Tbl_Strata count by distinct individual vs rows - gtsummary

Question

How can I use tbl_strata and get the output to show counts by distinct individual rather than rows?
Also, how can I change the order that displays for the variable I am putting in the by= section in tbl_summary?
I have a long table AND a wide table with one row per patient. Not sure how to apply the wide table to this code. I can apply the long table but getting row counts instead of distinct patient counts.
I have included an example of the Long table I have and the wide table and what I would like the output to look like in the picture.
Example code:
#Wide Table Example
df_Wide <- data.frame(patientICN =c(1, 2, 3, 4, 5)
,testtype =c("liquid", "tissue", "tissue", "liquid", "liquid")
,gene1 =c("unk", "pos", "neg", "neg", "unk")
,gene2 =c("pos", "neg", "pos", "unk", "neg")
,gene3 =c("neg", "unk", "unk", "pos", "pos"))
#Long Table Example
df_Long <- data.frame(patientICN =c(1, 1, 2, 2, 3)
,testtype =c("liquid", "tissue", "tissue", "liquid", "liquid")
,gene =c("Gene1", "Gene2", "Gene3", "Gene1", "Gene2")
,result=c("Positve", "Negative", "Unknown","Positive","Unknown"))
#Table Categorized by testtype and result for long table
df_Long %>%
select (result, gene, testtype)%>%
mutate(testcategory=paste("TestType",testtype))%>%
tbl_strata(
strata=testtype,
.tbl_fun =
~.x %>%
tbl_summary(by=result,missing="no")%>%
add_n(),
.header= "**{strata}**, N={n}"
)
##above is giving multiple Rows per patient counts

Answer 1

Is this what you're after? You can install the bstfun pkg from my R-universe: https://ddsjoberg.r-universe.dev/ui#packages
library(gtsummary)
library(dplyr)
#Long Table Example
df_Long <- data.frame(patientICN =c(1, 1, 2, 2, 3)
,testtype =c("liquid", "tissue", "tissue", "liquid", "liquid")
,gene =c("Gene1", "Gene2", "Gene3", "Gene1", "Gene2")
,result=c("Positive", "Negative", "Unknown","Positive","Unknown"))
tbl <-
df_Long %>%
tidyr::pivot_wider(
id_cols = c(patientICN, testtype),
names_from = gene,
values_from = result,
values_fill = "Unknown"
) %>%
mutate(across(starts_with('Gene'), ~factor(.x, levels = c("Positive", "Negative", "Unknown")))) %>%
tbl_strata(
strata = testtype,
~ .x %>%
bstfun::tbl_likert(
include = starts_with("Gene")
)
)
Created on 2022-10-06 with reprex v2.0.2