The V Lab

Alberta Health Data Atlas

A learning guide to nine commonly used Alberta administrative health datasets and clinical information systems: how each one is generated, what one record represents, which fields are typically available, and how to plan cohorts, linkage, access and responsible use. It is the reference Project 001 was designed from.

One care journey, many data traces

Each dataset sees a different part of a patient's path through the system. Choose sources by the part of the journey your question is about.

  1. 01

    Population & eligibility

    Who belongs in the cohort?

    REGISTRY

    The Population Registry gives a cohort a demographic and AHCIP-eligibility backbone, but it is not a census.

  2. 02

    Community & ambulatory

    Which services occurred?

    CLAIMSNACRS

    Claims capture billed services; NACRS captures emergency and selected ambulatory contacts. They represent different data traces.

  3. 03

    Emergency & inpatient

    How did care progress?

    NACRSDADCC / EPIC

    NACRS and DAD can describe the path from emergency care to admission; Connect Care adds richer clinical context.

  4. 04

    Tests & results

    What were the physiologic signals?

    PLDCC / EPIC

    Numeric results become comparable only after test codes, units, reference ranges, and source systems are harmonized.

  5. 05

    Community pharmacy

    Was medication dispensed?

    PIN

    PIN observes dispensing, not medication use. Match the research question to the behaviour the data can actually observe.

  6. 06

    Vital outcomes

    What happened over time?

    VITALDAD

    Mortality and hospitalization extend follow-up, but authorization, reporting lag, and small-cell disclosure must be designed in advance.

The nine datasets

Coverage dates marked * come from the UCalgary Centre for Health Informatics catalogue; confirm them with the data custodian before requesting data.

DADAdministrative data

Discharge Abstract Database

Captures demographic, diagnosis, intervention, administrative, and disposition information for acute inpatient separations—a core source for hospital outcomes and resource use.

Setting
Acute inpatient care (Alberta day surgery is generally in NACRS)
One record
One inpatient separation from a reporting facility
Coverage
CHI: Apr 2002–present*
Refresh
Monthly · ~1-month lag
Questions, linkage and caveats

Questions it can help answer

  • Which patients are readmitted within 30 days?
  • How does length of stay vary by procedure?

Linkage

Commonly linked with NACRS, Population Registry, Practitioner Claims, and Vital Statistics.

Caveat

One abstract is not one person or a complete episode of care; transfers, diagnosis types, and methodology years must be handled explicitly.

Source check

Alberta day surgery is generally reported through NACRS; confirm the CHI catalogue’s DAD day-surgery wording in the project specification.

Data dictionary notes

Official public workbook · 4 worksheets · 58 commonly requested rows · CCI and ICD reference tabs

Supports planning for separation, diagnosis, intervention, service, length-of-stay, and case-mix fields.

ConceptTypical fieldsNote
Record and linkageSEQNUM · ULI · PHN · CHARTNOSeparate abstract, person-linkage, and local chart identifiers.
Admission, discharge, and transferADMITDATE · DISDATE · DISP · INSTFROM · INSTTOA continuous transfer journey can create several abstracts.
Diagnoses and diagnosis typesDXCODE1–DXCODE25 · DXTYPE1–DXTYPE25Position and diagnosis type jointly determine interpretation.
Interventions and resourcesPROCCODE1–20 · ALL_DAYS · ACUTE_DAYS · CMG · RIWCCI, CMG, and RIW require documented methodology versions.

Reading order

  • State whether the analytic unit is a patient, stay, or transfer-linked episode.
  • Freeze ICD-10-CA, CCI, and grouper versions for the study period.
  • Confirm repeating-field limits and the final delivered structure.

Cautions

  • One DAD abstract is not one patient.
  • Alberta day surgery is generally reported through NACRS rather than the current DAD stream.
  • The public workbook is a request-planning aid, not the current custodian specification.
Data dictionary (UCalgary CHI)Source notesDeep dive
NACRSAdministrative data

National Ambulatory Care Reporting System

Covers emergency care, day surgery, urgent care, and selected outpatient or specialty-clinic visits; useful for studying emergency use and subsequent disposition.

Setting
Emergency · Urgent care · Selected outpatient clinics
One record
One emergency or ambulatory care visit
Coverage
Historical data from ~2002; NACRS format from 2010*
Refresh
Monthly · ~1-month lag
Questions, linkage and caveats

Questions it can help answer

  • How often do patients return to the emergency department within 72 hours?
  • Which visits result in hospital admission?

Linkage

Often linked with DAD, Population Registry, and Practitioner Claims to reconstruct care pathways.

Caveat

Strong ED and day-surgery reporting does not make every ambulatory setting complete; verify clinic, site, year, and submission level separately.

Source check

Alberta adopted the NACRS format on 2010-04-01; earlier AACRS/ACCS tables do not share one start date or field set.

Data dictionary notes

Official public workbook · 2 worksheets · 43 commonly requested rows

Covers core administrative and clinical elements for ED, day surgery, urgent care, and selected ambulatory activity.

ConceptTypical fieldsNote
Record, setting, and submissionSEQNUM · ABSTRACT_TYPE · MIS_CODE · ED_VISIT_INDICATORFirst distinguish true ED encounters from other ambulatory activity.
Triage and process timingTRIAGE_DATE/TIME · PIA_DATE/TIME · ED_DEPT_DATE/TIMEArrival, first assessment, disposition decision, and physical departure are different events.
Diagnoses and interventionsDXCODE1–10 · PROCCODE1–10Field depth depends on submission level and service type.
Disposition and resource groupingDISPOSITION · INSTTO · CACS_CODE · CACS_RIWThe full post-admission course usually requires DAD linkage.

