teal-clinical-modules

Create a teal_data object with your ADaM datasets inside within(), set join_keys with default_cdisc_join_keys, then pass modules like tm_t_summary, tm_t_events, and tm_g_km to init(). Each module needs a dataname and choices_selected selectors for variables.

Use tm_g_km with dataname set to your TTE dataset and provide the four required arguments: arm_var, paramcd, strata_var, and facet_var. Use value_choices for paramcd to select endpoints like OS, and pass selected = NULL for facet_var if no faceting is needed.

The arm_var column and other stratification variables must be factor type because the underlying tern package uses factor levels to define groups. Real ADaM data like pharmaverseadam stores these as character, so convert them with dplyr::mutate and as.factor inside within().

variable_choices selects column names from a dataset, such as choosing between ARM and ARMCD for the treatment variable. value_choices selects specific values within one column, such as picking a PARAMCD endpoint like OS or PFS.

Binary teal packages from Posit Package Manager can freeze a build-machine bslib path into the default theme, causing a white screen. Override it after loading teal with options(teal.bs_theme = bslib::bs_theme(version = 5)).

Converting character columns with as.factor drops variable label attributes, which teal's transform pipeline requires for analysis modules. Restore labels after conversion by copying the label attributes from the original dataset back onto each column.