Skip to contents

Calculate standard error from the reported confidence interval.

Source: R/bucher.R
find_SE_from_CI.Rd

Comparator studies often only report confidence interval of the treatment effects. This function calculates standard error of the treatment effect given the reported confidence interval. For relative treatment effect (i.e. hazard ratio, odds ratio, and risk ratio), the function would log the confidence interval. For risk difference and mean difference, we do not log the confidence interval. The option to log the confidence interval is controlled by 'logged' parameter.

Usage

find_SE_from_CI(
  CI_lower = NULL,
  CI_upper = NULL,
  CI_perc = 0.95,
  logged = TRUE
)

Arguments

CI_lower

Reported lower percentile value of the treatment effect

CI_upper

Reported upper percentile value of the treatment effect

CI_perc

Percentage of confidence interval reported

logged

Whether the confidence interval should be logged. For relative treatment effect, log should be applied because estimated log treatment effect is approximately normally distributed.

Value

Standard error of log relative treatment effect if 'logged' is true and standard error of the treatment effect if 'logged' is false

Examples

find_SE_from_CI(CI_lower = 0.55, CI_upper = 0.90, CI_perc = 0.95)
#> [1] 0.1256341