Skip to contents

Create Long-Format Survival Data from Exposure-Response Events

Source: R/survival_analysis.R
create_dsurv.Rd

THIS FUNCTION WILL BE DEPRICATED IN FUTURE. Processes raw exposure-response data into a long-format dataset suitable for survival analysis, including response, death, and censoring times. Supports use in Shiny with progress tracking.

Usage

create_dsurv(
  data,
  censoring_date = as.Date("2024-12-21"),
  filter_early_responses = FALSE,
  type = c("exp_to_resp", "resp_to_death")
)

Arguments

data

A data frame that must include columns: `ID`, `exp.GROUP`, `exp.DATE`, `resp.DATE`, `DATE_DEATH`, `DATE_MIGRATION`. These are used to calculate event and censoring times.

censoring_date

A `Date` object specifying the administrative censoring date. Default is `"2023-12-21"`.

filter_early_responses

Logical; if `TRUE`, responses occurring before the exposure date (negative time) are filtered out. If `FALSE`, they are included and recoded to 0. Default is `FALSE`.

type

character. Options: "exp_to_resp" or "resp_to_death"

Value

A data frame in long format with columns:

ID

Subject ID

name

Type of event: `"diagnose"`, `"dead"`, or `"censoring"`

value

Time (in days) from exposure to event

Details

This function is intended to support dynamic survival model construction. It handles competing risks (response, death, censoring) and can run inside a Shiny app with progress bar support via `.safe_inc_progress()`. Internally, it filters and summarizes the earliest valid event per subject.

Examples

if (FALSE) { # \dontrun{
dsurv <- create_dsurv(
  data = exposure_response_df,
  censoring_date = as.Date("2023-12-31"),
  filter_early_responses = TRUE
)
} # }