Skip to contents

Fake Swedish Registry Datasets

Source: R/data.R
fake_data.Rd

These datasets contain synthetic Swedish healthcare registry data for development, testing, and vignettes. They mimic the structure and format of real Swedish healthcare registries but contain completely fabricated data.

Usage

fake_person_ids

fake_demographics

fake_annual_family

fake_diagnoses

fake_prescriptions

fake_cod

Format

Various data structures matching real Swedish registries:

An object of class integer of length 1000.

An object of class data.table (inherits from data.frame) with 1000 rows and 3 columns.

An object of class data.table (inherits from data.frame) with 1000 rows and 2 columns.

An object of class data.table (inherits from data.frame) with 5000 rows and 49 columns.

An object of class data.table (inherits from data.frame) with 8000 rows and 37 columns.

An object of class data.table (inherits from data.frame) with 50 rows and 5 columns.

Source

Generated synthetic data based on real Swedish healthcare registry structures

Details

These datasets are created by dev/generate_fake_data.R and contain:

Key features:

  • Personal identifiers are numeric (e.g., 623334, 753064)

  • Prescription data uses column name "p444_lopnr_personnr"

  • ICD-10 codes include gender dysphoria (F64*), mental health (F20*, F32*, F40*), and physical health codes

  • ATC codes include hormone therapy (G03*), mental health medications (N05*, N06*)

  • Date ranges span 1978-2021 depending on registry

  • Realistic missing data patterns

  • SOURCE column in fake_diagnoses tracks data origin

Usage requirements:

  • Always apply swereg::make_lowercase_names() after loading data

  • Use appropriate identifier column names (lopnr vs p444_lopnr_personnr)

  • Follow Swedish registry conventions for date formats

  • Filter by SOURCE column when needed (e.g., SOURCE == "cancer" for ICD-O-3)

fake_person_ids

A numeric vector of 1000 fake personal identifiers (lopnr). Used as reference IDs across all other datasets.

fake_demographics

Demographics data (SCB format) with 1000 records:

lopnr

Personal identifier matching fake_person_ids

fodelseman

Birth year-month (YYYYMM format)

DodDatum

Death date (YYYYMMDD format) or empty string

fake_annual_family

Annual family status data (SCB format) with 1000 records:

LopNr

Personal identifier (mixed case as in real data)

FamTyp

Family type code (2-digit character)

fake_diagnoses

Combined diagnosis data with ~5000 records from three sources:

SOURCE

Data source: "inpatient", "outpatient", or "cancer"

LopNr

Personal identifier

AR

Year of care

INDATUMA

Admission date (YYYYMMDD character)

INDATUM

Admission date (Date class)

UTDATUMA

Discharge date (YYYYMMDD character, inpatient only)

UTDATUM

Discharge date (Date class, inpatient only)

HDIA

Main diagnosis (ICD-10 code)

DIA1-DIA30

Additional diagnoses

EKOD1-EKOD7

External cause codes

OP

Operation codes

ICDO3

ICD-O-3 morphology codes (populated for cancer source)

SNOMED3

SNOMED-CT version 3 codes

SNOMEDO10

SNOMED-CT version 10 codes

The SOURCE column identifies the registry origin:

  • "inpatient": NPR inpatient data (~2000 records)

  • "outpatient": NPR outpatient data (~2000 records)

  • "cancer": Cancer registry data (~1000 records, always has ICDO3)

fake_prescriptions

Prescription drug dispensing data (LMED format) with ~8000 records:

p444_lopnr_personnr

Personal identifier with p444 prefix

Fall

Case indicator

Kontroll

Control indicator

VARUNR

Product number

ATC

ATC classification code

ALDER

Age at prescription

LK

Healthcare county code

EDATUM

End date

FDATUM

Start date

OTYP

Origin type

...

Additional 27 columns matching real LMED structure

fake_cod

Cause of death data with ~50 records (Swedish registry format):

lopnr

Personal identifier

dodsdat

Date of death

ulorsak

Underlying cause of death (ICD-10) - Swedish variable name

morsak1

First multiple/contributory cause of death

morsak2

Second multiple/contributory cause of death

Examples

if (FALSE) { # \dontrun{
# Load fake data
data("fake_person_ids")
data("fake_demographics")
data("fake_diagnoses")

# CRITICAL: Apply lowercase names
swereg::make_lowercase_names(fake_demographics)
swereg::make_lowercase_names(fake_diagnoses, date_columns = "indatum")

# Check source distribution
table(fake_diagnoses$source)

# Filter by source
inpatient_only <- fake_diagnoses[source == "inpatient"]
cancer_only <- fake_diagnoses[source == "cancer"]

# Create skeleton with fake data
skeleton <- create_skeleton(
  ids = fake_person_ids[1:100],
  date_min = "2015-01-01",
  date_max = "2020-12-31"
)
} # }