Hospitalisations for mental health conditions and intentional self-harm in 2015–16
This technical note accompanies the Web update: Hospitalisations for mental health conditions and international self-harm in 2015–16.
The web update presents local rates of hospitalisations for the following mental health conditions and intentional self-harm:
- Anxiety and stress disorders
- Bipolar and mood disorders
- Depressive episodes
- Drug and alcohol use
- Schizophrenia and delusional disorders
Results are presented for the 31 Primary Health Network (PHN) areas that cover Australia, and for more than 300 smaller local areas called Statistical Areas Level 3 (SA3s), as defined by the Australian Bureau of Statistics (ABS) in 2011.
About the data source
Data for the web update were sourced from the AIHW’s National Hospital Morbidity Database (NHMD). The NHMD is a compilation of separation records from admitted patient morbidity data collections in Australian hospitals. The NHMD includes all episodes of care for admitted patients, including admissions for day only care, in nearly all public and private hospitals. The NHMD does not include separations of non-admitted patient care provided in outpatient clinics or emergency departments. Separations for unqualified newborn care, and records of posthumous organ procurement or hospital boarders are excluded.
In the web update, the counting unit is a ‘hospitalisation’ (or separation). This may be a complete hospital stay (to discharge, transfer, or death), or a part of the stay if there was a change of care type (for example from acute care to rehabilitation). As a record is included for each hospitalisation, not for each patient, patients hospitalised more than once or transferred between hospitals in the financial year will have more than one record.
A separation is classified as having specialised psychiatric care (specialised care) if the patient was reported as having spent one or more days in a specialised psychiatric unit or ward. Specialised bed days are the number of days a patient is admitted to a specialised psychiatric unit or ward.
A separation was flagged as being admitted to a private hospital if the establishment type was ‘private’ or ‘private free-standing day hospital facility’ (i.e. where esttype equals 2 or 5). Private hospital refers to the establishment not the patient’s insurance type.
Comparability over time
Caution should be used when comparing hospitalisations for mental health conditions and intentional self-harm over time, due to changes in coding practices and/or geographic concordances as well as admission practices which are outlined below. Note, for comparability purposes, the current release contains revised data for 2013–14 and 2014–15 accounting for these changes where relevant.
Information presented over time may be affected by changes to codes and coding standards defined in the International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM)1. The major changes affecting the interpretation of information presented in the web update are outlined below.
Data reported in previous Hospitalisations of mental health conditions and intentional self-harm in 2013–142 (released September 2016) and Web update: Hospitalisations of mental health conditions and intentional self-harm in 2014–153 (released February 2017) were calculated by applying a geographic concordance from postcode. The method has been updated to use the Statistical Area Level 2 (SA2) data item supplied within the NHMD. For comparability purposes, the current release contains data which were calculated by applying geographic concordance from SA2 for 2013–14 and 2014–15, which had a minor effect on previously reported rates.
New reporting of principal diagnoses for rehabilitation care hospitalisations
A change to the coding standard was made from 1 July 2015 to record the underlying condition requiring rehabilitation as the principal diagnosis, rather than the code Z50 - Care involving the use of rehabilitation procedures. The AIHW examined data provided for 2015–16. Overall numbers of mental health hospitalisations with principal diagnosis codes Z50.2 or Z50.3 decreased between 2014–15 and 2015–16, and a decrease in the number of mental health hospitalisations with care type Rehabilitation was also observed. These changes were likely due to the implementation of care type Mental Health (see below).
Implementation of the Mental health care type. The Mental health care type was introduced from 1 July 2015 to record separations in which the primary clinical purpose or treatment goal is improvement in the symptoms and/or psychosocial, environmental and physical functioning related to a patient’s mental disorder. Before 1 July 2015, records for which the current Mental health care type definition would have applied were assigned to another care type (for example, Acute, Rehabilitation, Psychogeriatric care or Geriatric evaluation and management). For more information about these issues, see the Data Quality Statement Summary, pages 238–263, Admitted patient care 2015–16: Australian hospital statistics.4
About the measures
The four measures in the release are described below. The metadata specifications for each measure are available in the Indicator specification section.
Age-standardised rates of mental health overnight hospitalisations
A mental health hospitalisation is defined according to the AIHW Admitted patient mental health-related care developed in consultation with the National Mental Health Working Group Information Strategy Committee (now called the Mental Health Information Strategy Standing Committee) and the Clinical Casemix Committee of Australia.
A mental health hospitalisation was classified as a separation if a principal diagnosis is listed in the section on ‘Mental and behavioural disorders’ in the International and Statistical Classification of Diseases and Related Health Problems, 10th revision, Australian Modification (ICD-10-AM)1 classification (codes F00-F99).
For more information on the rationale for the inclusion criteria of relevant mental health codes, see: https://www.aihw.gov.au/reports/mental-health-services/mental-health-services-in-australia/classifications-and-technical-notesExternal link, opens in a new window..
To improve comparability across jurisdictions, this measure is reported for overnight hospitalisations only. This is because same-day separations can include treatments requiring multiple, planned same-day admissions over a period of time.
In addition, same-day separations had the highest transfer rate. Transfers result in duplication of mental health hospitalisations. Across local areas, the percentage of mental health hospitalisations that results in a transfer to another hospital varied widely.
Age-standardised rates of bed days for mental health overnight hospitalisations
This measure reflects the number of hospital bed days used by patients admitted overnight for a mental health condition compared to the population in the patients’ area of residence.
Age-standardised rates of intentional self-harm hospitalisations
This measure reflects the number of hospital admissions (same-day and overnight) for a condition or injury resulting from intentional self-harm compared to the population in the patients’ area of residence. Note that this measure does not attempt to account for patient transfers between hospitals. When a patient is transferred to another hospital, another admission is recorded for the same separation of self-harm.
