Add a temporal (time-binned) covariates output to the plan

Produces FeatureExtraction-style sparse covariates binned into time windows relative to the cohort index date. Returns four symbols on the server: <name>.temporalCovariates, <name>.covariateRef, <name>.timeRef, and <name>.personRef. The last maps cohort episodes to pseudonymous persons. Requires a cohort to be set.

Usage

ds.omop.plan.temporal_covariates(
  plan,
  table,
  concept_set = NULL,
  bin_width = 30L,
  window_start = -365L,
  window_end = 0L,
  analyses = c("binary"),
  name = "temporal"
)

Arguments

plan

An omop_plan object.

table

Character; source OMOP table to extract covariates from.

concept_set

Optional concept IDs or an OHDSI-style concept-set spec with concepts, include_descendants, include_mapped, and exclude. When NULL, all concepts present in the bounded event stream are retained, subject to the server concept cap.

bin_width

Integer; width of each time bin in days.

window_start

Integer; start of the observation window in days relative to the cohort index date (negative = before index).

window_end

Integer; end of the observation window in days relative to the cohort index date (0 = index date).

analyses

Character vector; types of analyses to compute. Supported values include "binary" and "count".

name

Character; output name used as a key in the plan's outputs list.

Value

The modified omop_plan with the temporal covariates output appended.

See also

ds.omop.plan.intervals, ds.omop.plan.features

Examples

if (FALSE) { # \dontrun{
plan <- ds.omop.plan()
plan <- ds.omop.plan.cohort(plan, cohort_definition_id = 1)
plan <- ds.omop.plan.temporal_covariates(plan,
  table = "condition_occurrence",
  concept_set = c(201826, 443238),
  bin_width = 30L,
  window_start = -365L,
  window_end = 0L,
  analyses = c("binary", "count")
)
} # }