Extract Longitudinal Measurement Trajectory Features — extract_temporal

Leverages DuckDB's native regr_slope, regr_r2, and stddev_samp aggregations to extract trajectory features across clinical measurement series in [index_date - window_days, index_date].

Usage

extract_temporal_trajectories(
  con,
  cohort_table,
  measurement_concept_ids,
  window_days = 365L,
  index_date_col = "cohort_start_date",
  id_col = "person_id",
  offset_anchor = c("visit_start_date", "visit_end_date", "index_date"),
  anchor = NULL
)

Arguments

con: Active DBI/DuckDB connection.
cohort_table: Character table name or data frame containing cohort index dates.
measurement_concept_ids: Integer or numeric vector of measurement concept IDs.
window_days: Look-back window in days prior to index date (default 365L).
index_date_col: Column name for index date (default "cohort_start_date").
id_col: Column identifying patient (default "person_id").
offset_anchor: Milestone date to anchor lookback windows ('visit_start_date' | 'visit_end_date' | 'index_date').
anchor: Alias for offset_anchor.

Value

A data.frame with wide trajectory features per patient.