Patterns, trends, and increasing disparities in mortality for Aboriginal and non-Aboriginal infants born in Western Australia, 1980–2001: population database study

doi:10.1016/S0140-6736(06)68771-0

The Lancet

Volume 367, Issue 9524, 27 May–2 June 2006, Pages 1758-1766

https://doi.org/10.1016/S0140-6736(06)68771-0 Get rights and content

Summary

Background

Since there are known disparities between Aboriginal and non-Aboriginal populations in Australia, trends in infant mortality rates can be used to assess the effectiveness of programmes designed to improve the health of Aboriginal populations. We have examined mortality rates in these populations in Western Australia.

Methods

We used the most comprehensive and valid total population data available for an Australian state to determine all-cause and cause-specific mortality for Aboriginal and non-Aboriginal infants born in Western Australia from 1980 to 2001.

Findings

Overall, infant mortality rates fell in both populations, but less so in Aboriginal (from 25·0 in 1980–84 to 16·1 in 1998–2001) than in non-Aboriginal infants (from 8·4 in 1980–84 to 3·7 in 1998–2001) such that disparities between the two groups increased for all major causes of infant death. The relative risk for Aboriginal compared with non-Aboriginal infants rose from 3·0 (95% CI 2·5–3·6) to 4·4 (3·5–5·5), and there were significantly more potentially preventable deaths, such as those caused by infections (5·9 per 1000 livebirths vs 0·7 per 1000 livebirths, RR 8·5, 95% CI 7·1–10·2). Additionally, for Aboriginal infants, postneonatal mortality rates were higher than neonatal mortality rates (11·2 per 1000 livebirths vs 9·7 per 1000 livebirths), trend analyses showed that previous reductions in deaths due to preterm birth (4·3 per 1000 livebirths—1·4 per 1000 livebirths from 1980–97) were not sustained in the most recent years studied (3·5 per 1000 livebirths), and rates of sudden infant death syndrome did not fall significantly (4·9 per 1000 livebirths vs 4·7 per 1000 livebirths).

Interpretation

These increasing disparities between Aboriginal and non-Aboriginal infants, especially in remote areas, demand immediate action in partnership with Aboriginal communities, focusing on both access to primary health care and better living conditions. Implementation and assessment of policies to reduce the continuing social and economic disadvantage faced by Aboriginal families are vital.

Introduction

A child's risk of death is greatest around time of birth and in the first year of life. For the past 100 years or so, children who survived the first year had a good chance of living to adulthood. Since the most powerful determinants of infant mortality are social and economic, death in infancy is a good indicator of the social progress of a society, country, or group of people. Many of the causes of infant death, such as infections, are potentially preventable; others, such as low birthweight or preterm birth can be treated with good health care and services. Although we do not know the cause or causes of sudden infant death syndrome (SIDS), the frequency of this diagnosis fell substantially in non-Aboriginal infants after risk-intervention campaigns.¹ Thus infant mortality is also an important measure of the effectiveness and availability of health services for mothers and children. As a result, any disparities in infant mortality, as are seen between rich and poor countries or between Aboriginal and non-Aboriginal children in Australia, are indications of inequalities in social and economic status and in the availability of health care.²

Not only birthweight but also the pattern of infant mortality differs between rich and poor populations. Those who are advantaged and have good services have higher neonatal than postneonatal mortality rates. The main causes of neonatal deaths are birth defects and preterm birth, which are not very amenable to either prevention or treatment. Postneonatal deaths are now mostly preventable—eg, infections and SIDS.

The excess burden of premature deaths in Aboriginal Australians and the disparity in the rates of infant and childhood deaths between Aboriginal and non-Aboriginal Australians are well known.3, 4, 5 A comprehensive and accurate profile of mortality is vital to enable effective prevention strategies, including appropriate health and other government policies to redress this disparity.

