[Feature Request]: Option to specify .labels and .formats from tern

[crystalluckett-sanofi]

Feature description

As an example, it is possible to customize the labels and formats in the tern table functions:

library(tern)
dta_test <- data.frame(
  USUBJID = rep(1:6, each = 3),
  PARAMCD = rep("lab", 6 * 3),
  AVISIT  = rep(paste0("V", 1:3), 6),
  ARM     = rep(LETTERS[1:3], rep(6, 3)),
  AVAL    = c(9:1, rep(NA, 9))
)
l <- basic_table() %>%
  split_cols_by(var = "ARM") %>%
  split_rows_by(var = "AVISIT") %>%
  analyze_vars(
    vars = "AVAL",
    .formats = c("range" = "(xx.x, xx.x)"),
    .labels = c("range" = "(Min, Max)")
  )
build_table(l, df = dta_test)

However, there's no such option available in the TMC equivalent, tm_t_summary_by.
It would be nice if these could be specified across all table modules, and, even better, if a set of defaults could be specified that is applied to all tables, similar to tern:::summary_formats() and tern:::summary_labels().

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.

[Melkiades]

Thanks for the nice question! You can do what you want in two ways:

1. Add your personal formats and labels directly to analyze_vars in the template_summary_by, reinstall {teal.modules.clinical} and they will appear in your app. You can also check the stats bindings from the following

shiny::checkboxGroupInput(
            ns("numeric_stats"),
            label = "Choose the statistics to display for numeric variables",
            choices = c(
              "n" = "n",
              "Mean (SD)" = "mean_sd",
              "Mean 95% CI" = "mean_ci",
              "Geometric Mean" = "geom_mean",
              "Median" = "median",
              "Median 95% CI" = "median_ci",
              "25% and 75%-ile" = "quantiles",
              "Min - Max" = "range"
            ),
            selected = a$numeric_stats
          )

2. Keeping in mind the above binding labels (mean_sd etc) you can change the defaults on your local by overloading tern::tern_default_formats and/or tern::tern_default_labels (I would change only what you need so not to mess it up too much)

As a final note, I do not think that, for now, we have plans to make accessible from the UI to customize labels. Right @donyunardi?

[donyunardi]

Not yet, but this something that we're thinking as well.
We want the teal_module to be flexible for user to extend and customize the output.

This issue might be related with insightsengineering/teal#1384

[crystalluckett-sanofi]

@Melkiades sorry for the very late reply on this.

I guess the solution you've provided would technically work, but I was looking for something a little more functional.
For example, I may have one application where I want to round the mean to 2 decimals and another application where I want to round to 3 decimals instead. I would have to maintain multiple templates, one for each app?

Since there are options in tern, why not expose arguments in the module function as well? Perhaps with a new function argument like

tern_args = list(
   .formats = c("range" = "(xx.x, xx.x)"),
    .labels = c("range" = "(Min, Max)")
)

and then passing these args into analyze_vars?

[Melkiades]

@Melkiades sorry for the very late reply on this.

I guess the solution you've provided would technically work, but I was looking for something a little more functional. For example, I may have one application where I want to round the mean to 2 decimals and another application where I want to round to 3 decimals instead. I would have to maintain multiple templates, one for each app?

Since there are options in tern, why not expose arguments in the module function as well? Perhaps with a new function argument like

tern_args = list(
   .formats = c("range" = "(xx.x, xx.x)"),
    .labels = c("range" = "(Min, Max)")
)

and then passing these args into analyze_vars?

You can probably define different stats names (still referring to the same stats) and then change the formats. Still you need this info to be accessible from the GUI

[crystalluckett-sanofi]

I guess I'm not so much concerned with having these options available in the UI, but just in the actual function itself when configuring the module?