  |
|
|
||||
|
||||||
| Journal of the New Zealand Medical Association,
17-December-2004, Vol 117 No 1207
|
||||||
|
Indigenous disparities in disease-specific mortality, a
cross-country comparison: New Zealand, Australia, Canada, and the United
States
Dale Bramley, Paul Hebert, Rod Jackson, Mark
Chassin
BackgroundDisparities in health status for the
indigenous peoples of New Zealand, Australia, Canada, and the United States (US)
have been well documented.1–6 In each of
these ‘rich’ countries, the indigenous peoples invariably suffer
from poorer health, with an excess of early mortality and lower life expectancy
when compared to the non-indigenous population.
Numerically, the indigenous populations of each country
represent a small proportion of the total population. Maori represent
approximately 15% of the New Zealand population, Aboriginals and Torres Strait
Islanders represent 2–3% of the Australian population, American Indians
and Alaskan Natives represent 1–1.5% of the US population, and Aboriginal
Canadians represent 4% of the Canadian
population.7–10
Although gains in health status have been made for all the
indigenous peoples of these four countries, large disparities remain. In New
Zealand, a recent report has highlighted that although life expectancy has
improved dramatically for non-Maori non-Pacific people, Maori life expectancy
has remained largely static, leading to a relative increase in the life
expectancy disparity experienced by Maori.11 In
particular, Maori disease specific mortality disparities have increased for
cardiovascular disease and cancer when compared to
non-Maori.11
The aim of the research is to compare the disease specific
mortality rates of the indigenous population of New Zealand, Australia, Canada,
and the US with the non-indigenous population in each country. The New Zealand
population is the reference population for which all comparisons are made. The
size of the relative disparities (indigenous/non-indigenous) in disease-specific
mortality rates, within and between countries are compared.
MethodLife expectancy at birth data
for New Zealand (2000-2002) was obtained from Statistics New
Zealand.12 Life expectancy data for the US
(2001) was obtained from the Centers for Disease Control and Prevention
(CDC).13 Life expectancy data for Australia
(2000) was obtained from the Australian Institute of Health and
Welfare.14 Life expectancy data for Canada
(2000) was obtained from Health Canada.4
Mortality risk ratios for comparison are those accounting for the leading causes
of death in New Zealand and the US in
1999.15,16 These included malignant neoplasms
of the lung, bowel, cervix, female breast, and prostate; ischaemic heart
disease; cerebrovascular disease; chronic obstructive pulmonary disease (COPD);
intentional self harm; diabetes; human immunodeficiency virus (HIV); assault;
pneumonia and influenza.
Mortality data for New Zealand are compiled by the New
Zealand Health Information Service (NZHIS).15
Cause of death in 1999 was defined by International Disease
Classification–9th edition (ICD–9) codes (Appendix 1). Crude
mortality data (1999) for this study was obtained from the NZHIS. The national
mortality dataset in New Zealand contains ethnicity information. Since 1996, the
ethnicity question recorded on death registration certificates has been the same
as that asked in the 1996 national census of population and
dwellings.15 Mortality data for the US are
compiled by the National Center for Health Statistics of the US
CDC.17 Cause of death in 1999 was defined by
ICD-10 codes (Appendix 1). Crude mortality data for this study was obtained from
the CDC. The national mortality dataset in the US contains race
information.18
Mortality data for Australia is complied by the
Australian Bureau of Statistics.19 Currently,
there is incomplete coverage of indigenous deaths in some state and territory
registration systems in Australia. Therefore, the mortality data used for this
study was from those jurisdictions assessed by the Australian Bureau of
Statistics as having a sufficient level of coverage to enable statistics on
Aboriginal and Torres Strait Islanders to be produced. These states and
jurisdictions include Queensland, South Australia, Western Australia, and the
Northern Territory.8 The Australian Institute
of Health and Welfare supplied crude mortality data (1999) for this
study.20 Cause of death in 1999 was defined by
ICD–10 codes. The Australian population denominator values used in this
study were derived from the 2001
census.20
Statistics Canada supplied mortality data for Canada,
for the population group ‘all Canadians’. The national mortality
dataset held by Statistics Canada does not contain ethnicity data. At present
there is no mortality data available for off-reserve indigenous Canadians.
Indigenous mortality data (1999) was only available for First Nation on-reserve
indigenous Canadians.21 Crude mortality data
for First Nation on-reserve indigenous Canadians was obtained from Health Canada
(First Nations and Inuit Health Branch).22
Cause of death in 1999 was defined by ICD–9 codes. For New Zealand,
Australia, Canada, and the US, we complied and calculated from crude data ethnic
specific mortality rates by primary cause of death in 1999 for the indigenous
and non-indigenous populations. We calculated age-adjusted mortality rates,
using direct standardization23 and weights
based on the WHO world standard population.24
We also used New Zealand, Australian, Canadian, and US-based weights; and Segi
standard population-based weights15—and
found results similar to those presented here.