Reading order

  • Define ED, day-surgery, and other ambulatory cohorts separately.
  • Specify whether return-visit timing starts at registration, disposition decision, or physical departure.
  • Handle AACRS/ACCS and NACRS separately in analyses spanning 2010.

Cautions

  • Strong ED coverage does not imply complete specialty-clinic coverage.
  • The current catalogue and older workbook use different historical start points.
  • Missing or out-of-order timestamps require prespecified protocol rules.
Data dictionary (UCalgary CHI)Source notesDeep dive
CLAIMSAdministrative data

Practitioner Claims

Derived from Alberta Health Care Insurance Plan service claims and payments; supports analysis of physician service use across inpatient, outpatient, and primary care settings.

Setting
Physician services · Primary care
One record
One billed service claim
Coverage
CHI: Apr 1993–present*
Refresh
Quarterly · ~4-month lag
Questions, linkage and caveats

Questions it can help answer

  • How many physician services do patients with chronic disease use in a year?
  • Does service use differ by region?

Linkage

Commonly linked with Population Registry, DAD, NACRS, and PIN.

Caveat

Billing activity does not establish disease occurrence and is not a complete medical record; Schedule of Medical Benefits, shadow-billing, and diagnostic-coding rules change over time.

Source check

CHI and the older Alberta overview use different historical start dates and completeness lags.

Data dictionary notes

Official public workbook · 2 worksheets · 22 commonly requested rows

Focuses on service dates, billed services, diagnoses, providers, delivery settings, and payment fields.

ConceptTypical fieldsNote
Recipient and timingRCPT_ULI · SE_START_DATE · SE_END_DATE · FISC_YRService dates must not be confused with processing or payment dates.
Service and diagnosisHLTH_SRVC_CCPX_CODE · HLTH_DX_ICD9X_CODE_1–3Use code-list versions aligned with the service date.
Provider and settingDOCTOR_CLASS · PRVD_SKILL_TYPE_CLS · DELV_SITE_TYPE_CLSProvider role, skill, and delivery setting are distinct concepts.
Payment and alternate plansFRE_ACTUAL_PAID_AMT · PGM_APP_INDPayment fields are affected by shadow-billing and program rules.

Reading order

  • Identify adjustments, reversals, duplicates, and payment status first.
  • Interpret utilization with the SOMB/service code set in effect during service.
  • Preserve claim-line grain before aggregating to person-time.

Cautions

  • A billed service is not a complete clinical encounter or disease truth.
  • The current catalogue says about four months; the older workbook says six months or longer.
  • Research access and internal AHS QI follow different request routes.
Data dictionary (UCalgary CHI)Source notes
PLDAdministrative data

Provincial Laboratory Data

Aggregates common clinical chemistry, hematology, serology, and other laboratory results; useful for studying biomarkers, disease states, and testing trajectories.

Setting
Inpatient and outpatient laboratory testing
One record
One laboratory test and result record
Coverage
CHI: Apr 2012–present*
Refresh
Daily · ~1-week lag
Questions, linkage and caveats

Questions it can help answer

  • How soon after an abnormal result does hospitalization occur?
  • Does testing frequency differ across population groups?

Linkage

Commonly linked with DAD, NACRS, Practitioner Claims, and Connect Care.

Caveat

Data were historically consolidated from multiple regional systems; harmonize test codes, units, reference ranges, and source systems before comparing raw values.

Source check

Confirm available tests, sites, time periods, and data custodian for each request.

Data dictionary notes

Official public workbook · 3 worksheets · 14 commonly requested rows · 10,872-row test-name reference

Documents person keys, order/test identifiers, results, units, reference ranges, abnormal flags, and verification time.

ConceptTypical fieldsNote
Person and orderCLNT_ULI · CLNT_PHN · ORDR_TEST_CODE_CD/NMPanel orders and individual test results must be distinguished.
Test identityTEST_CD · TEST_NMSource-system names are not inherently comparable across sites.
Result and unitTEST_RSLT · TEST_UOFM · TEST_ABNRML_FLAGValues require units and result qualifiers for interpretation.
Reference range and timingTEST_REF_RNG · TEST_REF_RNG_NRML_LOW/HIGH · TEST_VRFY_DTTMReference ranges can vary by age, recorded sex, method, and system.

Reading order

  • Build a versioned crosswalk of test codes, names, units, and source systems first.
  • Preserve text results, < or > qualifiers, and corrected-result status.
  • Specify collection, performance, verification, or load time as the analytic anchor.

Cautions

  • Historical laboratory systems differ in codes, units, and workflows.
  • Raw values cannot be compared directly across methods or reference ranges.
  • Point-of-care testing and some specialized domains may be incomplete.
Data dictionary (UCalgary CHI)Source notes
PINAdministrative data

Pharmaceutical Information Network

Research extracts focus mainly on community-pharmacy dispensing records, supporting analysis of dispensing patterns, refill persistence, and post-discharge medication trajectories.

