Skip to contents

Add quality registry data to skeleton

Source: R/add_quality_registry.R
add_quality_registry.Rd

Creates boolean skeleton columns from filter expressions evaluated against quality registry data (e.g., Riksstroke, NKBC). Unlike add_diagnoses() which matches ICD code prefixes, this function evaluates arbitrary R expressions against the dataset columns.

Usage

add_quality_registry(skeleton, dataset, id_name, date_col, codes = list())

Arguments

skeleton

A data.table containing the main skeleton structure created by create_skeleton.

dataset

A data.table containing quality registry data with person ID and a date column.

id_name

Character string specifying the name of the ID variable in the dataset.

date_col

Character string specifying the name of the date column in the dataset (must be Date or IDate class).

codes

Named list of filter definitions. Names become variable names in skeleton. Values are either TRUE (event flag) or quote() expressions evaluated against dataset columns. Example: list("stroke_event" = TRUE, "stroke_tia" = quote(tia == 1))

Value

The skeleton data.table is modified by reference with boolean variables added. New columns are FALSE on non-event weeks, and FALSE when filter expressions evaluate to NA (missing data).

The function also returns the skeleton, invisibly. The return is the object the caller passed, while that object has a free column slot for every new column. Past that point data.table cannot grow the column list in place, so the new columns land on the returned table alone. A caller that passes an expression rather than a variable MUST use the return value.

Any other name that pointed at the same table before the call is stale afterwards. R cannot grow a list in place. A growth therefore leaves the caller's binding on a NEW object, and every other name on the old one. Take an alias after the call, never before it.

Details

Each entry in codes produces one boolean column on the skeleton:

  • TRUE: event indicator, TRUE on any week with a registry event

  • quote(expr): filter expression evaluated against dataset columns. TRUE when the expression evaluates to TRUE, FALSE otherwise (including NA).

Multiple events in the same person-week are aggregated with any().

See also

add_diagnoses for ICD code matching, add_rx for prescription data

Other data_integration: add_annual(), add_cancer_without_morphology(), add_cods(), add_diagnoses(), add_onetime(), add_operations(), add_rx()

Examples

# Create fake data
data("fake_person_ids", package = "swereg")
skeleton <- create_skeleton(fake_person_ids[1:5], 2020, "2020-53")

fake_registry <- data.table::data.table(
  lopnr = c(fake_person_ids[1], fake_person_ids[2]),
  event_date = as.Date(c("2020-03-15", "2020-06-20")),
  severity = c(5, 18),
  treated = c(1, 2)
)

add_quality_registry(skeleton, fake_registry, "lopnr",
  date_col = "event_date",
  codes = list(
    "reg_event" = TRUE,
    "reg_severe" = quote(severity >= 15),
    "reg_treated" = quote(treated == 1)
  )
)