ResultsLife expectancy overall for males
(76.6 years) and females (82.1 years) was highest in Australia (see Table 1).
Male indigenous life expectancy was highest in New Zealand (69.0 years) and
female indigenous life expectancy was highest in Canada (76.6 years). The lowest
life expectancy for indigenous peoples for both males (56 years) and females (63
years) was in Australia. Australian Aboriginals and Torres Strait Islanders,
therefore, experienced the greatest disparity in life expectancy, when compared
to the non-indigenous population.
Maori had the highest mortality rates among all population
groups (see Table 2), for ischaemic heart disease, COPD, total malignant
neoplasms and malignant neoplasm of the lung, female breast, prostate, and
cervix. Non-Maori New Zealanders had the highest mortality rate for malignant
neoplasm of the bowel among all population groups. The only three
disease-specific mortality rates measured where Maori mortality was lower than
non-Maori mortality occurred in malignant neoplasm of the bowel, pneumonia and
influenza, and intentional self-harm.
Australian Aboriginals and Torres Strait Islanders had the
highest mortality rates among all population groups for cerebrovascular disease
and diabetes. When indigenous mortality rates were compared with non-indigenous
mortality rates in Australia, Aboriginal and Torres Strait Islander mortality
rates were higher for every disease-specific mortality rate measured, except for
malignant neoplasm of the bowel.
Canadian First Nation peoples had the highest mortality rate
among all population groups for intentional self-harm and, pneumonia and
influenza. Indigenous mortality rates were lower than non-indigenous mortality
rates in Canada for total malignant neoplasms, malignant neoplasm of the lung
and female breast, ischaemic heart disease, cerebrovascular disease, and
COPD.
American Indians and Alaskan Natives had the highest
mortality rate among all the population groups for assault. Indigenous mortality
rates were lower than non-indigenous mortality rates in the US for total
malignant neoplasms and each of the individual neoplasms reported (lung, bowel,
female breast, cervix, and prostate), ischaemic heart disease, cerebrovascular
disease, HIV and COPD.
In terms of the size of the relative disparities that exist
between population groups within a country, New Zealand Maori and Australian
Aboriginals and Torres Strait Islanders experienced the highest levels of
disparities when compared to their respective non-indigenous population groups
(Figures 1–3).
The size of mortality risk ratio for
indigenous/non-indigenous populations groups (Table 2) across all four countries
was highest in New Zealand for total malignant neoplasms (risk ratio [RR] 1.6)
of the lung (RR 2.9), breast (RR 1.5), cervix (4.5), and prostate (RR1.5); HIV
(2.0) and ischaemic heart disease (RR of 1.9, which was the same as the
Australian indigenous/non-indigenous RR). In Australia, the size of mortality
risk ratio for indigenous/non-indigenous populations groups was the highest for
all four countries for: cerebrovascular disease (RR 2.1), COPD (RR 2.5),
pneumonia and influenza (RR 2.1), diabetes (RR 9.8), ischaemic heart disease (RR
1.9), and assault (RR 5.6).
Appendix 1. International Classification of Disease
(ICD) codes used for defining major causes of death
Across all four countries (New Zealand, Australia, Canada,
and the US) the indigenous peoples had higher mortality rates for diabetes and
assault when compared to their non-indigenous populations.
DiscussionThis paper compares indigenous and
non-indigenous disease-specific mortality rates and risk ratios (in New Zealand,
Australia, Canada, and the US) for the leading causes of death in New Zealand.
There have been a limited number of academic papers published comparing
indigenous disparities in mortality among rich
countries.25–28 In 1992, Hogg attempted
to place Australian Aboriginal mortality within the broader context of other
countries. Hogg found that Australian Aboriginals had higher age and
cause-specific death rates, and a strikingly different mortality profile overall
compared to indigenous peoples in New Zealand, the US, and
Canada.25
The main findings of this research are that:
Life expectancy in all four countries was lower for the
indigenous peoples—with Australian Aboriginals and Torres Strait Islanders
having the lowest life expectancy of all population groups and the greatest
relative disparity when compared to the non-indigenous population.
The highest disease-specific mortality rates for ischaemic
heart disease and malignant neoplasms are found in New Zealand Maori (except for
malignant neoplasm of the bowel where New Zealand non-Maori have the highest
rate). Canadian First Nation peoples have the highest mortality rates of all
population groups for intentional self-harm and pneumonia/influenza. American
Indians and Alaskan Natives have the highest mortality rates of all population
groups for assault. Non-indigenous Americans have the highest mortality rate for
HIV.