Setting
Community pharmacy dispensing
One record
One community-pharmacy dispense record
Coverage
Jan 2008–present*
Refresh
Weekly · ~1-week lag
Questions, linkage and caveats

Questions it can help answer

  • Was a prescription dispensed after discharge?
  • How do refill patterns differ across medications?

Linkage

Commonly linked with DAD, Practitioner Claims, Population Registry, and Vital Statistics.

Caveat

A dispense does not prove that a medication was taken; do not assume that allergies, alerts, or prescribing functions in the clinical PIN application are present in research extracts.

Source check

Alberta currently identifies the available research asset as PIN Dispenses, with no financial information.

Data dictionary notes

Official public workbook · 4 worksheets · 11 commonly requested rows · DIN and ATC reference tabs

Focuses on dispense date, cancellation, DIN/ATC, quantity, unit, expected days supplied, and person linkage.

ConceptTypical fieldsNote
Person and eventRCPT_ULI · RCPT_DOB · DSPN_DATE · DSPN_CANCEL_DATEExclude or correctly classify retracted fills first.
Drug identityDRUG_DIN · SUPP_DRUG_ATC_CODEDIN identifies a product; ATC is a therapeutic classification.
Quantity and unitDSPN_AMT_QTY · DSPN_AMT_UNT_MSR_CDDispensed quantity alone is not dose.
Days suppliedDSPN_DAY_SUPPLY_QTY · DSPN_DAY_SUPPLY_UNT_MSR_CDExpected coverage days are not observed medication-taking days.

Reading order

  • Freeze DIN/ATC lists, hierarchy level, and code-set version.
  • Prespecify overlapping fills, stockpiling, grace periods, and switching.
  • Interpret inpatient gaps with DAD because institutional supply may be absent.

Cautions

  • Dispensing is not prescribing and does not prove use.
  • A research extract does not automatically include the full clinical PIN feature set.
  • Alberta Health currently states that PIN Dispenses contains no financial information.
Data dictionary (UCalgary CHI)Source notesDeep dive
VITALAdministrative data

Vital Statistics

Records live births, stillbirths, and deaths occurring in Alberta; supports all-cause and cause-specific mortality, perinatal outcomes, and population-outcome research.

Setting
Live birth · Stillbirth · Death
One record
One registered live birth, stillbirth, or death event
Coverage
CHI: 1983–present*
Refresh
Annual · ~1-year lag
Questions, linkage and caveats

Questions it can help answer

  • What is the all-cause mortality risk in a cohort?
  • How do perinatal outcomes differ by region?

Linkage

Frequently used as an outcome source for cohorts built from DAD, Practitioner Claims, and PIN.

Caveat

Vital Statistics requires separate authorization; cause-of-death coding, registration lag, and small-cell disclosure risk must be addressed.

Source check

CHI’s 1983 start date differs from Alberta’s current standard request page; confirm historical coverage in advance.

Data dictionary notes

Partial: Death Registry only · 2 worksheets · 21 commonly requested rows · current link covers Death Registry only

The current public workbook documents death date, place, demographics, linkage keys, autopsy, and underlying cause.

ConceptTypical fieldsNote
Event and registrationDR_ID · DETHDATE · YEAR · FISCAL_YREvent year, fiscal year, and registration updates must not be conflated.
Person linkageSTKH_NUM_1–5 · BIRTH_DATE · SEXA person may have primary and secondary ULIs.
Place and contextPL_DETH · HOSP_ID · POSTCODE · PL_INJURYPlace of death, residence, and injury are different concepts.
Cause of deathU_CAUSE · AUTOPSYUnderlying cause is not the same as all contributing causes.

Reading order

  • Record event date together with cutoff, registration lag, and revision status.
  • Define ULI linkage success and multiple-ULI handling.
  • Document the cause-of-death ICD version for each period.

Cautions

  • The public link is not a complete birth and stillbirth dictionary.
  • Vital Statistics has separate authorization and small-cell disclosure requirements.
  • The workbook states that cause-of-death coding is not ICD-10-CA.
Data dictionary (UCalgary CHI)Source notes
REGISTRYAdministrative data

Provincial / Population Registry

Provides basic demographic, plan, and geographic information for AHCIP-eligible residents; commonly used for cohort definition, denominators, and data linkage.

Setting
AHCIP eligibility · Demographics and geography
One record
One registrant record for an eligibility period
Coverage
CHI: Apr 1994–present*
Refresh
Annual · ~1-year lag
Questions, linkage and caveats

Questions it can help answer

  • Who had active coverage during the study period?
  • How can a cohort be stratified by age, sex, and geography?

Linkage

Serves as a cohort backbone for nearly all administrative health datasets.

Caveat

The insured population is not equivalent to the census population; migration, address recency, and eligibility changes affect denominators.

Source check

CHI and the older Alberta overview differ by roughly one year on the historical start date; use the version available to the project.

Data dictionary notes

Official public workbook · 2 worksheets · 15 commonly requested rows

Documents year-end coverage, age group, migration, birth/death indicators, geography, and linkage identifiers.

ConceptTypical fieldsNote
Linkage and identityASN · PHN · BIRTH_DT · SEXConfirm which transformed linkage key will be disclosed.
Year-end eligibilityACTIVE_COVERAGE · FYE · PERS_REAP_END_DATEA year-end snapshot is not continuous eligibility throughout the year.
Population changeBIRTH_IND · DEATH_IND · IN_MIGRATION_IND · OUT_MIGRATION_INDMigration and reporting delays alter denominators.
Geography and stratificationPOSTAL_CD · RHA · AGE_GRP_CDMailing address may not be actual residence.