Hospitalisations for a condition or injury resulting from intentional self-harm, may reflect a range of causes including purposely self-inflicting poisoning or injury and attempted suicide. All hospitalisations where any ICD-10-AM external cause code of X60–X84 or Y87.0 is reported are included.
The ICD-10-AM external cause codes provide an indication of whether a poisoning or injury was accidental, self-inflicted, inflicted by another person, or undetermined.
The reliability of hospital records in reflecting the level of intentional self-harm in hospitalised patients is untested and it is not known to what extent record of hospitalisation reflects self-harm in the community. Most people who contact health services after a separation of intentional self-harm are seen by emergency departments. They may or may not be admitted as hospital inpatients, and the injury may or may not be recorded as intentional.
Age-standardised rates of bed days for intentional self-harm hospitalisations
This measure reflects the number of hospital bed days for a condition or injury resulting from intentional self-harm compared to the population in the patients’ area of residence.
About the methods
- Primary Health Network (PHN) areas – 31 geographic areas covering Australia, with boundaries defined by the Australian Government Department of Health5
- Statistical Areas Level 3 (SA3s) – 333 geographic areas covering Australia, with boundaries defined by the ABS.6
Using the Statistical Area Level 2 (SA2) of usual residence in the NHMD unit record data, rather than the location of the hospital where each patient was admitted, data by geographic area were calculated by applying a geographic concordance from SA2 to PHN area and SA3 area. Where an SA2 boundary overlapped a PHN area, records were attributed to a PHN area based on the percentage of the population within that SA2 that fell within each PHN area. Figures were rounded at the end of the calculations.
In tables presenting measures by geographic area, individual area results may not add to national totals due to missing location data, and also due to rounding.
The web update includes measures expressed as age-standardised rates per 10,000 population. Age-standardised rates are hypothetical rates that would have been observed if the populations studied had the same age distribution as the standard population. This facilitates comparisons between populations with different age structures. This adjustment is important because the rates of many health conditions vary with age.
The direct method of age-standardisation was applied to the data.7 Age-standardised rates were derived by calculating crude rates by five year age groupings of 0–4 years to 85+ years. If a patient’s age was recorded as over 116, they were excluded. These crude rates were then given a weight that reflected the age composition of the standard population. The standard population used is the ABS Estimated Resident Population (ERP) for Australia as at 30 June 2001.
Limiting the impact of long-stay patients on bed days
In one area there was a policy change to move patients with a mental health condition out of psychiatric facilities and into the community. This resulted in the discharge of a group of patients who had been resident in a hospital or psychiatric facility for many years. The bed day rate in the data included care for up to 20 years prior to 2013–14 and resulted in an extremely high rate for the measure of bed days in this one area. Sensitivity analyses showed the 2013–14 bed day rate was inconsistent with previous years.
For this reason, an upper bound of one year (365 days) was applied to the number of bed days in each separation for results in 2013–14, 2014–15 and 2015–16. This approach to data analysis is known as winsorisation and it reduces the impact of an unusual event such as the policy.
Suppression of results
All data for an area were suppressed (marked ‘NP’ – Not available for publication) if the number of rounded mental health condition or intentional self-harm hospitalisations ranged from 1 to 4.
Age-standardised rates were suppressed if any of the following suppression conditions were met:
- the number of rounded mental health condition or intentional self-harm hospitalisations (numerator) ranged from 1 to 19, or
- the population of an area (denominator) was less than 2,500.
- the population of an area (denominator) was less than 30 in any of the standard 5-year age groupings used to calculate the rate, then the rate was marked ‘Interpret with caution’ (i.e. ‘#’) as these rates are considered potentially more volatile than other published rates. For each of these flagged rates, the effect of an increase of one hospitalisation on the rank of the area was examined. If the rank changed so much that the area was on the cusp of changing two deciles, then the rate was suppressed (marked ‘NP’ – Not available for publication).
The percentages and bed days for specialised and private hospitalisations were suppressed (marked ‘NP’ – Not available for publication) if the number of rounded separations in the geographic area ranged from 1 to 19.
1. ACCD (Australian Consortium for Classification Development) 2015. The International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM)–9th edn.–tabular list of diseases, and Alphabetic index of diseases. Sydney: Independent Hospital Pricing Authority.
2. Australian Institute of Health and Welfare (AIHW) 2016. Healthy Communities: Hospitalisations for mental health conditions and intentional self-harm in 2013–14. Cat. No. HSE 177. Canberra: AIHW.
3. AIHW 2017. Healthy Communities – Web update: Hospitalisations for mental health conditions and international self-harm in 2014–15. Canberra: AIHW. Viewed 31 August 2017. http://myhealthycommunities.gov.au/our-reports/mental-health-and-intentional-selfharm/february-2017/web-update
4. AIHW 2017. Admitted patient care 2015–16: Australian hospital statistics. Health services series no.75. Cat. no. HSE 185. Canberra: AIHW.
5. ABS (Australian Bureau of Statistics) 2011. Australian Statistical Geography Standard (ASGS): Volume 1–Main structure and greater capital city statistical areas, ABS cat. no. 1270.0.55.001. Canberra: ABS.
6. Department of Health 2016. Primary Health Networks (PHNs). Canberra: Department of Health. Viewed 31 August 2017, http://www.health.gov.au/internet/main/publishing.nsf/Content/PHN-HomeExternal link, opens in a new window.
7. AIHW 2005. Age-standardised rate. Canberra: AIHW. Viewed 8 March 2017, http://meteor.aihw.gov.au/content/index.phtml/itemId/327276External link, opens in a new window.