Feature request: Should `add_overall.tbl_summary()` accept a statistic argument?

I had a question from a colleague who wanted to display a cross table with just percentages in the main table, but with % and number of observations per row in total column.

It could be done by creating two tables and then using tbl_merge(). Here an example.

library(questionr)
library(gtsummary)
#> #BlackLivesMatter
library(labelled)
#> Warning: le package 'labelled' a été compilé avec la version R 4.1.2

data(hdv2003)
hdv2003 <- hdv2003 %>%
  set_variable_labels(
    nivetud = "Niveau d'étude",
    sexe = "Sexe"
  )

theme_gtsummary_language("fr", decimal.mark = ",")
#> Setting theme `language: fr`

tbl_by <- hdv2003 %>%
  tbl_summary(
    by = relig,
    include = c(nivetud, sexe),
    percent = "row",
    statistic = ~ "{p}%",
    digits = ~ 1
  ) %>%
  bold_labels() %>%
  modify_header(all_stat_cols(stat_0 = FALSE) ~ "**{level}**")

tbl_total <- hdv2003 %>%
  tbl_summary(
    include = c(nivetud, sexe),
    percent = "row",
    statistic = ~ "{p}% (n={n})",
    digits = ~ c(0, 0)
  ) %>%
  modify_header(stat_0 ~ "**Total**")

tbl_merge(
  list(tbl_by, tbl_total),
  tab_spanner = c("**Rapport à la religion**", NA)
) %>%
  modify_footnote(all_stat_cols() ~ NA) %>%
  as_kable()

Caractéristique	Pratiquant regulier	Pratiquant occasionnel	Appartenance sans pratique	Ni croyance ni appartenance	Rejet	NSP ou NVPR	Total
Niveau d’étude
N’a jamais fait d’etudes	48,7%	20,5%	15,4%	7,7%	5,1%	2,6%	100% (n=39)
A arrete ses etudes, avant la derniere annee d’etudes primaires	23,3%	25,6%	38,4%	9,3%	2,3%	1,2%	100% (n=86)
Derniere annee d’etudes primaires	17,6%	26,7%	38,7%	11,7%	2,3%	2,9%	100% (n=341)
1er cycle	14,2%	25,5%	32,4%	20,1%	4,4%	3,4%	100% (n=204)
2eme cycle	13,7%	14,8%	39,3%	23,5%	6,0%	2,7%	100% (n=183)
Enseignement technique ou professionnel court	6,9%	24,6%	37,6%	23,5%	5,6%	1,7%	100% (n=463)
Enseignement technique ou professionnel long	8,4%	18,3%	43,5%	25,2%	3,1%	1,5%	100% (n=131)
Enseignement superieur y compris technique superieur	11,8%	19,7%	42,2%	19,7%	5,4%	1,1%	100% (n=441)
Manquant	18	17	34	35	7	1	112
Sexe
Homme	9,6%	18,9%	40,6%	23,4%	5,8%	1,8%	100% (n=899)
Femme	16,3%	24,7%	35,9%	17,2%	3,7%	2,2%	100% (n=1 101)

^{Created on 2021-12-06 by the reprex package (v2.0.1)}

Somehow, it would be much easier if possible, in add_overall() to overwrite the statistic and the digits argument., i.e.

tbl %>%
  add_overall(
    statistic = ~ "{p}% (n={n})",
    digits = ~ c(0, 0)
  )

ddsjoberg / gtsummary

Feature request: Should `add_overall.tbl_summary()` accept a statistic argument? #1047