Reading order

  • First confirm whether delivery is a year-end snapshot or eligibility interval.
  • Define continuous coverage, allowable gaps, migration, and person-time rules.
  • Align geography fields with boundary versions from the same period.

Cautions

  • The AHCIP registry is not a census.
  • The current catalogue says about one year; the older workbook says about six months.
  • The workbook warns that ALT_PREM_ARRANGEMENT quality declined after 2009.
Data dictionary (UCalgary CHI)Source notes
SCMClinical information systemLegacy system

Sunrise Clinical Manager

A historical Calgary Zone clinical information system containing demographics, orders, test results, and clinical notes; it should now be treated as a legacy source.

Setting
Historical Calgary inpatient EMR
One record
One clinical event, order, result, or note
Coverage
2011–confirm by site
Refresh
Legacy system · Confirm availability
Questions, linkage and caveats

Questions it can help answer

  • What clinical detail is available beyond administrative records?
  • Which notes are available for historical Calgary cohorts?

Linkage

Can be reconciled with DAD, NACRS, and PLD, but system migration must be understood.

Caveat

Connect Care progressively replaced SCM; access to historical data does not mean that new data continue to accrue.

Source check

CHI’s “current / no delay” wording should not be interpreted as real-time SCM accrual in 2026.

Data dictionary notes

No public field workbook · CHI currently links no public field workbook

This is a historical-system concept guide; actual tables, fields, sites, and cutoff dates require project-specific confirmation.

ConceptTypical fieldsNote
Person and encounterRequest-specificRequest the mapping of patient, encounter, ADT, and site keys.
Orders and resultsRequest-specificClarify relationships among orders, specimens, results, and revisions.
Documentation and medicationRequest-specificModules, text scope, and timestamp semantics must be requested explicitly.
Migration and cutoffRequest-specificRecord each Calgary site's migration date to Connect Care.

Reading order

  • Specify modules, entities, sites, years, and output grain in the data request.
  • Request the current field list, entity relationships, and value-set versions.
  • Validate SCM-to-Connect Care mappings during overlap periods.

Cautions

  • Historical availability does not imply ongoing capture.
  • Different modules do not share one universal record grain.
  • Do not treat CHI's older current/daily wording as a 2026 real-time status.
Source notes
CC / EPICClinical information system

Connect Care / Epic

A province-wide clinical information system connecting records, orders, results, and workflows; historical depth depends on each site’s go-live date.

Setting
Unified provincial clinical information system
One record
One event, order, result, or note within a clinical encounter
Coverage
Phased rollout, 2019–2024*
Refresh
Source system daily · Extract timing varies
Questions, linkage and caveats

Questions it can help answer

  • Which fine-grained clinical features characterize a cohort?
  • How do workflows and outcomes change before and after go-live?

Linkage

Can be linked with longitudinal administrative data and legacy EMRs.

Caveat

The phased 2019–2024 rollout creates important temporal and site-coverage bias; daily source refresh does not mean same-day research delivery.

Source check

AHS records the ninth and final launch on 2024-11-02, not full implementation in 2022 as stated on the CHI catalogue page.

Data dictionary notes

No public field workbook · CHI currently links no public research field workbook

This is an application-oriented system guide; projects must confirm Epic/Cogito sources, tables, fields, value sets, and extract-refresh rules.

ConceptTypical fieldsNote
Person, encounter, and ADTRequest-specificPreserve encounter context and transfer timelines.
Orders, results, and proceduresRequest-specificDistinguish ordered, performed, resulted, and verified times.
Medication and MARRequest-specificOrdering, dispensing, administration, and discontinuation are different events.
Notes, flowsheets, and provenanceRequest-specificDocument templates, copy-forward, and configuration versions.

Reading order

  • Build an observability matrix by site go-live date.
  • Define row grain, keys, timestamps, and statuses for each domain.
  • Confirm the Epic/Cogito layer and project-specific extraction logic.

Cautions

  • The 2019–2024 phased rollout creates site and time-coverage bias.
  • Daily source refresh does not mean same-day research-data delivery.
  • Workflow and local configuration changes alter how records are generated.
Source notes

Deep dives: DAD, NACRS and PIN

The three sources most often combined for emergency, inpatient and medication research.

DAD

Discharge Abstract Database

Acute inpatient separations

DAD turns the discharge abstract created at the end of a hospital stay into an analyzable record. It brings together demographic, admission, diagnosis, intervention, service, resource-use, and discharge information. It is especially useful for hospital events, procedures, length of stay, transfers, and readmissions—but it is not a minute-by-minute clinical chart. In Alberta, day surgery is generally reported through NACRS; confirm the project-specific extract.

