Skip to content

Add a regular episode-by-period panel to an extraction plan

Source: R/person_period.R
ds.omop.plan.person_period.Rd

Declares the regular relative-time bins that intersect the unique OMOP observation period covering each cohort index date. Event covariates are stored sparsely; a missing (rowId, timeId, covariateId) row represents zero only when that (rowId, timeId) exists in personPeriods. Each roster row keeps the requested startDay/endDay bin and adds inclusive observationStartDay, observationEndDay, and daysObserved for the observed part of the bin. The output contains no absolute dates or source event identifiers. It is a descriptive panel, not an inferred risk set: use survival/counting-process output when cohort end, death, or another censoring rule must define time at risk.

Usage

ds.omop.plan.person_period(
  plan,
  table,
  concept_set = NULL,
  bin_width = 30L,
  window_start = -365L,
  window_end = 0L,
  analyses = c("binary"),
  grain = "episode",
  time_origin = "index",
  name = "person_period"
)

Arguments

plan

An omop_plan object.

table

OMOP event table used for covariates.

concept_set

Optional concept IDs or an OHDSI-style concept-set spec. NULL requests all observed concepts subject to the server cap.

bin_width

Positive integer bin width in days.

window_start, window_end

Inclusive integer days from index.

analyses

Unique subset of "binary" and "count".

grain

Must be "episode".

time_origin

Must be "index".

name

Output name.

Value

The modified omop_plan.

See also

ds.omop.plan.temporal_covariates