Despite a large amount of data describing the poor health status of Aboriginal people in Australia, few states have comprehensive death data for more than a few years, including those where most Aboriginal people live.⁶ Thus in 2003, when the National Health and Medical Research Council of Australia developed a plan for Aboriginal health research, a main recommendation was to obtain more data for deaths and for neglected groups such as urban Aboriginal and young people.⁷

Epidemiological data are essential to contribute to the prevention of infant and child deaths and should not only include mortality trends over time but also measure potential risk factors; these include maternal and infant factors, specific causes of death, where families live, and where the infants die (as measures of access to health-care services). Complete death information is available in the unique total population linked data in Western Australia for all births since 1980. We have comprehensively reviewed all deaths to provide a complete and accurate epidemiological description of trends and potential risk factors.

Section snippets

Population data

The main source of data was the linked, total population Western Australian Maternal And Child Health Research Database (MCHRDB).⁸ The study population consisted of all livebirths in Western Australia from 1980 to 2001 inclusive that were linked to infant deaths. These data are more than 99% complete for both Aboriginal and non-Aboriginal infants.⁹ Descriptive data for causes of death from the office of the Registrar General of Western Australia and autopsy reports held at the forensic

Results

We assessed total-population mortality data for 534 922 births in Western Australia between 1980 and 2001, in this birth cohort there were 3713 deaths. Births to Aboriginal mothers accounted for 6% of all births in Western Australia, but for 17% of the deaths (table 1). The IMR for Aboriginal infants was over three times that for non-Aboriginal infants. By contrast with non-Aboriginal infants, the rate of postneonatal death in Aboriginal infants was higher than that of neonatal deaths, but in

Discussion

This research has extended the previous reports of infant mortality in Western Australia and described all-cause and cause-specific mortality to the end of 2002. We have shown that the main causes of infant death in Western Australia are SIDS and infection in Aboriginal infants and birth defects and prematurity in non-Aboriginal infants.⁵ Notably, the only significantly reduced rates in Aboriginal infants were overall neonatal deaths and those due to sequelae of prematurity up to birth years

References (34)

FJ Stanley et al.
A population maternal and child health research database for research and policy evaluation in Western Australia
Semin Neonatol
(1997)
CD Holman et al.
Population-based linkage of health records in Western Australia: development of a health services research linked database
Aust N Z J Public Health
(1999)
M Gracey et al.
Environmental health conditions in remote and rural Aboriginal communities in Western Australia
Aust N Z J Public Health
(1997)
LD Elam-Evans et al.
Trends in the percentage of women who receive no perinatal care in the United States, 1980–1992: contributions of the demographic and risk effects
Obstet Gynecol
(1996)
LD Elam-Evans et al.
Patterns of prenatal care initiation in Georgia, 1980–1992
Obstet Gynecol
(1997)
CJ Freemantle
CJ Freemantle et al.
N Thomson
A review of Aboriginal health status
CJ Freemantle
Indicators of infant and childhood mortality for Indigenous and non-Indigenous infants and children born in Western Australia from 1980 to 1997 inclusive
(2003)
LM Alessandri et al.
Perinatal and postneonatal mortality among Indigenous and non-Indigenous infants born in Western Australia, 1980–1998
Med J Aust
(2001)

The health and welfare of Australia's Aboriginal and Torres Strait Islander People

(2005)

The NHMRC road map: a strategic framework for improving Aboriginal and Torres Strait Islander health through research

(2003)

V Gee et al.

Validation study of the Western Australian midwives' notification system 1992

(1994)

C Bower et al.

Ascertainment of birth defects: the effect on completeness of adding a new source of data

J Paediatr Child Health

(2000)

Freemantle CJ, Read AW, de Klerk NH, McAullay D, Stanley FJ. Sudden infant death syndrome: Aboriginal and...

V Gee et al.

Perinatal statistics in Western Australia: fourteenth annual report of the Western Australia midwives' notification system

(1998)

AW Read

Comparison of Indigenous status at birth with that at hospital admission using the maternal and child health research database

(1999)

Cited by (82)