One record
One inpatient separation from a reporting facility
Alberta catalogue range
Apr 2002–present; confirm before requesting
Core classifications
ICD-10-CA diagnoses · CCI interventions
Best suited to
Admissions, procedures, resource use, and discharge outcomes

What a record traces

  1. Admission and origin
  2. Diagnoses and interventions
  3. Services and resource use
  4. Discharge, transfer, or death

Field groups

GroupExample fieldsWhat it supports
Identity, demographics, and geographyULI / PHN · AGE_ADMIT · SEX · POSTCODE · RCPT_ZONESupports approved linkage and stratification by age, recorded sex/gender, and residence. Direct identifiers are normally removed, substituted, or scrambled for disclosure.
Admission timing and entryADMITDATE · ADMITTIME · ADMITCAT · ENTRYCODE · ADMITBYAMBDescribes when the inpatient stay officially began, admission category, point of entry, and ambulance arrival.
Discharge and destinationDISDATE · DISTIME · DISP · INSTFROM · INSTTODefines the index discharge date, discharge status, and inter-facility transfers. A continuous transfer episode may create several abstracts.
DiagnosesDXCODE1–DXCODE25 · DXTYPE1–DXTYPE25Repeated ICD-10-CA diagnosis codes and diagnosis types. A case definition should consider position and type—not merely whether a code appears.
Interventions and proceduresPROCCODE1–PROCCODE20Uses CCI codes to describe interventions during the stay, supporting procedure cohorts and comparisons of care pathways.
Services and provider rolesMPSERV · DOCSVC1–8 · DOCTYPE1–8Describes the main patient service, provider types, and services. These are role and service classifications—not a full care-team record.
Length of stay and special careALL_DAYS · ACUTE_DAYS · ICU_HOURS · CCU_HOURS · SCUHOURSSupports analysis of total and acute length of stay plus ICU or special-care use. Confirm derivation and missing-value rules first.
Case mix and resource intensityCMG · MCC · RIW · COMORB_LVLCIHI grouping and resource-weight measures support case-mix and utilization comparisons, but version changes affect comparability over time.

What it captures

  • Acute inpatient separation abstracts
  • Diagnoses, interventions, and major services documented during the stay
  • Admission, discharge, transfer, and length-of-stay information
  • In-hospital outcomes such as death at discharge

What it cannot tell you

  • Minute-by-minute vital signs, full lab trajectories, or clinical notes
  • Whether medication was dispensed or taken after discharge
  • Alberta day surgery, non-admitted ED care, and other ambulatory activity; generally in NACRS
  • Severity and causality that cannot be established from codes alone
  • Actual patient-level cost; RIW is relative resource intensity, not a bill

A fictional example record

FieldValueMeaning
ADMITDATE2025-04-12Official admission date
DISDATE2025-04-16Index discharge date
DXCODE / DXTYPEI50.- / main diagnosis (illustrative)Heart-failure code family and its role
DISPHome (illustrative)Discharge disposition
ALL_DAYS4 daysDerived total length of stay

Values are invented for illustration and do not describe any real person.

Study design examples

Example A · 30-day readmission

What proportion of patients discharged alive after a heart-failure admission have an unplanned readmission within 30 days?

Cohort
Identify adult heart-failure admissions using a validated ICD-10-CA algorithm and require live discharge.
Index
The qualifying discharge date for each patient, after consolidating a continuous transfer chain into one episode.
Outcome
The next qualifying DAD admission on days 1–30.
Linkage
Use the Registry for eligibility, Vital Statistics for deaths, and NACRS for ED returns that do not become admissions.

The key is not merely finding the next DAD record; it is defining transfers, planned admissions, competing death, and the per-person index rule.

Example B · Resource use after hip replacement

Do length of stay and 30-day acute-care reuse after elective hip replacement differ across hospitals?

Cohort
Define elective procedures using CCI intervention codes, admission category, and age criteria.
Exposure
Treating facility or zone, subject to facility and small-cell disclosure rules.
Outcomes
ALL_DAYS or ACUTE_DAYS, in-hospital disposition, and 30-day DAD readmission or NACRS ED visit.
Adjustment
Age, diagnosis type, comorbidity, transfers, and case-mix methodology version.

Facility differences may reflect case mix, transfers, and coding practice; an observed difference is not automatically a causal quality effect.

Common pitfalls

  • Treating one abstract as one person; the same person can have many admissions.
  • Searching diagnosis codes without considering diagnosis type, position, and a validated case algorithm.
  • Counting an inter-facility transfer as a readmission or splitting one episode into independent events.
  • Mixing admission year, discharge year, and fiscal year and misaligning observation windows.
  • Comparing CMG, RIW, ICD, or CCI across years without recording version and field-rule changes.

Sources

NACRS

National Ambulatory Care Reporting System

Emergency, day surgery, and selected ambulatory care

NACRS records one hospital- or community-based ambulatory encounter. In Alberta, its most stable research uses are generally ED and day-surgery activity. It can connect arrival, triage, physician initial assessment, diagnoses, interventions, and visit disposition, making it useful for ED flow, waits, revisits, and admission. Completeness must be checked separately by care setting, site, year, and submission level.

