Social and racial inequalities in birth rates and infant outcomes
in Western Australia
Amanda Teresa Langridge
BSc(Mathematical Sciences), BSc(Hons)
School of Paediatrics and Child Health
University of Western Australia
This thesis is presented for the degree of Doctor of Philosophy of
The University of Western Australia
2009
xii
ABSTRACT
BACKGROUND
Despite many developed countries reaching unprecedented levels of national wealth in
recent decades, fertility rates have continued to decline and inequalities in children’s
health and development are increasing. A recent UNICEF report indicated that
between 1970 and 2006 the crude birth rate in developed countries fell by 35%, while a
bulletin from the World Health Organisation clearly states that inequalities in health are
not only persisting, but also increasing in some countries.
Research has shown that even prior to conception, a child’s health and development is
influenced by maternal factors and the various environments in which they grow and
learn, including the family and community in which the child is embedded. Children
from lower socioeconomic groups have higher mortality, more chronic health problems,
poorer general health and higher rates of developmental delay and hospitalisations. It
has also been established that poor birth outcomes, such as low birthweight, are
associated with an increased risk of behavioural problems and learning difficulties, as
well as poorer health outcomes during adulthood. These findings highlight the
importance of biological, psychological and social experiences before and during
pregnancy, as well as during the early years of life. Additionally, similar to adult health
outcomes, important early child health outcomes, such as low birth weight and preterm
birth, have a strong social gradient and striking racial disparity. However, little research
has attempted to explain why social and racial inequalities in birth outcomes exist, or
investigated how these inequalities have changed over time.
AIMS
• To examine trends in birth rates in Western Australia and investigate the effects of
a National monetary incentive introduced to increase birth rates,
xiii
• To develop a measure of individual-level socioeconomic status using routinely
collected parental occupation titles and investigate the differences between individual
and area-level socioeconomic status,
• To investigate racial inequalities in poor fetal growth and preterm birth, and
examine social inequalities in those outcomes by Aboriginality.
METHODS
The study population comprised all infants born in Western Australia between 1980-
2006. Data were sourced from individual record linked data from administrative
datasets, including the midwives’ notification and birth registrations. Statistical models
including negative binomial regression and multivariate logistic regression were applied
to assess trends over time whilst accounting for confounders and predictors associated
with birth rates and infant outcomes. Multilevel models were also applied to examine
the effects of individual and area-level socioeconomic measures on the outcome
variables.
RESULTS
After a steady decline from 1984, the birth rate increased significantly following the
introduction of a National monetary incentive in 2004 with increases among most
maternal age groups (15-49years), in both Aboriginal and non-Aboriginal women, in all
residential locations (urban, regional, rural) and among all community level
socioeconomic groups. Examination of the individual-level socioeconomic measure
developed from the routinely collected data on parental occupation at the time of birth
showed there was considerable variation in individual socioeconomic status within
area-level socioeconomic groups. It demonstrated the importance of including both
individual and area-level socioeconomic data when examining inequalities.
Prior to plateauing at 15.7% from 2003, there had been a steady decline in the
percentage of term singleton infants with poor fetal growth over the last 23 years. The
xiv
proportion of infants with poor fetal growth rose with increasing levels of social
disadvantage, with the difference between the most and least disadvantaged groups
increasing considerably over time, from 6% in 1984-1988 (16%-22%) to 10% in 2004-
2006 (13-23%). Aboriginal infants were significantly more likely to be growth restricted
than non-Aboriginal infants and examination of social inequality by Aboriginality
revealed that, although the disparity increased over time, it was greater among non-
Aboriginal infants. Findings also revealed that while the rate of preterm births remained
fairly static over the last 23 years, inequalities between Aboriginal and non-Aboriginal
infants increased. While a considerable portion of the disparity is attributable to
parental socioeconomic and demographic characteristics, it persists even after
adjustment of these factors.
