Continuous-value distribution over a cohort, in one call

Thin wrapper over ds.omop.analysis.run for the catalog's continuous-covariate analysis ("dsomop:fe.continuous"): per-covariate count and avg/sd/median/p10-p90 over a scoped cohort (measurement values, age, or time-in-cohort). It builds the params and delegates, inheriting scoping, pooling, optional plotting, and the ONE disclosure gate (which strips min/max and masks sub-threshold stats) unchanged.

Usage

ds.omop.distribution(
  cohort = NULL,
  metric = "measurement_value",
  domain = "measurement",
  top_n = 50,
  concept_id = NULL,
  tables = NULL,
  plot = FALSE,
  symbol = "omop",
  conns = NULL
)

Arguments

cohort

Cohort reference to scope to (handle, cohort_definition_id, or server-side table name). Required unless tables is given.

metric

Character; "measurement_value" (default), "age", or "time_in_cohort".

domain

Character; value domain for metric = "measurement_value" by name ("measurement" / "observation") or code ("3"/"4"). Default "measurement".

top_n

Integer; number of top covariates to return (default 50).

concept_id

Integer vector or NULL; narrow the gated result to these covariate concept id(s) (post-gate row subset). Default NULL.

tables

Optional character vector of omop.table symbol names to scope to; may be combined with cohort.

plot

Logical; build the entry's client-side plot (default FALSE).

symbol

Character; the session symbol (default "omop").

conns

DSI connection object(s) or NULL to use the session default.

Value

A dsomop_result (see ds.omop.analysis.run).

Details

As with ds.omop.prevalence, the cohort IS the population, so a cohort/tables scope is required; an un-scoped call errors clearly.

See also

ds.omop.analysis.run, ds.omop.prevalence

Examples

if (FALSE) { # \dontrun{
# Measurement-value distributions over a cohort, one call.
ds.omop.distribution(cohort = my_cohort)

# Age distribution of the cohort.
ds.omop.distribution(cohort = my_cohort, metric = "age")
} # }