One record
One ED, day-surgery, or other ambulatory visit
Alberta catalogue range
Historical data from ~2002; NACRS format from 2010
Core classifications
ICD-10-CA · CCI · CED-DxS / EDVS
Best suited to
ED flow, waiting, revisits, and disposition

What a record traces

  1. Registration and arrival
  2. Triage and first assessment
  3. Diagnoses and interventions
  4. Discharge, transfer, or admission

Field groups

GroupExample fieldsWhat it supports
Visit type and modeABSTRACT_TYPE · Submission Level · VISIT_MODE · ED_VISIT_INDICATOR · MIS_CODEDistinguishes ED, day surgery, clinic, scheduled/arranged activity, and contact mode. Separate care settings before combining them.
Registration and departure timingVISIT_DATE · VISIT_TIME · ED_DEPT_DATE · ED_DEPT_TIMEForms the basis for total ED length of stay and encounters crossing midnight. Missing or out-of-order timestamps require quality rules.
TriageTRIAGECODE · TRIAGE_DATE · TRIAGE_TIMERecords triage level and timing for case mix and wait analyses. Clinical severity should not be inferred from a single triage code alone.
Physician initial assessmentPIA_DATE · PIA_TIMECan be combined with registration or triage time to measure time to physician assessment, provided the start point and missingness rules are explicit.
Diagnoses and presenting complaintDXCODE1–DXCODE10 · Presenting Complaint · ED Discharge DiagnosisIncludes ICD-10-CA problem codes and CIHI ED pick-lists. Presenting complaint, ED discharge diagnosis, and final inpatient diagnosis are different concepts.
Interventions and servicesPROCCODE1–10 · PROVIDER_TYPE1–8 · PROVIDER_SVC1–8Describes recorded interventions and provider/service roles. The submission level can affect the amount of detail available.
Disposition and facility flowDISPOSITION · DISP_DATE · DISP_TIME · INSTFROM · INSTTODistinguishes discharge home, transfer, incomplete care, or admission and supports construction of a care episode.
Length of stay and case mixVISIT_LOS_MINUTES · ED_ER_MINUTES · CACS_CODE · CACS_RIW · MACSupports ED length-of-stay and relative-resource comparisons. Derived values, grouper versions, and valid timestamps all require verification.

What it captures

  • Alberta ED and day-surgery activity plus selected other ambulatory care
  • Registration, triage, physician initial assessment, and departure timing
  • Visit problems, diagnoses, interventions, and disposition
  • Ambulance arrival, transfers, and clues to inpatient admission

What it cannot tell you

  • It does not guarantee uniform completeness across clinics, sites, and years
  • The full inpatient stay after ED disposition; link to DAD
  • Detailed clinical notes, complete lab results, and medication administration
  • Waiting time alone cannot establish quality or causality

A fictional example record

FieldValueMeaning
VISIT_TIME14:08Registration time
TRIAGECODELevel 3 (illustrative)Triage level
PIA_TIME14:47Physician initial assessment
DISPOSITIONDischarged home (illustrative)Visit disposition
VISIT_LOS_MINUTES210 minutesDerived total visit length

Values are invented for illustration and do not describe any real person.

Study design examples

Example A · 72-hour ED revisit

Among ED patients discharged home, what proportion make an unplanned return visit within 72 hours?

Cohort
Select linkable, true ED index visits ending in discharge and exclude arranged visits.
Index
The date and time the patient left the ED, not registration date alone.
Outcome
The next qualifying NACRS ED record within 72 hours after departure.
Linkage
Use DAD for subsequent admission, Vital Statistics for death, and the Registry for eligibility.

Define “revisit” first: same facility or province-wide, arranged or unplanned, and which timestamp starts the 72-hour clock.

Example B · Acute care after day surgery

After adult day surgery, does an ED visit occur within 7 days or an inpatient admission within 30 days?

Cohort
Define day surgery with ABSTRACT_TYPE, the MIS setting, and a validated main CCI intervention; require discharge alive to home.
Index
The day-surgery service date. Hour-level work requires the applicable end-time field and must not automatically reuse an ED departure field.
Outcomes
A true NACRS ED encounter on days 1–7 and a DAD acute admission on days 1–30.
Linkage
Link the NACRS day surgery to later NACRS ED care, DAD admissions, and Registry eligibility; exclude a direct same-episode transfer.

Do not misclassify planned follow-up or another ambulatory contact as an ED return; record CCI and CACS grouper versions across years.

Common pitfalls

  • Treating high ED coverage as universal completeness; CIHI still identified incomplete periods at 19 Alberta facilities in 2024–2025, representing about 900 missing ED records.
  • Ignoring submission level, site, year, and Connect Care flow changes and treating a missing field as an absent event.
  • Treating presenting complaint, ED discharge diagnosis, and final inpatient diagnosis as equivalent.
  • Using dates without times and misclassifying overnight revisits, transfers, or waits.
  • Counting admission from the ED as a new independent event instead of reconciling the episode with DAD.

Sources

PIN

Pharmaceutical Information Network Dispenses

Alberta community-pharmacy dispensing

The PIN asset available for research is primarily PIN Dispenses: one record represents a dispense from a community pharmacy. It can show when a product was dispensed, its product coding, quantity, days supplied, and cancellation information, supporting studies of post-discharge fills, refills, and persistence. It observes dispensing—not prescribing intent, actual consumption, or inpatient administration.

