Skip to contents

Pre-Flight Validation for Counting-Process (Start/Stop) Survival Data

Source: R/surv_helpers.R
validate_start_stop.Rd

Inspects counting-process survival data before model tuning and reports subject, interval, event, and censored counts alongside structural checks for overlapping or gapped intervals, non-positive interval lengths (`stop <= start`), events occurring before a subject's last interval, rows appearing after an event, and tied stop/event times. Emits a prominent warning when no subjects are censored (e.g. 100

Usage

validate_start_stop(
  data,
  id = "id",
  start = "tstart",
  stop = "tstop",
  status = "status",
  warn_no_censoring = TRUE,
  error_on_invalid = FALSE
)

Arguments

data

A data frame containing counting-process survival intervals.

id, start, stop, status

Column names (as strings or bare symbols) for the subject identifier, interval start time, interval stop time, and event status indicator (`0` = censored, `1` = event). Defaults to `"id"`, `"tstart"`, `"tstop"`, and `"status"`.

warn_no_censoring

Logical (default `TRUE`). Emit a warning when every subject experiences an event (`n_censored == 0`).

error_on_invalid

Logical (default `FALSE`). When `TRUE`, stops with an error if any structural issue (missing values, `stop <= start`, overlapping intervals, events before the last interval, or rows after an event) is detected.

Value

A 1-row tibble of class `cbe_start_stop_validation` with columns: `n_subjects`, `n_intervals`, `n_events`, `n_subjects_with_event`, `n_censored`, `censoring_rate`, `n_overlapping`, `n_gaps`, `n_events_before_last`, `n_rows_after_event`, `n_nonpositive_length`, `n_tied_stop_times`, `n_tied_event_times`, `n_missing`, `has_censoring`, and `valid`. Detailed vectors of affected subject IDs are stored in `attr(x, "details")`.

See also

[surv_subject_truth()], [glmnet_IBS()]

Examples

df <- data.frame(
  id = c(1, 1, 2, 3),
  tstart = c(0, 5, 0, 0),
  tstop = c(5, 10, 8, 12),
  status = c(0, 1, 0, 1)
)
validate_start_stop(df, id = "id", start = "tstart", stop = "tstop", status = "status")
#> TempleCBE Start/Stop Survival Data Check
#> ----------------------------------------
#> Subjects:               3 (2 with event, 1 censored [33.3%])
#> Intervals (rows):       4 (events: 2, missing: 0)
#> Overlapping intervals:  0
#> Gapped intervals:       0
#> Events before last row: 0
#> Rows after an event:    0
#> Non-positive intervals: 0
#> Tied times:             0 stop time(s), 0 event time(s)
#> Status:                 VALID