In 2005, the Australian Institute of Health and Welfare released a report on the quality of Indigenous identification in hospital separations data (AIHW 2005a). It advised restricting analyses of hospital separations data for Indigenous persons to states and territories with a level of Indigenous identification that was adequate for analysis purposes (Queensland, South Australia, Western Australia, and the Northern Territory (public hospitals only)). It also advised against performing analyses of hospital separations data by remoteness for Indigenous persons, as variation in identification levels by remoteness could have biased the results.
This report presents the results of studies conducted between 2006 and 2008 by the AIHW, in collaboration with the state and territory health authorities, in a follow-up assessment of the quality of Indigenous identification in public hospitals. It presents revised recommendations for analysis of hospital separations data based on the results of the studies, and estimates of correction factors that can be applied to the data for analysis purposes.
An estimated 89% of Indigenous patients were correctly identified in Australian public hospital admission records in 2007-08. In other words, 11% of Indigenous patients were not identified, and the 'true' number of hospital admissions for Indigenous persons was about 12% higher than reported.
States and territories
While there is still scope for improvement in the identification of Indigenous persons in hospital separations data, the results of the studies support expanding national reporting to include data for New South Wales, Victoria, Queensland, South Australia, Western Australia, and the Northern Territory (public hospitals only). Levels of Indigenous identification were 80% or higher for those jurisdictions.
For Tasmania and the Australian Capital Territory, the levels of Indigenous identification were not considered acceptable for analysis purposes.
Since the 2005 report, the levels of Indigenous identification for public hospitals:
- increased in New South Wales and Victoria to 88% and 84% respectively, and were considered to have been at 'acceptable' levels from the 2004-05 data year
- increased for Queensland, Western Australia, and the Northern Territory
- decreased (but were still regarded as acceptable) for South Australia.
The studies show that there were acceptable levels of Indigenous identification for all remoteness areas, ranging from 80% in Major cities to 97% in Remote and Very remote areas.
Therefore the quality of the data supports analyses by remoteness areas, in aggregate, across states and territories for which the levels of identification were considered to be acceptable.
The sample size was insufficient to allow assessment of the quality of Indigenous identification by remoteness area within jurisdictions.