One record
One community-pharmacy dispense event
Alberta request range
2008–present; PIN Dispenses
Product classifications
DIN product identifier · ATC therapeutic class
Best suited to
Fills, refills, days covered, and persistence

What a record traces

  1. Identify the target drug
  2. Dispense date and quantity
  3. Days supplied and overlap
  4. Refill, gap, or cancellation

Field groups

GroupExample fieldsWhat it supports
Approved linkage keyRCPT_ULISupports approved person-level linkage to DAD, Claims, or the Registry. Research disclosures normally do not retain the raw direct identifier.
Basic demographicsRCPT_DOB · RCPT_GENDER_CDCan support age and recorded sex/gender checks. Analyses often prefer specification-confirmed demographic fields from the Registry.
Dispense dateDSPN_DATEThe core timing field for first fills, post-discharge windows, and refill sequences.
Cancellation or reversalDSPN_CANCEL_DATEIdentifies a dispense that was later retracted. Ignoring cancellations can count an invalid event as a fill.
Product identifierDRUG_DINThe DIN is an 8-digit identifier for a specific authorized drug product and should be stored as text to preserve leading zeros. Longitudinal work needs a versioned product list and rules for new or discontinued products.
Therapeutic classificationSUPP_DRUG_ATC_CODEATC is a five-level therapeutic hierarchy. Pre-specify the analysis level and version; it is not interchangeable with a DIN.
Dispensed quantity and unitDSPN_AMT_QTY · DSPN_AMT_UNT_MSR_CDQuantity must be interpreted with its unit and formulation. Tablets, millilitres, and devices are not directly comparable scales.
Days suppliedDSPN_DAY_SUPPLY_QTY · DSPN_DAY_SUPPLY_UNT_MSR_CDOften used to construct days covered, gaps, and persistence, with explicit rules for missing/extreme values, early refills, and overlaps.

What it captures

  • Dispense events reported by Alberta community pharmacies
  • Product DIN, ATC class, quantity, and unit
  • Dispense date, days supplied, and cancellation date
  • Longitudinal refill sequences after approved linkage

What it cannot tell you

  • The source prescription and prescribing intent; no dispense cannot distinguish not prescribed from not filled
  • Whether the patient took the medication as directed
  • Inpatient, institutional, continuing-care, public-health, and other non-community supply
  • Drug cost or payment information in the current Alberta research asset

A fictional example record

FieldValueMeaning
DSPN_DATE2025-04-18Community-pharmacy dispense date
DRUG_DINExample DINSpecific drug product
SUPP_DRUG_ATC_CODEC09… (illustrative)Therapeutic class
DSPN_AMT_QTY30 tablets (illustrative)Quantity interpreted with its unit
DSPN_DAY_SUPPLY_QTY30 daysEstimated coverage period

Values are invented for illustration and do not describe any real person.

Study design examples

Example A · Fill within 7 days of discharge

After a heart-failure hospitalization, does the patient fill the target drug class within 7 days of discharge?

Cohort
Adults meeting a validated DAD case definition, discharged alive, and observable during follow-up.
Index
The DAD discharge date, with an explicit rule for whether day 0 counts.
Outcome
The first non-cancelled PIN dispense on days 0–7 whose DIN or ATC belongs to a pre-versioned target list.
Linkage
Link DAD to PIN using the approved ULI; use the Registry for eligibility and Vital Statistics for death.

Describe the outcome as an observed dispense—not treatment initiation or medication adherence.

Example B · 180-day refill persistence

After a new start in a drug class, what proportion have a gap of more than 30 uncovered days within 180 days?

New users
Require a sufficient washout period with no valid dispense in the target DIN or ATC class.
Coverage
Build coverage intervals from DSPN_DATE plus days supplied and pre-specify how early refills carry forward, truncate, or overlap.
Outcome
The first gap longer than 30 uncovered days within 180 days, with separate rules for switching within or outside the class.
Competing events
Death, out-migration, prolonged hospitalization, and loss of eligibility can make later community dispensing unobservable.

Persistence estimates are highly sensitive to rules for days supplied, cancellations, overlaps, switching, and unobservable time.

Common pitfalls

  • Describing a dispense as consumption or adherence; PIN cannot directly observe ingestion.
  • Failing to exclude cancellations/reversals or reconcile same-day duplicate dispenses.
  • Treating quantity as dose or ignoring units such as tablets, millilitres, and devices.
  • Interchanging DIN and ATC, or applying a current list to all years without preserving hierarchy, leading zeros, and version.
  • Calling an inpatient period a treatment gap when inpatient administration is generally outside PIN Dispenses.

Sources

Six steps for designing a study

From defining the unit of analysis to building authorization into the design.

  1. Separate people, records, and events

    One person can leave many records during one care journey. Define the analytic grain before counting people, visits, or rates.

    Example: one emergency visit followed by admission may appear in both NACRS and DAD.

  2. Break the question into P-I-O-T

    Specify the Population, Index event, Outcome, and Time window, then place the exposure or comparison group within that structure.

    Example: age 65+ / discharge date / 30-day readmission / 2019–2024.

  3. Test feasibility with the data dictionary

    Map every concept to fields, code systems, available years, and missing-data rules. A field can exist without being fit for the question.

    Example: diagnosis position and type can change a case definition.

  4. Design linkage and observation windows

    Define keys, the index date, look-back, follow-up, and duplicate handling. The linkage rate is itself a quality result.

    Example: use the Population Registry to assess eligibility during follow-up.

  5. Recognize migrations and bias

    Site go-lives, coding changes, billing incentives, and geographic coverage can create artificial trends. Map data generation before interpreting a curve.

    Example: Connect Care rolled out in phases from 2019 to 2024.

  6. Build authorization into the design

    Minimum-necessary fields, secure environments, output review, and destruction plans are part of the method—not closing paperwork.

    Example: REB approval does not itself authorize a custodian to disclose data.

Research scenarios

Three common questions, the datasets each needs, and the trap to avoid.