In terms of the size of the relative disparities that exist
for disease-specific mortality, New Zealand Maori, Australian Aboriginals and
Torres Strait Islanders have the highest levels of disparities when compared to
their non-indigenous population groups.
Diabetes is a powerful determinant of health outcome and for
indigenous peoples across the four countries, diabetes related mortality is
high. Australian Aboriginals and Torres Strait Islanders, in particular, have
very high mortality rates associated with diabetes—and the
indigenous/non-indigenous risk ratio of 9.8 was the highest reported. The
prevalence of diagnosed diabetes has in recent years been increasing in all four
indigenous populations.29–32 Also, the
prevalence of obesity is increasing in some indigenous
populations,30,32 this will result in a rise in
diabetes related mortality in the near future.
The current high levels of indigenous mortality and
disparities that exist in New Zealand and Australia are not acceptable. In
comparison, indigenous mortality in Canada and the US is lower in many of the
disease-specific areas reported in this study when compared to their
non-indigenous counterparts.
There are several cross-country and country-specific lessons
that should be explored following these analyses. For example, cancer deaths in
indigenous Americans (and to a lesser extent, cancer deaths in indigenous
Canadians) are very low—these findings are consistent with other published
reports. Low indigenous mortality rates for lung cancer in the US may be
partially explained by the rarity of habitual cigarette smoking among Southwest
tribes but reasons for the low rates of other cancers are not so
evident.33 Cobb, in a recent report on
indigenous cancer deaths in the US, stated that further research is required to
elucidate why American Indians have low cancer mortality. This research may have
significant implications for cancer prevention in other ethnic
groups.33
Although disparities are large in New Zealand for death
from assault, the absolute rates are lower than in the US. For example, the
Maori age-standardised mortality rate from assault is 3.9 per 100,000 (RR of 3.9
compared to non-Maori) compared to the non-indigenous rate of 6.4 per 100,000 in
the US. Further research should be undertaken to explore how the national
response to violence differs between countries. A review of factors that have
been successful in keeping death related to assault comparatively low in New
Zealand may have implications for policy development in the US.
New Zealand has a low annual incidence of new HIV infections
and subsequent low mortality rates as reflected in the study findings (although
new infections have been increasing in recent years). The New Zealand response
to the HIV epidemic has been viewed as a public health success story. The New
Zealand response was characterised by law change (the Homosexual Law Reform Act
was passed), national coordination of a policy response (the National Council on
AIDS and a medical advisory committee were formed), and empowerment of affected
communities (groups such as the AIDS Foundations, Injecting Drug User Community
Groups, and the New Zealand Prostitutes Collective were
formed).34 Such a public health approach could
be undertaken to protect the health of indigenous and non-indigenous populations
in other countries.
The publication of comparative data such as this should
stimulate increased cross-country learning, research, and policy
development.
The quality of indigenous mortality dataThere are several common issues that
adversely affect the quality of indigenous mortality data. These include the
lack of an accurate denominator value for the indigenous population concerned
(mainly due to undercounting) and the lack of agreement over which population
denominator values to use if they do exist (e.g., whether to use single ethnic
response groups as the denominator value vs the multiple ethnic response
groups).
Denominator values for the indigenous population in all four
countries are usually derived from census data. However, in Australia,
estimating the size of the Aboriginal and Torres Strait Islander population has
proved difficult due to uncertainties attached to interpreting indigenous
population counts from the 5-yearly census.8
Between 1996 and 2001, the Australian indigenous population
increased 16 %, however the expected increase based on natural increase (births
minus deaths) was 12%.8 This variance is in
part due to the increased propensity of indigenous people to self-identify as
indigenous on the census forms. As it is not possible to estimate how these
factors may change over time, it is therefore problematic to estimate the
inter-census population denominator counts that are needed to calculate annual
death rates.
There is a lack of agreement as to how official agencies
define indigenous status and the way in which ethnic specific mortality data is
recorded. In New Zealand (as in other countries), there has been frequent
modification of the ethnicity question recorded in the censuses and it was not
until 1991 census that the biological concept of ethnic origin was replaced with
that of self-identified ethnicity.35 These
frequent changes in the census ethnicity question has led to difficulty
comparing mortality trends over time and have also produced difficulties in
estimating inter-census population denominator counts.
