Skip to contents

Emergency Department Registry Dataset (SVAR, 2017-2018)

svar_emergency_registry_skane.Rd

This dataset contains information about emergency department visits in Skåne region during 2017-2018, including patient demographics, visit timings, triage levels, vital parameters, and outcomes. The data was extracted from the Akutvårdsregistrering system.

Usage

svar_emergency_registry_skane()

Format

A data frame with 702,551 observations and multiple variables:

patient_id

Integer. Patient pseudonym identifier, serves as a foreign key to link with other patient-level data. Available for approximately 76.7% of records after linkage. Original field name: Mapped from visit_id via key_data

hospital

Character. Hospital or center name where the visit occurred. Original field name: CentridTRIMJUST

visit_id

Character. Unique identifier for each emergency department visit. May correlate with the "Nyckel" ID in other datasets. Original field name: BesöksIDAkutvårdsregistrering

patient_birth_date

Character. Patient's birth date. Original field name: Födelsedatum

age_at_arrival

Numeric. Patient's age in years at arrival. Original field name: Åldervidankomsttillakutmottagning

gender

Character. Patient's gender (standardized to "Male"/"Female"). Original field name: Identitetstyp

arrival_datetime

POSIXct. Date and time of arrival at the emergency department. Original field name: AnkomsttillakutmottagningAkutvårdsregistrering

arrival_date

Date. Date of arrival at the emergency department. Derived from arrival_datetime.

arrival_hour

Numeric. Hour of arrival (0-23). Original field name: Ankomsttimme

planned_return_visit

Logical. Whether this was a planned return visit. Original field name: PlaneratåterbesökAkutvårdsregistrering

arrival_method

Character. How the patient arrived (ambulance, walking.) Original field name: AnkomstsättAkutvårdsregistrering

referral

Character. Whether the patient was referred and by whom. Original field name: RemissAkutvårdsregistrering

chief_complaint

Character. Reason for the emergency department visit. NOTE: This field is known to have accuracy issues in many cases. Original field name: BesöksorsakAkutvårdsregistrering

referral_elsewhere

Logical. Whether the patient was referred elsewhere. NOTE: This field has substantial missing data (~50K+ records). that is, SVAR does not contain alot of reffered patient visits. the NA's in this file are unknown. Original field name: HänvisningAkutvårdsregistrering

triage_level

Character. Assigned triage level (Red, Orange, Yellow, Green, Blue). NOTE: Triage data is known to be inaccurate, particularly for walk-in patients (95%+ labeled as Green) and entirely at Helsingborg and Ystad hospitals. Original field name: TriagenivåAkutvårdsregistrering

nurse_time

POSIXct. Time when the patient was first seen by a nurse. Original field name: SjukskötersketidAkutvårdsregistrering

doctor_time

POSIXct. Time when the patient was first seen by a doctor. Original field name: LäkartidAkutvårdsregistrering

completion_time

POSIXct. Time when care was completed. NOTE: The definition of this metric is unclear. Original field name: FärdigtidAkutvårdsregistrering

discharge_time

POSIXct. Time when the patient was discharged. Original field name: UttidAkutvårdsregistrering

left_without_being_seen

Logical. Whether the patient left without being seen. Original field name: AvvikitAkutvårdsregistrering

discharged_to

Character. Where the patient was discharged to. Original field name: UtskriventillAkutvårdsregistrering

discharge_diagnosis

Character. ICD-10 diagnosis code at discharge. Original field name: UtdiagnosAkutvårdsregistrering

time_to_nurse

Numeric. Time in minutes from arrival to nurse assessment. Recalculated from arrival_datetime and nurse_time.

time_to_doctor

Numeric. Time in minutes from arrival to doctor assessment. Recalculated from arrival_datetime and doctor_time.

total_visit_time

Numeric. Total time in minutes from arrival to discharge. Recalculated from arrival_datetime and discharge_time.

death_days_after_visit

Numeric. Number of days from visit to death, if applicable. Set to 0 if discharged_to is "Avlidit" (Deceased). Original field name: Dödefter…dagar

death_within_7days

Logical. Whether the patient died within 7 days of visit. recalculated due to missing registries from death_days_after_visit Original field name: Dödinom7dagar

death_within_30days

Logical. Whether the patient died within 30 days of visit. recalculated due to missing registries from death_days_after_visit Original field name: Dödinom30dagar

systolic_bp

Numeric. Systolic blood pressure. NOTE: Substantial missing data; values may not be the first measurement. Original field name: SystolisktblodtryckVitalparameter

diastolic_bp

Numeric. Diastolic blood pressure. NOTE: Substantial missing data; values may not be the first measurement. Original field name: DiastolisktblodtryckVitalparameter

heart_rate

Numeric. Heart rate. NOTE: Substantial missing data; values may not be the first measurement. Original field name: Hjärtfrekvens Vitalparameter

respiratory_rate

Numeric. Respiratory rate. NOTE: Substantial missing data; values may not be the first measurement. Original field name: Andningsfrekvens Vitalparameter

oxygen_saturation

Numeric. Oxygen saturation (%). NOTE: Substantial missing data; values may not be the first measurement. Original field name: SaturationVitalparameter

temperature

Numeric. Body temperature. NOTE: Substantial missing data; values may not be the first measurement. Original field name: Kroppstemperatur Vitalparameter

consciousness_rls

Character. Consciousness level on RLS scale. NOTE: Substantial missing data. Original field name: MedvetandegradRLSVitalparameter

consciousness_gcs

Numeric. Glasgow Coma Scale score. NOTE: Substantial missing data. Original field name: MedvetandegradenligtGlascowComaScale Vitalparameter

Source

SVAR (Svenska Akutvårdsregistret), Skåne Region

Details

This dataset captures emergency department visits across Skåne region hospitals in 2017-2018. It includes patient demographics, visit timings, triage levels, vital signs, and outcomes. Several fields have quality issues noted in their descriptions.

The dataset uses multiple missing value indicators including NA, "#NULL!", and "#SKÄRNING!". In the processed data, these are standardized to NA values. Time calculations (time_to_nurse, time_to_doctor, total_visit_time) are recalculated from the datetime fields rather than using the potentially erroneous provided values.

The original SVAR data has been linked with patient_id from the key_data mapping file, providing identification for approximately 76.7% of SVAR records. This linkage allows integration with other datasets in the SEM cohort study.

Note

  • Many fields contain substantial missing data, especially the vital parameter measurements

  • Triage levels are generally unreliable, with 95%+ of walk-ins marked as Green and completely unreliable data from Helsingborg and Ystad hospitals

  • Visit reasons (chief_complaint) are known to be inaccurate in many cases

  • Vital parameter values may not represent the first measurement taken

  • The death_days_after_visit metric is set to 0 if discharged_to is "Avlidit" (Deceased)

  • The referral_elsewhere field is missing data for approximately 50,000 records

  • Visit IDs may correlate with the "Nyckel" ID in other datasets

  • IMPORTANT: For most analyses, the patientliggaren dataset is recommended over SVAR due to its better linkage with other tables in the cohort study. SVAR should primarily be used when vital parameters are required, though their quality is dubious

  • Only 76.7% of SVAR records could be linked to a patient_id, further limiting its utility compared to the patientliggaren dataset

Examples