Admission after ED discharge

Who is admitted to hospital within 30 days after discharge from the emergency department?

NACRSDADREGISTRY
Cohort
People with an identifiable NACRS emergency-discharge record in the study period and 30 observable days afterward.
Timeline
Index: emergency discharge → Follow-up: days 1–30 → Outcome: DAD inpatient admission.
Roles
NACRS: index emergency visit
DAD: subsequent admission
Registry: eligibility and demographic strata

Watch for: A direct transfer from emergency to inpatient care is not a post-discharge admission. Exclude the same episode of care and consider death as a competing event.

Post-discharge dispensing

Was the target medication dispensed within 7 days after a heart-failure discharge?

DADPINREGISTRY
Cohort
Adults discharged alive from a DAD admission that meets a validated case definition.
Timeline
Index: discharge date → Follow-up: days 0–7 → Outcome: target medication dispensed in PIN.
Roles
DAD: admission and diagnosis
PIN: dispensing event
Registry: eligibility and geography

Watch for: Dispensing does not prove medication use, and community-pharmacy records do not fully reflect medications administered in hospital.

Chronic-disease utilization

Has physician-service and emergency use changed over time among people with diabetes?

CLAIMSNACRSDADREGISTRY
Cohort
Cases identified with a validated multi-year diagnosis algorithm and an adequate look-back period.
Timeline
Look-back: 2-year case identification → Annual windows → Measures: services and emergency visits per person-year.
Roles
Claims: community services
NACRS: emergency care
DAD: inpatient care
Registry: denominator

Watch for: Changes in coding and billing practice can mimic changes in prevalence or utilization; annual denominators must account for migration and death.

Responsible-use checklist

What to settle at each stage, from application to sharing results.

Before applying

  • Clearly define whether the purpose is research, quality improvement, planning, or operations, and confirm the applicable approval pathway
  • Confirm the current data custodian, application route, and available data years
  • Determine whether approval from a designated Research Ethics Board is required and which agreement applies
  • Request only the minimum data elements, shortest time period, and smallest cohort needed to answer the question

Before analysis

  • Map each concept to data-dictionary fields, coding systems, and versions
  • Document the extraction date, refresh lag, site coverage, and system migrations

During analysis

  • Work only in the approved secure environment, and do not attempt to re-identify individuals
  • Quantify missingness, linkage failures, coding changes, and bias arising from population coverage

Before sharing

  • Aggregate results as required, apply small-cell suppression rules, and complete any output disclosure review
  • Report the data-generation process, limitations in comparability, and uncertainty
  • Retain, archive, or destroy the data and derived files as required by the agreement

Where to go next

Public data to practise on, official access routes and the rules that govern them.

Open

  • Alberta Interactive Health Data Government of Alberta

    Use public aggregate tables, maps, and dashboards to practise interpreting rates, trends, geographic comparisons, and health inequities.

  • Alberta Open Government Government of Alberta

    Find downloadable public health datasets and publications for practice in data cleaning, citation, and reproducible analysis.

  • CIHI Data Holdings Canadian Institute for Health Information

    Review the pan-Canadian coverage of holdings such as DAD and NACRS, and use public indicators for interprovincial comparisons.

Application required

  • Catalogue of 9 Commonly Used AHS Datasets UCalgary Centre for Health Informatics

    The starting point for this site: dataset summaries, CHI metadata, and selected data dictionaries.

  • Health data access Government of Alberta

    Review current administrative health-data holdings, application pathways for research and QI, data-specification requirements, and the disclosure process.

  • Requesting AHS Data Resources Alberta Health Services

    Learn the requirements for analyst-prepared extracts, direct system access, REB review, and applicable data-disclosure agreements.

Secure environment

  • CIHI Custom Data Request Canadian Institute for Health Information

    Request customized aggregate or record-level data; record-level analysis generally takes place in CIHI’s Secure Access Environment.

  • Statistics Canada RDCs Statistics Canada

    After project and security approval, use microdata to study social determinants of health and health inequities in Alberta.

Guidance

  • Health Information Act Government of Alberta

    Understand foundational rules concerning custodians, the minimum-necessary principle, non-identifying health information, disclosure for research, and data matching.

  • ARECCI Decision Support Tools Alberta Innovates

    Helps distinguish research, quality improvement, and evaluation; results from the tool do not constitute formal ethics approval.

Glossary

Grain
What one row or analytic unit represents: a person, encounter, stay, service, or test.
Index event
The event or date that starts an observation window, such as a discharge.
Look-back
Time before the index event used to identify history, exposures, or eligibility.
Follow-up
Time after the index event during which outcomes are observed.
Coverage
How well the data capture the target population, sites, care settings, and years.
Reporting lag
Time from an event to its appearance in an analytic source; not the turnaround time for a data request.
Linkage
Connecting the same person or event across sources using approved identifiers or keys.
Non-identifying information
Information from which identity cannot be readily determined; applicable agreements and risk controls still apply.
REB
A board that reviews research ethics; REB approval does not itself authorize a data custodian to disclose data.
DDA
An agreement governing approved purpose, access, retention, disclosure, and destruction of data.

An independent educational resource. It contains no patient-level data, does not copy the official UCalgary workbooks, and does not replace legal, ethics, custodian or data-access advice. It is not an official publication of Alberta Health Services, the University of Calgary, the Government of Alberta or CIHI. Source repository: HealthcareData_Knowledge.