Perhaps the most important issue in regards to the quality
of indigenous mortality data is the undercounting of deaths (the numerator for
mortality data). In each of the four countries, the undercounting of indigenous
deaths is likely to lead to an underestimation of the relative size of
disparities that exist between indigenous and non-indigenous populations. In
Australia, for example, the Australian Bureau of Statistics (which administers
the national mortality database), recommends that the coverage of indigenous
mortality data is of sufficient quality to be used for research purposes only
from the jurisdictions of Queensland, South Australia, Western Australia, and
the Northern Territory.8 This is primarily due
to the fact that indigenous ethnicity status on death certificates is not always
recorded, or recorded incorrectly, leading to an undercounting of the number of
indigenous deaths.8
In New Zealand, research has been undertaken that attempts
to adjust for this undercounting by a process of probabilistic record linkage of
death registration data with census data. This research has produced estimates
of the considerable extent of the undercounting of Maori
deaths.36,37 Unfortunately, this data could not
be used for this study as there was no similar ‘corrected’ mortality
data available from the US, Australia, or Canada.
An issue that is unique to Canada is that the national
mortality database administered by Statistics Canada does not contain ethnicity
data. The regional offices of Health Canada collect mortality data for the
indigenous, on-reserve, First Nations population. Via a series of partnerships
with each provincial vital statistics registrar, First Nations specific death
certificate information is sent to the regional First Nations and Inuit Health
Branch regional office. However, in a number of areas no such relationships
exist (for example the Atlantic, Ontario, and Quebec regions), and therefore
data is obtained directly from the local communities, or not at
all.21 The availability of indigenous mortality
data in Canada is further limited by the lack of information that is available
for off-reserve, or non-status, indigenous peoples.
Methodological considerationsThe varying degrees of completeness
and accuracy of the indigenous mortality databases that exist within the four
countries are likely to affect these findings. For example research by Ajwani
(2003), has reported that during 1996-1999, 7% more decedents identified Maori
as one of their ethnic groups on the 1996 census compared with mortality
data.38 Therefore the accuracy of the Maori
deaths rates used in this study is relatively high. This level of accuracy is
unlikely to be present in the three other countries.
In the US, some estimates of the under-reporting of American
Indian deaths have ranged from 11% to 25%.39,40
In Canada, it is difficult to determine an accurate overview of indigenous
mortality for the reasons reported already, and due to the fact that only
limited information is available regarding indigenous people that reside in
urban areas.
The implication of these findings may therefore be that when
New Zealand indigenous/non-indigenous mortality risk ratios are compared with
indigenous/non-indigenous mortality risk ratios from these countries (Figures
1–3), the results may be somewhat improved to that described.
Although it is impossible to quantify the exact amount of
the measurement bias that may exist in our calculations, the data presented here
is the most reliable currently available. Differences in the calculation of life
expectancy and in ICD coding practices between countries could bias the
findings, but this is likely to have a minimal effect on the relative
differences in mortality between indigenous and non-indigenous populations
within a country, which is the main focus of this paper. It should also be noted
that grouping of data for indigenous peoples may obscure important differences
that may exist between large tribal grouping, an issue that may be particularly
important in North America where mortality and other health status indicators
vary widely between tribal and geographical indigenous populations. Further,
this analysis was for a 1-year period, if a longer period were available for
analysis, this could increase the consistency of the rates reported.
ConclusionThe indigenous peoples of New
Zealand and Australia suffer from high disease-specific mortality rates. The
relative size of indigenous/non-indigenous mortality disparities are highest in
New Zealand and Australia. There appears to be a number of common issues that
adversely affect the quality of the mortality data that is available in the four
countries. Action is required to address indigenous health disparities and to
improve the quality of indigenous mortality data
Author information:
Dale Bramley, Senior Lecturer, Section of Epidemiology and Biostatistics,
University of Auckland, Auckland; Paul Hebert, Assistant Professor, Department
of Health Policy, Mount Sinai School of Medicine, New York, NY, USA; Rod T
Jackson, Professor and Head of Section, Section of Epidemiology and
Biostatistics, University of Auckland, Auckland; Mark Chassin, Professor and
Chairman of the Department, Department of Health Policy, Mount Sinai School of
Medicine, New York, NY, USA
Acknowledgements:
During the research, Dale Bramley was on a Harkness Fellowship in healthcare
policy, with funding provided by the Commonwealth Fund of New York (USA). The
views presented here are those of the authors and not necessarily those of The
Commonwealth Fund, its director, officers, or staff.
We thank the following people who provided epidemiological
information for the purposes of compiling this paper: Adam Probert (First Nation
and Inuit Health Branch, Health Canada), Dr Fadwa Al-Yaman, (Australian
Institute of Health and Welfare), and Thomas D Dunn (Centers for Disease
Control, Atlanta, USA).
References:
|
|
||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|