CONCLUSIONS
This thesis highlights that aggregated data used by policy makers may be misleading,
as results demonstrate how overall trends mask increases in social and racial
inequalities. The thesis further reveals the importance of regularly monitoring trends in
birth outcomes by social characteristics. Failure to do this may result in incorrect or
misguided implementation of policies and interventions that do not target the very
populations that need them, and may lead to poorer infant health outcomes and greater
health disparities.
xv
ACKNOWLEDGEMENTS There are many people who provided support and encouragement throughout my PhD,
without whom this thesis may not have been possible. I am particularly thankful for my
two primary supervisors, Jianghong Li and Natasha Nassar, for their continued support,
guidance, motivation and dedication. I am also extremely grateful for their mentorship
and support and encouragement to take opportunities to enhance my professional
development, including travelling to national and international conferences. This has
been an invaluable experience, allowing me to develop my knowledge and research
skills as part of my journey towards becoming a successful independent researcher.
I would also like to thank Fiona Stanley and the Developmental Pathways Project
Team for the opportunity to be part of a unique and pioneering project. Being actively
involved with the progression and expansion of the Developmental Pathways Project
over the last four years has been a truly amazing experience. I would like to give a
special thanks to Rebecca Glauert, the Developmental Pathways Project Coordinator,
who has been fantastic, providing support, advice and motivation.
To my friends, the last four years would have been so much harder without your
friendship, patience and encouragement. To my fellow PhD student, Melissa
O’Donnell, you made this journey so much more bearable and fun.
Finally, I am forever grateful to my incredible family - Mum, Dad, Alan, Steven and
Stacey. Your unconditional love, support, encouragement and devotion has been
amazing – you made this journey so much easier!
1
TABLE OF CONTENTS
Abstract ....................................................................................................................... xii
Background .............................................................................................................. xii
Aims ........................................................................................................................ xii
Methods .................................................................................................................. xiii
Results .................................................................................................................... xiii
Conclusions ............................................................................................................ xiv
Acknowledgements ...................................................................................................... xv
Table of Contents ......................................................................................................... 1
List of Figures ............................................................................................................... 4
List of Tables ................................................................................................................ 6
Chapter One: Introduction ............................................................................................. 8
The Bioecological model ........................................................................................... 9
Developmental health ............................................................................................. 11
Social determinants of health .................................................................................. 13
Contextual vs. compositional differences ................................................................ 18
Area vs. individual-level socioeconomic measures .................................................. 19
Measuring socioeconomic position ......................................................................... 20
Racial inequalities ................................................................................................... 21
Aboriginal health and inequalities ............................................................................ 22
The importance of investigating and monitoring social and racial inequalities ......... 23
Chapter Two: General Methods .................................................................................. 26
Introduction ............................................................................................................. 26
Data sources ........................................................................................................... 26
Midwives’ Notification System ............................................................................. 26
Birth Registrations ............................................................................................... 27
Hospital Morbidity Data System .......................................................................... 28
Census Data ....................................................................................................... 28
Data Linkage in Western Australia .......................................................................... 29
The Linkage Process .......................................................................................... 30
Linkage without consent ...................................................................................... 34
Ethics Approvals ..................................................................................................... 35
Statistical methods .................................................................................................. 36
Chapter Three: Birth rates and parental characteristics .............................................. 37
Introduction ............................................................................................................. 37
Data sources ........................................................................................................... 37
2
Data analysis ...................................................................................................... 37
Calculation of birth rates ..................................................................................... 38
Birth rates in Western Australia 1984 to 2006 ......................................................... 39
Birth rates by area-level social disadvantage .......................................................... 40
Birth rates by maternal age group ........................................................................... 41
Birth rates by maternal Aboriginality ........................................................................ 42
Plurality ................................................................................................................... 43
Maternal median age at first birth ............................................................................ 43
Maternal residential location at time of birth ............................................................ 44
Parity ...................................................................................................................... 44
Maternal smoking during pregnancy ....................................................................... 44
Paternal age ........................................................................................................... 45
