sankey_list_maker converts timed sankey data at the patient level to summarized counts in
the nested hierarchical list format needed for the nswidgets:sankey function.
sankey_list_maker(
id_data,
cohort,
stage_labels = id_data$stages,
id_var = "patient_id",
index_var = "index_date",
weight = F,
censor_vars = NULL,
absorbing_vars = NULL,
none = "None",
gofl_formula = NULL,
collapse_states = NULL,
force = F
)List. One element for patient level order data, one element with states found in that data, one element with
stages. Returned from sankey_id_data. Pro tip: if you want to reduce number of stages, edit the stages vector.
Cohort data.
How do you want the stages (time points) labelled in the widget? Default is stages in id_data.
String. Variable name of the id/grouping variable.
Date. Indicates cohort_file date or day variable name per subject when event ordering should begin.
Logical. Should Sankey transitions be weighted? Default is FALSE.
Character vector. Variable names in cohort that indicate censoring date/day.
If multiple, sankey_list_maker will use the minimum. If weight = FALSE the Sankey will use the names of the vector to
label the states shown. If unnamed will group together under "Censored".
Named character vector. Variables in cohort that indicate date/day of states that members cannot transition out of. Names of the vector will be names of the states shown in the Sankey.
Character. How to label the "empty" state. Default is "None".
Formula. How should the Sankey be filtered? Uses gofl to create groupings. Default is NULL. Overall Sankey is
always created.
How to redefine states if combinations are anticipated. As a named list it allows users to identify how states
should be grouped with unique identifiers. As a named vector it assigns a hierarchy in the order that the states appear. Default NULL.
Force to calculate transitions even if they exceed guidelines.
Nested list with all the treatment pathways + sizes at each level.