under the alternative
alt.med	The median under the alternative
alt.survp	The survival probability at tp under the alternative

Details

Accrual is assumed to be uniform over the accrual period, so censoring is uniform on the interval add.fu to acc.per + add.fu. The power is based on the Wald test for the log failure rate parameter unless nonpar=TRUE, in which case the test is based on the nonparametric estimate of the cumulative hazard at tp. Simulations suggest that the distribution approximation to the latter test is anti-conservative, so the test (and the corresponding power calculations should be used with extreme caution). Exponential failures are assumed for the calculations, even for the nonparametric test.

Any combination of parameters determining the null and alternative event rates can be specified.

Value

A vector giving the sample size (n), the number of failures needed (nd), the proportion of failures (fail.prob), the hazard ratio of the null to alternative hazard rates (haz.ratio), the input values of pct.imp and control.rate, the hazard rate under the alternative (alt.rate) and the input values of alpha and beta.

Examples

surv1samp(-1.333 * log(0.96) / 5, 33.3, 4, 3, alpha = 0.1, power = 0.95)
surv1samp(-log(0.95) / 5, 100 * (log(0.95) / log(0.96) - 1), 4, 3,
          alpha = 0.1, power = 0.95)
surv1samp(-log(0.955) / 5, 100 * (log(0.955) / log(0.965) - 1), 4, 3,
          alpha = 0.1, power = 0.95)

raredd/desmon documentation built on May 7, 2024, 3:46 p.m.