Paternal Aboriginality .............................................................................................. 45
Paternal median age at first birth ............................................................................ 46
Marital status at time of birth ................................................................................... 47
Conclusion .............................................................................................................. 48
Chapter Four: Do monetary incentives increase birth rates in socially
disadvantaged groups? ............................................................................................... 49
Introduction ............................................................................................................. 49
Methods .................................................................................................................. 51
Data Sources ...................................................................................................... 51
Independent Variables ........................................................................................ 51
Analyses ............................................................................................................. 52
Ethics .................................................................................................................. 54
Results .................................................................................................................... 54
Discussion .............................................................................................................. 60
Chapter Five: improving the monitoring of health inequalities using routinely
collected data to obtain an individual-level socioeconomic indicator ........................... 65
Introduction ............................................................................................................. 65
Methods .................................................................................................................. 67
Data Sources ...................................................................................................... 67
Genealogical Data .............................................................................................. 68
Parental Occupation ............................................................................................ 68
Validation of Occupation Codes .......................................................................... 70
Area-level Socioeconomic Status ........................................................................ 70
Poor fetal growth ................................................................................................. 70
Statistical Analyses ............................................................................................. 71
Ethics .................................................................................................................. 72
3
Results .................................................................................................................... 72
Discussion .............................................................................................................. 83
Chapter Six: Rising social and racial inequalities amongst infants with poor
fetal growth in Western Australia, 1984 to 2006 .......................................................... 88
Introduction ............................................................................................................. 88
Methods .................................................................................................................. 89
Data Sources ...................................................................................................... 89
Low birth weight .................................................................................................. 90
Independent Variables ........................................................................................ 91
Ethics .................................................................................................................. 93
Results .................................................................................................................... 94
Discussion ............................................................................................................ 104
Chapter Seven: Social and racial inequalities in preterm births in Western
Australia, 1984 to 2006 ..............................................................................................109
Introduction ........................................................................................................... 109
Methods ................................................................................................................ 111
Data Sources .................................................................................................... 111
Variables ........................................................................................................... 111
Statistical Analyses ........................................................................................... 113
Ethics ................................................................................................................ 114
Results .................................................................................................................. 114
Discussion ............................................................................................................ 121
Chapter Eight: Summary of findings and future directions ..........................................125
References ................................................................................................................125
Appendix A: Socioeconomic Indexes For Areas (SEIFA) ...........................................178
Constructing The Socioeconomic Indexes For Areas ............................................ 179
SEIFA Variables ................................................................................................... 180
Appendix B: Occupation Groups ................................................................................184
4
LIST OF FIGURES
Figure 1: Correlation between income inequality and the UNICEF index of child
wellbeing in 23 rich countries (Adapted from (Pickett & Wilkinson 2007)) ................... 18
Figure 2: WA Data Linkage System as of April 2009 (Modified (Department of Health)
2009) .......................................................................................................................... 33
Figure 3: The Linking Process - Steps 1, 2 and 3 (Modified (Kelman et al. 2002b)) ... 34
Figure 4: The Linking Process - Step 4 (Modified (Kelman et al. 2002b)) .................... 35
Figure 5: Birth rates in Western Australia .................................................................... 39
Figure 6: Birth rates by area-level social disadvantage ............................................... 40
Figure 7: Birth rates by maternal age group ................................................................ 41
Figure 8: Birth rates by maternal Aboriginality ............................................................. 42
Figure 9: Percentage of births by marital status .......................................................... 47
Figure 10: Actual birth rates and fitted trend ............................................................... 55
Figure 11: A - Adjusted^ birth rate for maternal age group by year; B - Adjusted^ birth
rate for Aboriginality by year; C - Adjusted^ birth rate for location by year; D - Adjusted^
