To advance and promote the discipline and professional practice
of public health epidemiology in Ontario

Hospitalization Data

The Discharge Abstract Database (DAD) captures administrative, clinical and demographic information on hospital discharges (including deaths, sign-outs and transfers).  


Original Source:

Discharge Abstract Database (DAD), Canadian Institute for Health Information (CIHI)

Distributed By:

Ontario Ministry of Health and Long-Term Care (MOHLTC): IntelliHEALTH ONTARIO (IntelliHEALTH)

Suggested Citation:

Inpatient Discharges [years], intelliHEALTH ONTARIO, Ontario Ministry of Health and Lont-Term Care, Date Extracted: [date].

Note: Specify in the report whether in-patient discharges or day procedures or both.  

See Data Citation Notes.

 

Data Notes

Background
  • Hospital separation records are historically the most comprehensive and accessible source of morbidity information. Other components of morbidity include visits to emergency departments and physicians. Physician visits or Medical Services data (from OHIP Profiles System Detailed Claims files) are also available through intelliHEALTH, starting at 2001. Emergency department visits from the National Ambulatory Care Reporting System (NACRS) are available through intelliHEALTH with data starting at 2002. Note: NACRS also contains a ‘Disposition Status' indicator - which indicates the flow of patients after they leave the emergency department, e.g. if the patient is discharged home or admitted to hospital. Disposition status equal to 6 or 7 (transfers to inpatient care in the reporting hospital) can be used to provide more timely data on hospitalization rates since hospital stays are only reported at discharge. It also allows reporting of hospitalizations for patients that are admitted to an adult psychiatric bed.
  • A separation may be due to death, discharge home, or transfer to another facility.
  • The main diagnostic code gives the primary reason for the hospital stay or "most responsible diagnosis". Subsequent diagnosis codes are not as reliable. However, depending upon the query, all codes may need to be included.
  • A second set of codes, external cause or "e-codes" are used to classify the environmental events, circumstances and conditions that cause an injury, e.g., motor vehicle traffic injury. While the ecodes are the principal means for classifying injury deaths, they are not used as a MRD for hospitalizations so they need to be examined separately.
  • Prior to March 31 2001, the MRD and e-codes were coded using the Ninth Revision of the International Classification of Diseases (ICD-9). Since April 1, 2001, the MRD and e-codes were coded using the Tenth Revision of the International Classification of Diseases Canada (ICD-10-CA). Comparison of trends for specific causes of hospitalization from 2001 onward with earlier rates must therefore be interpreted with caution.
  • As of April 2003, Day Procedures/Surgery are captured in the National Ambulatory Care Reporting System (NACRS).
  • Co-morbidity contributes uncertainty to classifying the most responsible diagnosis.
  • Since a person may not be hospitalized, or may be hospitalized several times for the same disease or injury event, or discharged from more than one hospital (when transferred) for the same injury event, hospitalization data provide only a crude measure of the prevalence of a cause.
  • Data are influenced by factors that are unrelated to health status such as availability and accessibility of care, and administrative policies and procedures. This may influence comparisons between areas and over time.
  • Data are collected based on location of hospital but are generally analyzed by the residence of the patient for health status purposes.
  • Data can be analyzed by fiscal or calendar year. Calendar year is generally used by public health units for health status purposes.  
  • Ontario residents treated outside of the province are excluded. Although less than 0.5% of all procedures performed for Ontario residents are out-of-province, areas bordering other provinces may be more affected.The IntelliHEALTH licensing agreement does not require suppression of small cells, but limits reporting at a level that could identify individuals, e.g. reporting at the postal code level by age and sex, regardless of the cell size. Aggregation (e.g. combining years, age groups, categories) should also be done when small numbers result in unstable rates.
  • Ontario residents treated outside of the province are excluded. Although less than 0.5% of all procedures performed for Ontario residents are out-of-province, areas bordering other provinces may be more affected.
  • The IntelliHEALTH licensing agreement does not require suppression of small cells, but limits reporting at a level that could identify individuals, e.g. reporting at the postal code level by age and sex, regardless of the cell size. Aggregation (e.g. combining years, age groups, categories) should also be done when small numbers result in unstable rates.

Live Births and Stillbirths
  • Live births and stillbirths can be identified in hospitalization data by the Entry code = "N” for newborn or "S” for stillbirth. The date of admission should be used rather than discharge date because it will correspond to the birth date. Newborns with a long length of stay could have a discharge date very different from the birth date. Live birth records begin in 1997 calendar year in intelliHEALTH data but stillbirth records are only available beginning with the 2002 calendar year. Note that hospital data are updated by fiscal year of discharge; births where the calendar year of admission is different than the fiscal year of discharge (e.g. born in March, but discharged in April) will not be included in the earlier year of data.   Because of this, the previous year’s total births may be higher, if the data are run again after the subsequent year’s data update.  

