Data Source

The Data Source | Limitations of the data | Scope of the database

The Data Source

The source of these data is the AIHW National Hospital Morbidity Database. Other data on the  principal diagnoses and DRGs of hospitalised patients is available in Australian Hospital Statistics 2007-08 and Mental Health Services in Australia 2006-07. Some data have been revised since previously published. For confidentiality, some data have been suppressed. 

The National Hospital Morbidity Database is compiled by the Institute from data supplied by the State and Territory health authorities. It is a collection of electronic confidentialised summary records for admitted patients separated from public and private hospitals in Australia in the years 1993-94 to 2007-08. The total number of records for 2007-08 was 7.9 million.

Almost all hospitals in Australia are included: public acute, public psychiatric hospitals, private acute and psychiatric hospitals, and private free-standing day hospital facilities. See the notes on the Scope of the database for further information.

The National Health Data Dictionary definitions form the basis of the database, ensuring a high standard of data comparability. Definitions that are used in the interactive data presented on this Internet site are explained on the definitions page.

Limitations of the data

• Although the National Health Data Dictionary forms the basis of the data presented here, the actual definitions used may have varied among the data providers and from one year to another. Also, fine details of the scope of the collection have varied from year to year.
• Not all private hospital separations are included in the National Hospital Morbidity Database. In 2001-02, there were about 118,064 separations (4.6%) fewer private hospital separations reported to the Database than to the Australian Bureau of Statistics' Private Health Establishments Collection, which has wider coverage.
• Records for each financial year are for hospital separations (discharges, transfers, deaths or changes in care type) which occurred in the period 1 July to 30 June. For example, for 2002-03 this includes records where the separation date occurred between 1 July 2002 and 30 June 2003. Data on patients who were admitted on any date before 1 July 2002 are included, provided that they also separated between 1 July 2002 and 30 June 2003. A record is included for each separation, not for each patient, so patients who separated more than once in the year have more than one record in the database.
• Data are not generally available on the number of patients who receive admitted patient care each year. This is because information is not generally available to determine how many multiple admissions occur for individual patients who have multiple admissions, for example for chronic conditions.
• Separations do not always represent periods of 'hospitalisation' for patients because a new separation record follows a change in care type (which can occur with a transfer from, for example, a medical ward to a rehabilitation unit within a hospital), or a transfer from one hospital to another (for example from an acute care hospital to a rehabilitation hospital). In 2002-03, there were 66,911 separations that began with a 'statistical' admission following a change in care type (1.0% of the total). There were also 256,983 separations that began with a transfer of an admitted patient from another hospital (3.9% of the total). If the 'hospitalisations' had not been split into more than one separation with a 'statistical admission', the average length of stay would have been calculated as about 3.6 days, instead of 3.5 days. The average length of stay would have been calculated as about 3.7 days if they had not been split with a transfer from one hospital to another, or if they had not been split by either of these types of transfer.
• Patient day statistics can be used to provide information on hospital activity that, unlike separation statistics, account for differences in length of stay. As the database contains records for patients separating from hospital during the year, this means that not all patient days reported will have occurred in the reporting period (for example 1 July 2002 to 30 June 2003 for the 2002-03 data). It is expected, however, that patient days for patients who separated during the current financial year, but who were admitted in the previous financial year, would be counterbalanced overall by the patient days for patients who remained in hospital on 30 June and who will separate in future reporting periods. The numbers of separations and patient days can be a less accurate measure of the activity for establishments such as public psychiatric hospitals, and for patients receiving care other than acute care, for which more variable lengths of stay are reported.
• The number and pattern of hospitalisations can be affected by differing availability of other health care services. They can also be influenced by admission practices, which can vary among health service providers and over time. For example, over the past few years there has been a gradual reclassification of chemotherapy patients from admitted patients to non-admitted patients (outpatients) in New South Wales public hospitals.
• Although data on separations from the National Hospital Morbidity Database can reflect an aspect of the burden of disease in the community, they do not usually provide measures of the incidence or prevalence of conditions. This is because not all persons with a type or severity of illness are treated in hospital, and patients can have multiple admissions for some chronic conditions.
• For 1993-94 to 1997-98, diagnoses were recorded using ICD-9-CM, the Australian Version of the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). For 1998-99, they were recorded using ICD-9-CM by South Australia, Western Australia, Queensland and Tasmania, and by New South Wales, Victoria, the Australian Capital Territory and the Northern Territory using the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM). The Institute mapped ICD-9-CM data provided for 1998-99 forward to ICD-10-AM. The data for 1998-99 may therefore not be exactly comparable with data for 1993-94 to 1997-98. For 1999-00, diagnoses were reported to the National Hospital Morbidity Database for all states and territories except South Australia, using the first edition of ICD-10-AM (National Centre for Classification in Health 1998). South Australia mapped the data collected using the 1st Edition of ICD-10-AM forward to codes of the 2nd Edition of ICD-10-AM before providing them to the Institute. Where mapped codes could be identified (because they were invalid 1st Edition codes), the Institute mapped the South Australian data backward to 1st Edition codes so that national data could be presented in a single classification. For 2000-01, diagnoses were reported to the National Hospital Morbidity Database for all states and territories using the second edition of ICD-10-AM (National Centre for Classification in Health 2000). For 2001-02, diagnoses were reported to the National Hospital Morbidity Database for all states and territories except South Australia, using the second edition of ICD-10-AM. South Australia mapped the data collected using the 2nd Edition of ICD-10-AM forward to codes of the 3rd Edition of ICD-10-AM (National Centre for Classification in Health 2002) before providing them to the Institute. Where mapped codes could be identified (because they were invalid 2nd Edition codes), the Institute mapped the South Australian data backward to 2nd Edition codes so that national data could be presented in a single classification. For 2002-03, diagnoses were reported to the National Hospital Morbidity Database for all states and territories using the third edition of ICD-10-AM.
• For 1997-98 to 1999-00, the diagnosis related groups (DRGs) were reported by states and territories in a number of different versions of the Australian Refined Diagnosis Related Groups (AR-DRG) classification. For these years, the Institute regrouped the diagnosis related groups (DRGs) for all records in the database using AR-DRG version 4.0. Similarly, for 2000-01 to 2002-03, the Institute regrouped all records in the database using AR-DRG version 4.2. Following the introduction of AR-DRG version 5.0, the Institute retrospectively regrouped all records for the years 1998-99 to 2002-03 using AR-DRG version 5.0.

