You may also use value argument in tbl_summary like below. First convert the NA in outcome to an explicit level "unknown". Then modify your table to show the counts of "Yes" in your outcome variable.

df |> 
  mutate(outcome = if_else(is.na(outcome), "unknown", outcome)) |> 
  tbl_summary(
    by = group,
    percent = "column",
    value = list(outcome ~ "Yes")
  )