The health and wellbeing of Indigenous adolescents: a global collective for an equitable and sustainable future
2022, The Lancet
Maternal fetal programming of birthweight among Australian Aboriginal infants: a population-based data linkage study
2019, The Lancet Global Health
Low birthweight, which is common among Australian Aboriginal infants, has been found to persist across generations because of shared genetic and environmental factors and possibly fetal programming. Fetal programming refers to the response of a fetus to hostile uterine conditions with lifelong effects and possibly, in turn, providing a poorer uterine environment for future offspring. Fetal programming might have a greater effect in populations that have undergone rapid lifestyle transitions—for example, Indigenous populations. Disentangling causal effects is difficult, but family-based approaches could provide insights. We explored whether poor maternal fetal growth caused low birthweight in Aboriginal infants.
In this data linkage study, we used linked administrative health records of 12 865 singleton Aboriginal infants born in Western Australia between 1980 and 2010 and their relatives (including siblings born in 2011). Electronic birth records included all births since 1980 with at least 20 weeks completed gestation or a birthweight of 400 g. We compared parental–offspring birthweight associations using three approaches—a regression analysis of the complete sample, adjusting for confounding variables; a comparison of the maternal–offspring and paternal–offspring associations; and a within-cousin group comparison. We used binary and continuous measures of birthweight. We categorised infants and their parents as small for gestational age (SGA) if their birthweight was below the first decile of birthweights for all singleton livebirths of the same sex and gestational age in Australia between 1998 and 2007.
The relative risk (RR) of SGA birth was higher for infants with SGA mothers than for those with non-SGA mothers (RR 1·65, 95% CI 1·49 to 1·83), after adjusting for grandmaternal parity. After additional adjustment for maternal height, the risk remained higher for those with non-SGA mothers (RR 1·51, 1·36 to 1·68). The maternal birthweight Z score coefficient was 0·17 (95% CI 0·14 to 0·20), compared with 0·13 (0·10 to 0·16) for paternal birthweight, a difference of 0·03 (−0·01 to 0·08). In the cousin analysis, the maternal–offspring association was fully attenuated (0·00, 95% CI −0·05 to 0·06). Conditions in the current pregnancy were strongly associated with offspring birthweight Z score. Smoking was associated with a mean decrease of 0·39 (95% CI −0·45 to −0·34) in offspring birthweight Z score, drug misuse with a decrease of 0·31 (−0·43 to −0·20), and diabetes with an increase of 0·58 (0·39 to 0·77).
We found little support for maternal fetal programming causing low offspring birthweight. The similar maternal and paternal influence on birthweight and our cousin analysis suggested transmission of genetic and environmental factors could explain much of the maternal–offspring birthweight association. Compared with other risk factors in the current pregnancy, fetal programming appears to have little or no role in the high numbers of infants with low birthweight among Aboriginal populations.
National Health and Medical Research Council of Australia and Bellberry Ltd.
New Frameworks for Understanding Sudden Unexpected Deaths in Infancy (SUDI) in Socially Vulnerable Families
2017, Journal of Pediatric Nursing
Citation Excerpt :
For many years, there have been known disparities in neonatal and postneonatal deaths between the richest 20% and poorest 20% of the population, and these disparities are greatest for postneonatal deaths (Lawn, Cousens, & Zupan, 2005; Luo et al., 2004). Deaths in the postneonatal period are, in the main, considered preventable, and largely associated with social determinants of health (Freemantle et al., 2006; Kleinman & Kiely, 1990). Approximately 94% of Sudden Unexpected Death in Infancy (SUDI) occur in the postneonatal period where it is the leading category of death (Queensland Family and Child Commission, 2016).
Sociological frameworks may enhance understanding of the complex and multidimensional nature of disadvantage, which is prevalent among families who experience Sudden Unexpected Death in Infancy (SUDI).
SUDI is the largest category of postneonatal death and largely associated with the social determinants of health. The highly successful ‘Back to Sleep’ campaign has resulted in a more than 85% decrease in SUDI. However, social inequalities have accompanied this decrease, and the burden of SUDI now lies with the most disadvantaged and socially vulnerable families. A considerable body of research on the phenomena of SUDI and disadvantage has been published over the last decade, demonstrating the widening social gradient in SUDI, and the importance in recognising structural factors and the multifactorial nature of disadvantage. Gaps in understanding of risk factors and scepticism about the received wisdom of health professionals have emerged as central themes in understanding why socially vulnerable families may adopt unsafe infant care practices. The direct impact of social disadvantage on infant care has also been recognised.