Obstetric Deliveries
  • Obstetric deliveries include births and stillbirths, and count multiple births as one delivery. Because the number of multiple births and stillbirths is relatively small, the number of deliveries is close to the number of newborns. Obstetric deliveries can be identified through a variety of ways, but ideally using any diagnosis of ICD10 code Z37. In the past, other methods have been used including Main Patient Service = 51 (OBS DELIVERED) or CMG codes, both of which appear to undercount deliveries. Z-codes have been in use since fiscal year 2002; as a result, counts of obstetric deliveries are available as of calendar year 2003. 
  • For obstetric deliveries, the Z37 codes are:
    • Z37 Outcome of delivery
    • Z37.0 Single live birth
    • Z37.1 Single stillbirth
    • Z37.2 Twins, both liveborn
    • Z37.3 Twins, one liveborn and one stillborn
    • Z37.4 Twins, both stillborn
    • Z37.5 Other multiple births, all liveborn
    • Z37.6 Other multiple births, some liveborn
    • Z37.7 Other multiple births, all stillborn
    • Z37.9 Outcome of delivery, unspecified
  • Note that there are no decimal places for ICD-10 codes in IntelliHEALTH, i.e., Z37.1 is Z371.
  • Information about whether the birth occurred through spontaneous ovulation and conception or through Artificial Reproductive Technologies (ART) is also available through these Z codes, beginning with the 2009 fiscal year (first complete calendar year is 2010).
  • Hospital records for newborns and mothers are separate. As a result, it is not possible to do a cross-tabulation of information such as birth weight (from the newborn record) with mother’s age (from the maternal record). Although there is the potential to link the records, linking is not complete, particularly for older data.


Analysis Checklist

General Checklist for All Associated Indicators
  • Consider aggregation of data values and/or cell suppression when dealing with small numbers to avoid risk of confidentiality breach. A new resource is currently under development to provide more detailed information on this issue.
  • For in-patient separations in PHPDB (through Intellihealth): use Inpatient Discharge Main Table data source from the ‘05 Inpatient Discharges' folder.
    • Filter for Admit Entry Type not equal to ‘N' and ‘S' - to exclude healthy newborns and stillbirths from your counts, as ICES and CIHI do for their publications.
    • Filter for Hospital Type = AT (Acute Treatment) or AP (Acute Psychiatric) - to include only acute care hospitals.
    • Filter for ICD10 Chapter not equal to 5 (V Mental and Behavioural Disorders) and ICD9 Chapter not equal to 5 (V Mental Disorders) - to exclude stays with a mental disorder as the most responsible diagnosis.
    • Select # Dschgs measure.
  • In your report, select the appropriate geography of patient (public health unit, LHIN) and use the pre-defined filter to select your area before you run your report. Hospital information (hospital name, PHU or LHIN) can also be selected in your report.
  • Note: inpatient data are reported by fiscal year (April 1 - March31). Any changes in the source data occur on a fiscal year basis (e.g., ICD10 reporting began on April 1 2002) and will affect reporting by calendar year.

References

General References
  • Canadian Institute for Health Information. Abstracting Manual. Ottawa, ON: Canadian Institute for Health Information, 1995.
  • Appendix: A Summary of Studies on the Quality of Health Care Administrative Databases in Canada. In: Goel V, Williams JI, Anderson GM, Blackstien-Hirsch P, Fooks C, Naylor CD, editors. Patterns of Health Care in Ontario. The ICES Practice Atlas, 2nd edition. Ottawa: Canadian Medical Association; 1996. p. 339-45.3.
  • Coyte PC, Young W, Williams JI. Devolution of Hip and Knee Replacement Surgery? ICES Working Paper, 1995;38;1-35.

Revision History

 This Core Indicator Product webpage is maintained by the Health Outcomes Subgroup.  
Date Review Type Author Changes PDF
September 20, 2012 Date of Last Revision

December 9, 2019]
Website Update:
No Content Review
Sherri Deamond,
on behalf of the CIWG
    Migrated to new website structure and format, including:
  • Reorganized content to provide high-level information at a glance, and move in-depth analytic information into dedicated sections for users to access when needed.  
  • Added short data source description in header.  
  • Moved data-specific Analysis Checklist items from the indicator pages to reduce duplication.  
  • Added Revision History table, with PDF copy of previous version for reference.  

     
APHEO's Core Indicators Project has been developed through collaboration across the field of public health in Ontario, 
to provide standardized methodology for population health assessment, to measure complex concepts of individual and community health.
Please contact core.indicators@apheo.ca for further information.