ci_general: Noninferiority confidence intervals

In NourHawila/EC: Exact-corrected confidence interval and other useful functions to analyze noninferiority binomial trials

Description

This function computes the confidence interval limits given an ordering criterion with or without the exact-correction

Usage

ci_general(
  x.T,
  x.C,
  N.T,
  N.C,
  delta0,
  method = "MN",
  EC,
  alpha = 0.05,
  tol = 1e-10
)

Arguments

x.T	positive integer representing the observed number of responders in the treatment group
x.C	positive integer representing the observed number of responders in the control group
N.T	positive integer representing the sample size in the treatment group
N.C	positive integer representing the sample size in the control group
delta0	numeric between 0 and 1 representing the noninferiority margin
method	character representing the method for ordering criterion("MN","FM","SS","Blackwelder")
EC	logical. TRUE for the exact-corrected confidence limits. FALSE for default method without exact-correction
alpha	numeric between 0 and 1 representing the significance level
tol	positive numeric representing the tolerance for convergence

Value

list of length 2 (ci.lower, ci.upper) representing the lower and upper confidence limits

References

\insertRef

Hawila:21EC

\insertRef

Miettinen:85EC

\insertRef

Farrington:90EC

Examples

#These two examples demonstrate the confidence intervals for the
#Rodary et al. study with and without the exact-correction.
ci_general(x.T=83,x.C=69,N.T=88,N.C=76,delta0=0.1,method="MN", EC=TRUE)
ci_general(x.T=83,x.C=69,N.T=88,N.C=76,delta0=0.1,method="MN", EC=FALSE)

NourHawila/EC documentation built on Feb. 24, 2022, 8:24 a.m.