The translation of epidemiological findings regarding SUDI risk into public health recommendations for health professionals and families alike has to date focused on eliminating individual level risk behaviours. Unfortunately, such a model largely ignores the broader social, cultural, and structural contexts in which such behaviours occur. Translating the new knowledge offered by sociological frameworks and the principles of behavioural economics into evidence based interventions may assist in the reduction of SUDI mortality in our most socially vulnerable families.
Twenty-Five Year Survival of Children with Intellectual Disability in Western Australia
2017, Journal of Pediatrics
Citation Excerpt :
Although Aboriginal children are more likely to have an intellectual disability, with a greater risk of mild or moderate intellectual disability rather than severe intellectual disability,5 our study found that Aboriginal children with intellectual disability were not at any greater risk of mortality, compared with non-Aboriginal children with intellectual disability. However, within the cohort of children and young people without intellectual disability, we showed lower survival for Aboriginal compared with Caucasian children at all ages, as has been documented previously in the Australian population.16,33 This contradiction may be due to the fact that many Aboriginal children with intellectual disability have mild or moderate intellectual disability not necessarily associated with comorbid health conditions.5
To investigate survival up to early adulthood for children with intellectual disability and compare their risk of mortality with that of children without intellectual disability.
This was a retrospective cohort study of all live births in Western Australia between January 1, 1983 and December 31, 2010. Children with an intellectual disability (n = 10 593) were identified from the Western Australian Intellectual Disability Exploring Answers Database. Vital status was determined from linkage to the Western Australian Mortality database. Kaplan-Meier product limit estimates and 95% CIs were computed by level of intellectual disability. Hazard ratios (HRs) and 95% CIs were calculated from Cox proportional hazard regression models adjusting for potential confounders.
After adjusting for potential confounders, compared with those without intellectual disability, children with intellectual disability had a 6-fold increased risk of mortality at 1-5 years of age (adjusted HR [aHR] = 6.0, 95%CI: 4.8, 7.6), a 12-fold increased risk at 6-10 years of age (aHR = 12.6, 95% CI: 9.0, 17.7) and a 5-fold increased risk at 11-25 years of age (aHR = 4.9, 95% CI: 3.9, 6.1). Children with severe intellectual disability were at even greater risk. No difference in survival was observed for Aboriginal children with intellectual disability compared with non-Aboriginal children with intellectual disability.
Although children with intellectual disability experience higher mortality at all ages compared with those without intellectual disability, the greatest burden is for those with severe intellectual disability. However, even children with mild to moderate intellectual disability have increased risk of death compared with unaffected children.
Association of gestational age and growth measures at birth with infection-related admissions to hospital throughout childhood: a population-based, data-linkage study from Western Australia
2016, The Lancet Infectious Diseases
Reduced gestational age and low birthweight are associated with an increased risk of neonatal infections. However, the long-term risk of infection, especially in late preterm infants or those at near-normal birthweight, is unknown. We estimated whether rates of infection-related admissions to hospital for children in Western Australia were associated with age, gestational age, birthweight, and birth length.
We did a population-based, data-linkage study using total-linked, registry data from the Western Australia Birth Register of all liveborn, non-Indigenous Australian singleton births recorded from Jan 1, 1980, to Dec 31, 2010. We followed up individuals from birth-related hospital discharge to age 18 years, death, or end of 2010, and linked to data about subsequent admissions to hospital or death registrations. Gestational age was assessed from both the last menstrual period and from estimates based on ultrasonography. We categorised birthweight by 500 g bands and birth length by 5 cm bands, and approximated the reference ranges for both to the 50th percentile. Because size at birth and gestational age are strongly associated, we calculated Z scores for gestational-specific and sex-specific birthweight, birth length, and ponderal index. Our primary outcomes were the number and type of infection-related admissions to hospital. We used multilevel negative binomial regression to generate rate ratios (RR) for such admissions, identified by codes from the International Classification of Diseases, versions 9 and 10-AM. We adjusted the RRs for maternal age at delivery, birth year, birth season, parity, sex, 5-min Apgar score, delivery method, socioeconomic status, and bronchopulmonary dysplasia.