birth rate for index of education and occupation by year ............................................. 59
Figure 12: Maternal skill level distribution ................................................................... 73
Figure 13: Paternal skill level distribution .................................................................... 73
Figure 14: Adjusted^ rate of poor fetal growth for Father’s skill level by area-level
disadvantage .............................................................................................................. 82
Figure 15: Adjusted^ rate of poor fetal growth for Mother’s skill level by area-level
disadvantage .............................................................................................................. 83
Figure 16: Trends in poor fetal growth (POBW) in Western Australia: 1984-2006 ....... 94
Figure 17: Inequalities in poor fetal growth (POBW) by area-level disadvantage in
Western Australia: 1984-2006 ..................................................................................... 95
Figure 18: Poor fetal growth by Aboriginality and social disadvantage in Western
Australia: 1984-2006 ................................................................................................... 96
5
Figure 19: Trends and inequalities in preterm births by area-level social disadvantage
in Western Australia, 1984-2006 ............................................................................... 115
Figure 20: Preterm births by Aboriginality and area-level social disadvantage in
Western Australia, 1984-2006 ................................................................................... 116
6
LIST OF TABLES Table 1: Plurality ......................................................................................................... 43
Table 2: Maternal median age at first birth .................................................................. 43
Table 3: Maternal characteristics ................................................................................ 45
Table 4: Paternal Characteristics ................................................................................ 46
Table 5: Paternal median age at first birth................................................................... 46
Table 6: Socioeconomic and demographic characteristics of mothers who gave birth in
2004 to 2006 ............................................................................................................... 56
Table 7: ANZSCO skill levels ...................................................................................... 69
Table 8: Paternal skill level by maternal skill level ....................................................... 74
Table 9: Maternal skill level at first birth compared to highest skill level reported across
all births ...................................................................................................................... 75
Table 10: Paternal skill level at first birth compared to highest skill level reported across
all births ...................................................................................................................... 76
Table 11: Maternal and paternal skill level by area level social disadvantage ............. 77
Table 12: Unadjusted Odds Ratios predicting poor fetal growth (Births: 1999 – 2006)78
Table 13: Maternal and paternal skill level and area-level social disadvantage
predicting poor fetal growth (Births: 1999 – 2006) ....................................................... 79
Table 14: Racial disparities in poor fetal growth and contributing factors in Western
Australia: 1984-2006 ................................................................................................... 99
Table 15: Absolute and relative differences in poor fetal growth births by area-level
disadvantage in non-Aboriginal infants ..................................................................... 100
Table 16: Absolute and relative differences in poor fetal growth births by area-level
disadvantage in Aboriginal infants ............................................................................ 102
Table 17: Racial disparities in preterm births and contributing factors in Western
Australia: 1984-2006 ................................................................................................. 117
Table 18: Social inequalities in preterm birth and contributing factors: non-Aboriginal
infants ....................................................................................................................... 119
7
Table 19: Social inequalities in preterm birth and contributing factors: Aboriginal infants
................................................................................................................................. 120
Table 20: Index of Relative Socioeconomic Disadvantage ........................................ 181
Table 21: Index of Economic Resources ................................................................... 182
Table 22: Index of Education and Occupation ........................................................... 183
8
CHAPTER ONE: INTRODUCTION
“The enjoyment of the highest attainable standard of health is one of the
fundamental rights of every human being without distinction of race, religion,
political belief, economic or social condition.”
– WHO, 1948
Despite many developed countries reaching unprecedented levels of national wealth in
recent decades, fertility rates have continued to decline and research clearly
demonstrates that inequalities in children’s health and development are increasing. A
recent UNICEF report indicated that between 1970 and 2006 the crude birth rate in
developed countries fell by 35% (United Nations Children’s Fund (UNICEF) 2007),
while a bulletin from the World Health Organisation states that health inequalities are
not only persisting, but in some countries, increasing (WHO 2004). Within the context
of these important trends, this thesis has two purposes. The first is to examine trends in
birth rates in Western Australia and determine whether the fertility policy introduced by
the Australian Government in 2004 is associated with an increase in birth rates. The
second aim of this thesis is to investigate contributing factors to social and racial
inequalities in poor fetal growth and preterm birth and to determine whether inequalities
in these infant outcomes have increased over the last two decades.
This introduction is organised around several key areas. Firstly, research has shown
that even prior to conception, a child’s health and development is influenced by
parental factors and the various environments in which they grow and learn, including
the family and community environment. There is still a critical need to investigate
factors that impact on infant health and development the most and whether these
factors have changed over time. Secondly, research has shown significant social and
racial inequalities in poor health outcomes. However, we know little about how social
9
and racial inequalities have changed over time and how these changes have impacted
on infant health and development. Finally, research previously conducted in this area
has been limited due to a lack of population-level longitudinal data, and a lack of
information on the social environment at various levels, such as the child, family and
community level. Not only has this limitation prevented the examination of how
changing social and racial inequalities have affected child health and development, but
it has also restricted the ability to examine how the various levels of social organisation
jointly influence perinatal outcomes, as well as the ability to monitor inequalities over
time.
