Data sources
National Integrated Health Services Information Analysis Asset
The National Integrated Health Services Information (NIHSI) analysis asset is a person-focused, de-identified analytical asset that allows for a range of research and analysis on many aspects of population health through access to standardised, integrated health administrative datasets. The NIHSI is the only enduring linked data asset that includes linked hospital data including admitted patient care services, emergency department services and outpatient services in public hospitals for NSW, Vic, Qld, SA, Tas and the ACT. The NIHSI includes various national Commonwealth government datasets, including the Pharmaceutical Benefits Scheme, Repatriation Pharmaceutical Benefits Scheme data, Medicare Benefits Schedule data, Residential Aged Care Services data, and the National Death Index data. The NIHSI version 1.0 analysed for this report contains data from 2010 until 2018–19.
Data linkage for the NIHSI was undertaken using probabilistic linkage. This linkage procedure involves creating record pairs by combining records from one data set with records from another data set based on similarities in characteristics such as first and last name(s); day, month and year of birth; and sex. The Medicare Consumer Directory and the National Death Index were first linked to create the AIHW linkage spine which was used to create the NIHSI.
Pharmaceutical Benefits Scheme
The Pharmaceutical Benefits Scheme (PBS) contains claims for Australian Government subsidised medicines. PBS item claims are mapped to the Anatomical Therapeutic Classification (ATC) index, a classification system that is recommended by the World Health Organization as the international standard for presenting and comparing drug usage data. The ATC groups medicines according to the body organ or system on which they act, as well as their therapeutic and chemical characteristics.
The version of the PBS included in NIHSI 2018–19 includes Repatriation Pharmaceutical Benefits Scheme data, covers all jurisdictions, and spans the period 2010–11 to 2018–19.
This analysis uses PBS data from June 2010 to December 2017 to identify people who were dispensed dementia-specific medication (Table 1).
Admitted Patient Care
The Admitted Patient Care (APC) is part of the National Hospital Morbidity Database. APC data contains episode-level records from admitted patient morbidity data collection systems in Australian hospitals. The version of the APC included in NIHSI 2018–19 includes public hospital data for 6 states/territories: NSW, Vic, Qld, SA, Tas and the ACT for the years 2010–11 to 2018–19, and select private hospital data for: Vic (2010–11 to 2016–17), Qld (2010–11 to 2018–19) and the ACT (2010–11 to 2018–19).
It is important to note that the version of the APC included in NIHSI 2018–19 does not include the supplementary codes for chronic conditions. Supplementary codes for chronic conditions provide information about chronic conditions where it may not be appropriate to include this information as an additional diagnosis. They were introduced into the APC data in 2015 and have been found to improve the identification of people living with dementia (AIHW 2023).
Diagnostic information in APC data can be recorded as a principal diagnosis or additional diagnosis. These are coded using the International Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM).
This analysis uses APC data from June 2010 to December 2017 to identify people who were hospitalised with a principal or additional diagnosis of dementia (Table 2), and data from June 2016 to December 2018 to understand people’s hospital use and the characteristics of their first hospitalisation.
National Non-Admitted Patient Emergency Department Care
The National Non-Admitted Patient Emergency Department Care Database (ED) is part of the National Hospital Morbidity Database. The ED data contains information on non-admitted emergency department activity in public hospitals and is available in NIHSI 2018–19 for all jurisdictions except WA and NT for the years 2010–11 to 2018–19, although diagnostic information is only available for 2013–14 to 2018–19.
Diagnostic information for emergency department presentations can be recorded in the principal or additional diagnosis. Diagnostic information in the ED data is not consistently recorded and is reported using several classification systems. The following were used from 2013–14 to 2018–19:
- Systematised Nomenclature of Medicine—Clinical Terms—Australian version, Emergency Department Reference Set (SNOMED CT-AU EDRS)
- International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM).
- ICD-10-AM.
This analysis uses ED data from June 2013 to December 2017 to identify people who were had an ED presentation with a principal or additional diagnosis of dementia (Table 3). Data from June 2016 to December 2018 is used to combine discrete hospital episodes into hospital stays and to understand people’s hospital use during their first hospitalisation and 12-months afterwards.
Aged Care Funding Instrument
The Aged Care Funding Instrument (ACFI) is a resource allocation tool that was used in government-funded residential aged care facilities from March 2008 to September 2022 to assess the care needs of each resident and determine the funding of each facility. After admission to residential aged care, the ACFI was used to categorise residents core care needs in each funding domain and included two sections for previously diagnosed health conditions that most affect the person’s care needs. Information on mental or behavioural conditions was used to identify whether people had dementia in this analysis (Table 4).
The version of the ACFI included in NIHSI 2018–19 includes data for all jurisdictions for the years 2010–11 to 2018–19. This analysis uses ACFI data from June 2010 to December 2017 to identify people had dementia listed as a mental or behavioural condition.
National Death Index
The NDI contains records of all deaths occurring in Australia since 1980. The underlying cause of death and all other associated causes of death are recorded. These are coded using the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10).
The version of the NDI in NIHSI 2018–19 contains data on fact of death for all jurisdictions from July 2010 to June 2019, and information on cause of death from July 2010 to late 2018 (with complete cause of death information from July 2010 to June 2017).
This analysis uses NDI data from 2017 to identify people who had dementia listed as an underlying or associated cause of death (Table 5) and NDI data from 2017 and 2018 to report whether people died after their hospitalisation.
Residential Aged Care activity
The Residential Aged Care activity data (RAC activity data) contains episode-level records from all government-subsidised residential aged care services, including permanent residential aged care and respite residential aged care. The version of the RAC activity data included in NIHSI 2018–19 includes data for all jurisdictions for the years 2010–11 to 2018–19. This analysis uses RAC activity data from 2016 to 2018 to understand whether people lived in residential aged care as a permanent resident or accessed residential respite care before and after their hospitalisation.
Medicare Benefits Scheme claims data
The Medicare Benefits Scheme claims data (MBS) contains MBS claims data for Medicare services subsidised by the Australian Government. The version of the MBS data included in NIHSI 2018–19 includes data for all jurisdictions for the years 2010–11 to 2018–19. This analysis uses MBS data from 2017 and 2018 to understand people’s use of Medicare-subsidised healthcare services after their hospitalisation.
Demographic data
Information on a person’s sex and age in years is based on information held in the MBS or PBS. Where date of birth or sex information is not available from the MBS or PBS data, it is obtained from the RAC, NDI or APC data.
People who did not have age or sex information in the linked data were excluded from analysis.