[Bug]: `tm_t_summary()` show NA values when they should be omitted

[llrs-roche]

What happened?

Output doesn't respect the option to show and hide NAs.
This was first reported internally by one user for tm_t_summary() and tm_t_summary_by().

This might be related to using df_explicit_na() inside the module, that forces all NAs to be present as a new factor.

data <- within(teal_data(), {
  ADSL <- tmc_ex_adsl
  ADSL$EOSDY[1] <- NA_integer_
  ind <- 1:100
  ADSL$SEX[ind] <- NA
  # Get code::
  set.seed(1000)
  ind <- sample(seq_len(nrow(ADSL)), 100)
  ADSL$RACE[ind] <- NA
})
join_keys(data) <- default_cdisc_join_keys[names(data)]

app <- init(
  data = data,
  modules = modules(
    tm_t_summary(
      label = "Demographic Table",
      dataname = "ADSL",
      arm_var = choices_selected(c("ARM", "ARMCD"), "ARM"),
      add_total = TRUE,
      summarize_vars = choices_selected(
        c("SEX", "RACE", "BMRKR2", "EOSDY", "DCSREAS", "AGE"),
        c("SEX", "RACE")
      ),
      useNA = "ifany"
    ),
    tm_t_summary_by(
      label = "Summary by Row Groups Table",
      dataname = "ADSL",
      arm_var = choices_selected(c("ARM", "ARMCD"), "ARM"),
      add_total = TRUE,
      summarize_vars = choices_selected(
        choices = variable_choices("ADSL", c('EOSDY')),
        selected = c('EOSDY')
      ),
      by_vars = choices_selected(
        c("SEX", "RACE", "BMRKR2", "EOSDY", "DCSREAS", "AGE"),
        c("SEX", "RACE")
      ),
      useNA = "ifany"
    )
  )
) |> runApp()

sessionInfo()

Relevant log output

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

[llrs-roche]

Opened an issue on tern about analyze_vars() not working when some cells have NAs and others don't: insightsengineering/tern#1440

Depending on how it is fixed we could set NA as df_explicit_NA when there are NAs and useNA = "ifany" is selected.
For excluding NA when useNA = "no" is enough to remove those df_explicit_NA calls.