While there is an abundance of literature available regarding the social determinants of
health and social inequalities, the purpose of this review is not to be exhaustive, but to
provide an overview of the relevant theoretical background behind this thesis and the
questions it seeks to answer. Detailed literature reviews specific to each of the aims of
this thesis can be found within each chapter.
THE BIOECOLOGICAL MODEL
In 1977, Bronfenbrenner proposed an ecological model of development. Based on this
model the developing child is nested within a series of systems, which interacts with
the developing child. Bronfenbrenner has termed these systems the microsystem,
mesosystem and exosystem. Bronfenbrenner (1977) describes the microsystem as the
interaction between the developing child and their immediate setting, such as their
home or school, the mesosystem as a collection of microsystems and the interactions
between them, and the exosystem as an extension of the mesosystem that contains
other social structures. However, unlike the micro and mesosystem, the developing
child is not an active participant in the exosystem. The ecological model also includes
the macrosystem, which differs from the other three systems, as it, ‘refers to the
overarching institutional patterns of the culture or subculture, the economic, social,
10
educational, legal, and political systems, of which micro-, meso- and exosystems are
the concrete manifestations,’ (Bronfenbrenner 1977).
While Bronfenbrenner’s original ecological model of development considers the
interactions that occur between the developing person and their surroundings, it does
not account for the biological aspect of development. As a result of this, in 1994,
Bronfenbrenner and Ceci extended on the ecological model and proposed the
bioecological model. The bioecological model ‘proposes and provides for the
assessment of mechanisms, called proximal processes, through which genetic
potential are actualised’ (Bronfenbrenner & Ceci 1994).
Within each system of the bioecological model, there are factors that influence the
developing child, including the resources that are available to that child. In 1995,
Brooks-Gunn and colleagues, proposed the family and community resources
framework, making an important contribution to developmental health research
(Brooks-Gunn et al. 1995). This framework was based on work by sociologists
Coleman (1988; Coleman 1990) and Haveman and Wolfe (Haveman & Wolfe 1994), as
well as economist Becker (Becker 1964), and identifies resources within the family and
community that they see as being critical for the developing child. At the family level,
these resources include income, time, human capital and psychological capital, while at
the community level these resources include child care, school, peer groups and
community groups (Brooks-Gunn 1995). Human capital includes parental knowledge
(including educational attainment), occupation, skills, beliefs and attitudes, while
psychological capital includes parental characteristics, parenting behaviour and mental
health (Becker 1964). With the development of this framework and the identification of
resources influencing child development, researchers are now able to measure these
resources at multiple levels, including the child, family, community and school.
11
Both the ecological model and the family and community resource framework are two
cornerstones of the developmental health perspective, in recognising that multilevel
models are critical for a better understanding of the complex causal pathways of
optimal human development from conception to birth and later life. These theoretical
frameworks inspire the multilevel analytical approach of this thesis, in an attempt to
understand poor birth outcomes and social and racial inequalities in them.
DEVELOPMENTAL HEALTH
Developmental health was a term first used by Keating and Hertzman (Keating &
Hertzman 1999) to describe the full range of developmental outcomes, including
physical and mental health, behavioural and educational, that demonstrate the social
gradient. Developmental health integrates the knowledge and understanding of a
variety of disciplines, including social epidemiology, psychology and biology (Keating &
Hertzman 1999) and is essentially concerned with the study of individual, family and
area level factors that contribute to healthy human development. It is primarily
interested in the prevention of problems through early intervention. As a result of this
focus on early intervention, understanding the pathways through which factors
influence development in the early years of life is of utmost interest and importance to
developmental health researchers.
Research from a variety of disciplines has established that human development is
determined by complex interactions of biological, psychological and social experiences
(Keating & Hertzman 1999); (Rutter 2002). Furthermore, research has shown that
some of these complex interactions have lasting effects, with maternal behaviours and
health prior to conception affecting the unborn foetus in significant ways (Kramer et al.