Of 719 311 liveborn singletons included in the analysis and followed up for 8 824 093 person-years, 365 867 infection-related admissions to hospital occurred for 213 683 (30%) children. Of the 719 311 children included in the analysis, 137 124 (19%) had one infection-related admission to hospital, 43 796 (6%) had two, 16 679 (2%) had three, and 16 084 (2%) had four or more. The 365 867 admissions to hospital included a diagnosis of infection of the upper respiratory tract for 174 653 (48%), the lower respiratory tract for 74 297 (20%), the gastrointestinal tract for 44 755 (12%), and a viral infection for 37 213 (10%). Infection-related rates of admissions to hospital increased by 12% for each week reduction in gestational age less than 39–40 weeks (RR 1·12, 95% CI 1·12–1·13), by 19% for each 500 g reduction in birthweight less than 3000–3500 g (1·19, 1·18–1·21), and by 41% for each 5 cm reduction in birth length less than 45–50 cm (1·41, 1·38–1·45). Gestational age-specific and sex-specific birthweight Z scores lower than the 25th to 50th percentile and birth length Z scores lower than the 10th to 25th percentile were associated with increased rates of infection-related admissions to hospital (eg, 1st–5th percentile RR 1·15, 95% CI 1·12–1·19, and 1·11, 1·07–1·14, respectively). Ponderal index Z scores lower than the 25th to 50th percentile were also associated with increased rates of infection-related admissions (eg, 1st–5th percentile RR 1·08, 95% CI 1·04–1·12). A gestational age of 41 weeks or later, a birthweight or birth length Z score above the 50th percentile, or a ponderal index Z score between the 75th and 95th percentile, were associated with modestly reduced rates of infection-related admissions to hospital.
Children who were born with reduced gestational age, birthweight, and birth length have persistently increased rates of infection-related admissions to hospital until age 18 years. Pregnancy outcomes should be optimised to prevent infection occurring in this population, especially in resource-limited settings where suboptimum intrauterine growth and moderate prematurity are common.
Australian National Health and Medical Research Council.
It looks like a breadbox': A pilot study investigating implementation of the Pepi-Pod® program with Aboriginal families in metropolitan South Australia
2021, Primary Health Care Research and Development
This cross-sectional study is the first one to explore the collaboration of the influencing factors thereof amongst general practitioners (GPs) and gynaecologists (Gyns) working in primary care in urban and rural settings in Germany.
The number of women aged ≥ 50 years is predicted to increase in the next years in Germany. This coincides with the ageing of primary care specialists providing outpatient care. Whereas delegation of tasks to nurses as a form of interprofessional collaboration has been the target of recent studies, there is no data regarding collaboration amongst physicians in different specialisations working in primary care. We explored collaboration amongst GPs and Gyn regarding the healthcare provision to women aged ≥ 50 years.
A quantitative postal survey was administered to GPs and Gyns in three federal states in Germany, focusing on care provision to women aged ≥ 50 years. A total of 4545 physicians, comprising 3514 GPs (67% of the total GP population) randomly selected, and all 1031 Gyns practicing in these states received the postal survey in March 2018. A single reminder was sent in April 2018 with data collection ending in June 2018. Multiple logistic regressions were performed for collaboration, adjusted by age and sex, alongside descriptive methods.
The overall response rate was 31% (1389 respondents): 861 GPs (25%) and 528 Gyns (51%), with the mean respondent age being 54.4 years. Seventy-two per cent were female. Key competencies of collaboration are associated with working in rural federal states and with network participation. Physicians from rural states [odds ratio (OR) = 1.5, 95% confidence interval (CI) = 1.2, 1.9] and physicians in networks (OR = 3.0, CI = 2.3, 3.9) were more satisfied with collaboration. Collaboration to deliver services for women aged ≥ 50 years is more systematic amongst GPs and Gyns who are members of a network; increased networking could improve collaboration, and ultimately, outcomes too.

View all citing articles on Scopus

View full text

ArticlesPatterns, trends, and increasing disparities in mortality for Aboriginal and non-Aboriginal infants born in Western Australia, 1980–2001: population database study