---

Scope of the database

The National Hospital Morbidity Database includes data from public acute and (until 1997-98) Department of Veterans' Affairs hospitals, public psychiatric hospitals, private acute and psychiatric hospitals, and private free standing day hospital facilities.

Exceptions within the public sector are public hospitals not within the jurisdiction of a State or Territory health authority or the Department of Veterans' Affairs (that is, hospitals operated by the Department of Defence, for example, and hospitals located in off-shore territories). In addition data are not available for some years for a few small public hospitals in some jurisdictions (detailed below).

In the private sector data are not available for some years for a few small private hospitals and free-standing day hospital facilities in some jurisdictions (detailed below).

A different methodology for collecting information on the scope was used from 1998-99 compared with earlier years. Therefore, the information below for 2002-03, 2001-02, 2000-01, 1999-00 and 1998-99 is likely to be more accurate than that provided for the earlier years, especially for private hospitals.

2002-03

Public sector hospitals that were not included were a mothercraft hospital in the Australian Capital Territory, a small rural hospital in New South Wales and a dental hospital in Victoria. 

Private sector hospitals that were not included were 3 free-standing day hospital facilities and 3 other private hospitals in Victoria and some other private hospitals in Victoria were unable to submit complete data. Data were also not included for all private free-standing day hospital facilities in the Australian Capital Territory, and the private free-standing hospital in the Northern Territory. For South Australia, data were not available for one private free-standing day hospital facility for one month. For Tasmania, data were not available for one small free-standing day hospital facility.

2001-02

Public sector hospitals that were not included were a mothercraft hospital in the Australian Capital Territory.

Private sector hospitals that were not included were 5 free-standing day hospital facilities and 3 other private hospitals in Victoria (some other private hospitals in Victoria were unable to submit complete data), all private free-standing day hospital facilities in the Australian Capital Territory, and the private hospital and the private free-standing hospital in the Northern Territory. For South Australia, data were not available for one private free-standing day hospital facility and were missing for January 2002 for another. Data were also missing for February to June 2002 for one private hospital (non-day only) and for January 2002 for another private hospital in South Australia.

