Skip to contents

cce_spec() defines the minimal schema needed to convert normalized input tables into an analysis-ready dataset. The returned object is intentionally simple and can be stored as YAML for reproducible runs.

Usage

cce_spec(
  covariates,
  subgroup_biomarker = NULL,
  endpoint = "os",
  id_col = "patient_id",
  index_date_col = "index_date",
  regimen_col = "regimen_name",
  treatment_start_col = "start_date",
  index_flag_col = "is_index_treatment",
  endpoint_col = "endpoint",
  time_col = "time",
  event_col = "event",
  follow_up_col = "last_follow_up_date",
  biomarker_name_col = "biomarker_name",
  biomarker_value_col = "biomarker_value",
  biomarker_baseline_flag_col = "is_baseline",
  arm_map = c(SOC = "SOC", A = "A"),
  missing_strategy = "complete_case",
  time_zero_tolerance_days = 0L
)

Arguments

covariates

Character vector of baseline covariate column names.

subgroup_biomarker

Optional biomarker name used to derive a subgroup column from the biomarkers table.

endpoint

Endpoint name to retain from the outcomes table.

id_col

Patient identifier column name.

index_date_col

Baseline index-date column name.

regimen_col

Treatment regimen column name.

treatment_start_col

Index treatment start-date column name.

index_flag_col

Logical column marking the index treatment row.

endpoint_col

Outcome endpoint-name column.

time_col

Follow-up time column.

event_col

Event indicator column.

follow_up_col

Last follow-up date column.

biomarker_name_col

Biomarker-name column.

biomarker_value_col

Biomarker-value column.

biomarker_baseline_flag_col

Logical baseline-biomarker flag column.

arm_map

Named character vector mapping raw regimen labels to output labels. The output labels must be exactly c("SOC", "A").

missing_strategy

Missing-data rule. Only "complete_case" is implemented in v0.1.0.

time_zero_tolerance_days

Allowed difference between index_date and the index treatment start date.

Value

A cce_spec object.