Extract Temporal Baseline and Prior Utilization Features — extract_temporal

Extracts baseline historical utilization (lookback encounter counts by category, recency, prior 30-day readmissions), stay characteristics (index LOS, discharge destination), and demographics directly in DuckDB SQL.

Usage

extract_temporal_features(
  con,
  cohort_table = "cohort",
  cohort_id = NULL,
  lookback_days = 365,
  encounter_categories = NULL,
  include_demographics = TRUE,
  include_stay_characteristics = TRUE,
  format = "df"
)

Arguments

con: Active DuckDB DBI connection.
cohort_table: Character string of cohort table or view (default "cohort").
cohort_id: Optional integer cohort definition ID filter.
lookback_days: Integer lookback window in days prior to admission (default 365).
encounter_categories: Named list of integer vectors mapping category names to visit concept IDs.
include_demographics: Logical. Include age, age group, sex, race, ethnicity (default TRUE).
include_stay_characteristics: Logical. Include index LOS and discharge category (default TRUE).
format: Output format: "df" (data.frame) or "arrow".

Value

A data.frame or Arrow Table.