pssm.power: A function to calculate the power for 'pssm'

In pssm: Piecewise Exponential Model for Time to Progression and Time from Progression to Death

Description

This function calculates the power, the sample size required for a given power, or the value of the parameter(delta) for a given sample size and power for testing for treatment effect on time to progression and for test for non-inferiority in terms of the survival at a fixed time point

Usage

pssm.power(parameter=c("N","Delta","Power","Power-Simulation"),
endpoint=c("Progression","Survival"),progression.delta=0.2 ,nsamp=100,alpha=0.025,
pow=0.80,tsurv=3,h0=0,prior.surv=0,alloc=c(1,1),phaz=log(.2),shaz=log(.2),accrual=3,
followup=2,inc=0.05,delta=0,intervals=1,nsims=500,n=5000,seed=NULL)

Arguments

parameter	A character vector of parameters to be calculated. For "N", "Delta" and "Power" the calculation proceeds by first calculating the standard error of the parameter measuring the treatment effect in a trial with n (default 3000) patients it then multiplies by the square root of n to get the proper factor to calculate the parameter called for. For "Power-Simulation" it runs a simulation to calculate the power for a specified sample size nsamp and progression.delta. This was included as a check on the asymptotics which is the basis of using one simulation to get a standard error and then scaling up or down to calculate the sample size or detectable difference.
endpoint	A character vector with either "Progression" or "Survival". In the first case it assumes that endpoint for which the sample size is being calculated is the parameter measuring the effect of treatment on the time to progression, which is log of the hazard ratio of placebo to active treatment. In the second it is the difference in the survival curves at the time tsurv
progression.delta	The log of the hazard ratio of placebo to active treatment. This used to calculate the sample size "N" and "Power". When Survival is the endpoint it is used to calculate the survival difference due to the effect of time to progression on overall survival.
nsamp	This is the sample size used to calculate "Power" or "Delta".
alpha	The one sided significance level to be used.
pow	The power required. This is used in the calculation of "N" and "Delta".
tsurv	The time used for the survival difference.
h0	The non-inferiority boundary for the survival difference given as a positive number.
prior.surv	The prior precision of the parameter that measures the effect of treatment on survival after progression. Note that this is 1/σ^2 where σ is the prior standard deviation of this parameter.
alloc	Allocation between control and treatment group c(1,1) is one to one allocation.
phaz	Log hazard of progression in the placebo group
shaz	Log hazard of survival after progression in the placebo and treatment group
accrual	Accrual period
followup	Follow up period.
inc	Time period between visits for acessing progression.
delta	Variation around visit times
intervals	Number of intervals to fit hazards
nsims	Number of simulations
n	Sample size used to estimate the standard Error.
seed	Seed for the random number generator if you don't want the data that is analyzed to change.

Value

A data frame is returned with variables, "parameter", "endpoint" and "value".

Author(s)

David A. Schoenfeld

See Also

pssm-class, pssm-package, pssm.generate.data, pssm.object, pssm, pssm.simulate, pssm.survivalcurv, plot-methods

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.
pssm.power(parameter="Power",progression.delta=.4,nsamp=450,h0=0.04,nsims=100)

pssm documentation built on May 2, 2019, 11:12 a.m.