Proportion difference estimation

The analysis function a_proportion_diff_j() can be used to create a layout element to estimate the difference in proportion of responders within a studied population. The primary analysis variable, vars, is a logical variable indicating whether a response has occurred for each record. See the method parameter for options of methods to use when constructing the confidence interval of the proportion difference. A stratification variable can be supplied via the strata element of the variables argument.

Usage

a_proportion_diff_j(
  df,
  .var,
  ref_path,
  .spl_context,
  ...,
  conf_level = 0.95,
  label = paste0("% Difference (", f_conf_level(conf_level), ")"),
  .stats = NULL,
  .formats = NULL,
  .labels = NULL,
  .indent_mods = NULL
)

Arguments

df

(data.frame)
input data frame.

.var

(string)
name of the response variable.

ref_path

(character)
path to the reference group.

.spl_context

(environment)
split context environment.

...

Additional arguments passed to the statistics function.

conf_level

(numeric)
confidence level for the confidence interval.

label

(string)
label for the diff_est_ci statistic in a_proportion_diff_j(). Defaults to "% Difference (conf_level CI)". Can be overridden by the caller.

.stats

(character)
statistics to calculate.

.formats

(list)
formats for the statistics.

.labels

(list)
labels for the statistics.

.indent_mods

(list)
indentation modifications for the statistics.

Value

• a_proportion_diff_j() returns the corresponding list with formatted rtables::CellValue().

Functions

• a_proportion_diff_j(): Formatted analysis function which is used as afun in estimate_proportion_diff().

Note

The a_proportion_diff_j() function has the _j suffix to distinguish it from tern::a_proportion_diff(). It is a thin wrapper around tern::s_proportion_diff() with the following additions:

• The diff_est_ci statistic (returned natively by tern since tern#1523) is relabelled with a shorter junco-style label ("% Difference (conf_level CI)").

• ref_path needs to be provided as extra argument to specify the control group column.

Examples

nex <- 100
dta <- data.frame(
  "rsp" = sample(c(TRUE, FALSE), nex, TRUE),
  "grp" = sample(c("A", "B"), nex, TRUE),
  "f1" = sample(c("a1", "a2"), nex, TRUE),
  "f2" = sample(c("x", "y", "z"), nex, TRUE),
  stringsAsFactors = TRUE
)

l <- basic_table() |>
  split_cols_by(var = "grp") |>
  analyze(
    vars = "rsp",
    afun = a_proportion_diff_j,
    show_labels = "hidden",
    na_str = tern::default_na_str(),
    extra_args = list(
      conf_level = 0.9,
      method = "ha",
      ref_path = c("grp", "B")
    )
  )

build_table(l, df = dta)
#>                                           A           B
#> ———————————————————————————————————————————————————————
#> Difference in Response rate (%)         12.0           
#>   90% CI (Anderson-Hauck)           (-5.4, 29.4)       
#> % Difference (90% CI)             12.0 (-5.4, 29.4)