Skip to contents

Bucher method for combining treatment effects

Source: R/bucher.R
bucher.Rd

Given two treatment effects of A vs. C and B vs. C derive the treatment effects of A vs. B using the Bucher method. Two-sided confidence interval and Z-test p-value are also calculated. Treatment effects and standard errors should be in log scale for hazard ratio, odds ratio, and risk ratio. Treatment effects and standard errors should be in natural scale for risk difference and mean difference.

Usage

bucher(trt, com, conf_lv = 0.95)

# S3 method for class 'maicplus_bucher'
print(x, ci_digits = 2, pval_digits = 3, exponentiate = FALSE, ...)

Arguments

trt

a list of two scalars for the study with the experimental arm. 'est' is the point estimate and 'se' is the standard error of the treatment effect. For time-to-event data, 'est' and 'se' should be point estimate and standard error of the log hazard ratio. For binary data, 'est' and 'se' should be point estimate and standard error of the log odds ratio, log risk ratio, or risk difference. For continuous data, 'est' and 'se' should be point estimate and standard error of the mean difference.

com

same as trt, but for the study with the control arm

conf_lv

a numerical scalar, prescribe confidence level to derive two-sided confidence interval for the treatment effect

x

maicplus_bucher object

ci_digits

an integer, number of decimal places for point estimate and derived confidence limits

pval_digits

an integer, number of decimal places to display Z-test p-value

exponentiate

whether the treatment effect and confidence interval should be exponentiated. This applies to relative treatment effects. Default is set to false.

...

not used

Value

a list with 5 elements,

est

a scalar, point estimate of the treatment effect

se

a scalar, standard error of the treatment effect

ci_l

a scalar, lower confidence limit of a two-sided CI with prescribed nominal level by conf_lv

ci_u

a scalar, upper confidence limit of a two-sided CI with prescribed nominal level by conf_lv

pval

p-value of Z-test, with null hypothesis that est is zero

Methods (by generic)

  • print(maicplus_bucher): Print method for maicplus_bucher objects

Examples

trt <- list(est = log(1.1), se = 0.2)
com <- list(est = log(1.3), se = 0.18)
result <- bucher(trt, com, conf_lv = 0.9)
print(result, ci_digits = 3, pval_digits = 3)
#>                   result                   pvalue 
#> "-0.167 [-0.610; 0.276]"                  "0.535"