2000-01

Public sector hospitals that were not included were one small 'outpatient clinic' in Queensland, a small rural hospital and a forensic hospital in Tasmania, and a mothercraft hospital in the Australian Capital Territory. Private sector hospitals that were not included were 11 free-standing day hospital facilities in Victoria, all private free-standing day hospital facilities in the Australian Capital Territory, and the one private hospital in the Northern Territory. For South Australia, data were not available for one private free-standing day hospital facility and were missing for January to June 2001 for another, and for May to June 2001 for one private hospital (non-day only). Data have only been provided for the periods from August 2000 to June 2001, January 2001 to June 2001 and April 2001 to June 2001 respectively for three other South Australian private free-standing day hospital facilities.

1999-00

Public sector hospitals that were not included were one small 'outpatient clinic' in Queensland, a forensic hospital in Tasmania, and a mothercraft hospital in the Australian Capital Territory.

Private sector hospitals that were not included were 17 free-standing day hospital facilities and one other private hospital in Victoria, all private free-standing day hospital facilities in the Australian Capital Territory, and the one private hospital in the Northern Territory. For South Australia, data were not available for three private free-standing day hospital facilities (one of which commenced operation in September 1999) and data were missing for March to June 2000, for May to June 2000 and for June 2000, respectively, for three others. For Tasmania, data were not available for one private free-standing day hospital facility and one other private hospital, and were missing for April to June 2000, December 1999 to June 2000, January and February 2000 and June 2000, respectively, for four other private hospitals.

1998-99

Public sector hospitals that were not included were one small outpatient clinic' in Queensland, a forensic hospital in Tasmania, and a mothercraft hospital in the Australian Capital Territory.

Private sector hospitals that were not included were 12 private free-standing day hospital facilities and one other private hospital in Victoria, three private free-standing day hospital facilities in South Australia, one private free-standing day hospital facility and four other private hospitals in Tasmania, six private free-standing day hospital facilities and one private hospital in the Australian Capital Territory, and the one private hospital in the Northern Territory. In addition, about 5.6% of private hospital separations data for Western Australia were not included (mainly for hospitals other than free-standing day hospital facilities).

1997-98

In 1997-98, public hospital data were not available for a mothercraft hospital in the Australian Capital Territory, one small outpatient clinic' in Queensland and most separations for three small district public hospitals in Tasmania.

In the private sector, about 4,500 hospital separations were not included for NSW private hospitals, and separations were not available for two private free-standing day hospital facilities and one other private hospital in Tasmania, private free-standing day hospital facilities in the Australian Capital Territory and the private hospital in the Northern Territory.

1996-97

In 1996-97 public hospital data were incomplete for three small district public hospitals in Tasmania, and were not available for some lodges attached to public hospitals in Western Australia and a mothercraft hospital in the Australian Capital Territory.

In the private sector, separations were not available for private free-standing day hospital facilities in the Australian Capital Territory and the private hospital in the Northern Territory.

1995-96

In 1995-96 public hospital data were incomplete for three small district public hospitals in Tasmania, and were not available for some lodges attached to public hospitals in Western Australia, a mothercraft hospital in the Australian Capital Territory and for public psychiatric hospitals in Queensland.

In the private sector, separations were not available for private free-standing day hospital facilities in the Australian Capital Territory and the private hospital in the Northern Territory. In addition, data for one of the private hospitals in the Australian Capital Territory were only supplied for the final 9 months of the year.

1994-95

In 1994-95 public hospital data were incomplete for three small district public hospitals in Tasmania, and were not available for some lodges attached to public hospitals in Western Australia, a mothercraft hospital in the Australian Capital Territory and for public psychiatric hospitals in all jurisdictions.

In the private sector, separations were not available for private free-standing day hospital facilities in the Australian Capital Territory and the private hospital in the Northern Territory. In addition, about 2% of separations from Victorian private hospitals were not included.

1993-94

In 1993-94 public hospital data were not available for some lodges attached to public hospitals in Western Australia, a mothercraft hospital in the Australian Capital Territory and for public psychiatric hospitals in all jurisdictions.

In the private sector, separations were not available for private free-standing day hospital facilities in the Australian Capital Territory and the private hospital in the Northern Territory. In addition, about 19% of separations from Victorian private hospitals were not included.