spass (version 1.2)

n.gee.1subgroup: Sample Size estimation for longitudinal GEE models

Description

n.gee.1subgroup calculates the required sample size for proving a desired alternative when testing a regression coefficients in a full and/or a subpopulation. See 'Details' for more information.

Usage

n.gee.1subgroup(alpha, tail = "both", beta = NULL, delta, sigma,
  tau = 0.5, k = 1, npow = NULL, nmax = Inf)

Arguments

alpha

level (type I error) to which the hypothesis is tested.

tail

which type of test is used, e.g. which quartile und H0 is calculated.

beta

type II error (power=1-beta) to which an alternative should be proven.

delta

vector of estimated treatment effect in overall and sub population, c(overall population, only subpopulation).

sigma

vector of estimated standard deviations, c(full population, subpopulation). See 'Details'.

tau

subgroup prevalence.

k

sample size allocation factor between control and treatment: see 'Details'.

npow

calculates power of a test if npow is a sample size.

nmax

maximum total sample size.

Value

n.gee.1subgroup returns the required sample size within the control group and treatment group.

Details

This function performs a sample size estimation in a design with a nested subgroup within an overall population. To calculate the required sample only the value of tested regressor needs to inserted as delta. sigma is the variance of that regressor. The power for the global null hypothesis is given by 1-beta and alpha specifies the false positve level for rejecting \(H_0: \Delta_F=\Delta_S=0\) to level alpha.

Here argument k denotes the sample size allocation factor between treatment groups, i.e. \(k = n_T/n_C\).

See Also

bssr.1subgroup for blinded sample size re-estimation within a running trial and sandwich for estimating asymptotic covarianc mtrices in GEE models.

Examples

Run this code
# NOT RUN {
#Calculate required sample size to correctly reject Null with
#80% probability when testing global Nullhypothesis H_0: Delta_F=Delta_S = 0, while
#assuming the coefficient in and outside of the subgroup is Delta=c(0.1,0,1) with a
#subgroup-prevalence of tau=0.4.
#The variances of regressors in delta when variances are unequal sigma=c(0.8,0.4).

estimate<-n.gee.1subgroup(alpha=0.05,beta=0.2,delta=c(0.1,0.1),sigma=c(0.8,0.4),tau=0.4, k=1)
summary(estimate)

#Alternatively we can estimate the power our study would have
#if we know the effect in and outside our subgroup as
#well as the variance of the regressors. Here we
#estimate that only 300 Patiens total can be recruited and we are interested
#in the power that would give us.

n.gee.1subgroup(alpha=0.05,delta=c(0.1,0.1),sigma=c(0.8,0.4),tau=0.4, k=1, npow=300)

# }

Run the code above in your browser using DataCamp Workspace