Comparing life expectancy of Indigenous people in Australia, New Zealand, Canada and the United States
Numerous studies have indicated that the gap between Indigenous and non Indigenous people’s life expectancy in Australia is greater than in New Zealand, Canada and the United States of America (USA).
However, difficulties related to concepts, data and methods behind such estimates throw doubt on conclusions drawn from country comparison studies. This paper suggests that the uncertainty associated with indigenous life expectancy estimates could be quite large.
How do different countries count their indigenous population?
Australia, New Zealand, Canada, and the USA differ in how they define and measure their Indigenous population. The two main differences relate to the 'inclusiveness' of the underlying Indigenous population and the accuracy of Indigenous deaths recorded.
- Australian estimates of the Indigenous population are based on self-identification of Indigenous status from the Census. While Indigenous status is reported in the national mortality database, there are known problems with non-reporting and incorrect reporting of Indigenous deaths.
- In New Zealand, Census respondents are asked what 'ethnic group' they belong to and multiple answers are permitted. The Maori life tables from 1990–1992 include all who identify as Maori, including those who identify with other ethnicities. Since 1995, Maori deaths have been identified in a similar way.
- Life expectancy estimates in the USA are restricted to American Indians and Alaska Natives (AI/ANs) who live 'on or near reservations'. Based on 2002 Census data, this was around 57% of the estimated AI/AN population. AI/AN identification is available on official death certificates, however, unlike the Census, multiple race information was not available before 2003. As of 2007, only 27 of the 52 registration areas collected multiple race information.
- Canadian estimates of the Aboriginal population are obtained from a 20% sample of the five-yearly population Census. Aboriginal status is not collected on the death registration form. Data linkages are used to ascertain Aboriginal deaths. Published life expectancy estimates are restricted to Registered Indians, who account for around one-half of the Aboriginal population.
Countries that adopted the least inclusive definitions of 'Indigenous' in their health and population data collection systems, tended to report the best life expectancy estimates.
How large is the uncertainty associated with Indigenous life expectancy estimates?
In general, cross-country comparisons give little or no consideration to the level of uncertainty associated with the data and methodologies used.
- Most publications present and discuss differences between country estimates of Indigenous life expectancy as though they represent real differences.
- Only a few publications provide general cautions about drawing conclusions in the presence of uncertainties of largely unknown magnitude.
Based on an informed evaluation, this paper suggests that the uncertainty associated with Indigenous life expectancy estimates could be quite large.
How could Indigenous life expectancy estimates be improved?
Improving the reliability of comparisons would involve all countries moving towards an 'inclusive' definition, and developing robust methods to test and adjust the population and death data for inadequate reporting.
- In New Zealand, Australia and the USA, only incremental change would be required: existing census and death data systems could be strengthened, and new validation and adjustment methods developed.
- In Canada, major system changes would be required to obtain national data on the population and deaths of all the Indigenous subgroups.
More information
Comparing life expectancy of Indigenous people in Australia, New Zealand, Canada and the United States (748KB PDF) contains more detailed information.