2002; Perry & Lumey 2004; Barker 1998). The importance of the maternal-fetal
interactions during pregnancy was highlighted by the work of David Barker in the
1990’s, resulting in the formation of Barker’s Hypothesis. Barker (Barker 1995)
12
suggests that if a foetus receives a limited supply of nutrients, then it will adapt,
resulting in permanent changes to its structure and metabolism. According to Barker’s
hypothesis, it is these changes that are responsible for the increased risk of chronic
diseases later in life, such as cardiovascular disease, diabetes, obesity and
hypertension (Barker 1998; Rich-Edwards et al. 1997). Consistent with this hypothesis,
Wadsworth (Wadsworth 1999; Wadsworth & Butterworth 2006) highlighted the
importance of biological, psychological and social experiences both before and during
pregnancy, as well during the early years of life: “Research…shows the vital role
played by social circumstances in the development of health in early life, beginning with
the mother’s health in the prenatal years, as well as during pregnancy, and in infancy
and childhood...the effects of the interactions during the early life period set the scene
for what follows, and form the basis of the individual’s biological, psychological, and
human capital” (Wadsworth & Butterworth 2006).
While the concept that early-life experiences have long-term health and social
implications is not new, little is known about how these experiences affect the biology
of the child which then results in the long-term consequences for human development
(Hertzman 2002). One explanation that has been suggested for how early life
experiences impact on health and well-being later in life is biological embedding. The
idea of biological embedding’ was proposed by Hertzman and Weins (1996) and is the
process by which experience affects the health and well-being across the life course,
and is seen to be the key link between human development and population health
(Keating & Hertzman 1999; Hertzman & Wiens 1996). Central to the idea of biological
embedding are the critical and sensitive periods that occur in the early years of life.
Keating and Miller (1999a) describe critical periods as those developmental stages
when the presence or absence of certain experiences impact upon the functionality of
the developing biological system. Outside of these critical periods, the experiences
have little or no effect on the developing system. One such example of a biological
system that has a critical period is the visual system (Cynader & Frost 1999). Unlike
13
critical periods, sensitive periods are those periods in which the experiences are
important for setting a base for the developing system, but where experiences do not
have the all-or-nothing effect that is seen with the critical periods (Keating & Miller
1999a). Examples include emotional and attention regulation (Keating & Miller 1999b).
In attempting to understand the process by which social circumstances affect
developmental health outcomes, Hertzman has proposed two explanatory models.
These models are the latency model and pathways model. The latency model is closely
tied to the idea of critical and sensitive periods and proposes that the psychosocial and
socioeconomic conditions in early life have a strong impact in later life, ‘independent of
intervening experiences’ (Hertzman 1999). The pathways model however, proposes
that there is a ‘cumulative effect’, and it is the addition of psychosocial and
socioeconomic conditions throughout the life course that impact in later life (Hertzman
1999).
SOCIAL DETERMINANTS OF HEALTH
The social determinants of health approach maintains that the social and economic
conditions affect the health and well-being of individuals, communities and populations.
Broadly speaking, there are three general categories of the social determinants of
health. These are the socioeconomic determinants, such as race, class, gender,
education, occupation and income, psychosocial risk factors, such as poor social
networks, self-esteem, chronic stress and isolation, and community and societal
characteristics, such as community participation, poverty and income inequality. It is
these social and economic circumstances that affect health throughout the life course
and produce pervasive social gradients in health. The social gradient has been well
documented and clearly shows that health improves with increasing socioeconomic
position and that health inequalities that exist are not simply differences between the
‘rich’ and ‘poor’ (Marmot et al. 1991). The Black Report (Townsend et al. 1988) and the
14
Whitehall study (Marmot et al. 1984) were among the first studies to provide solid
evidence that each socioeconomic position has slightly better health than the position
below it, and slightly worse health than the position above it. Since these two studies,
many studies have confirmed that finding, demonstrating that the social gradient exists
across time and the lifespan for many health and developmental outcomes, including
the early years (Brooks-Gunn et al. 1999). Research has shown that the social gradient
exists in infant mortality and low birthweight (Petrou et al. 2006; Quine & Quine 1993;
Quine 1991; Mortensen et al. 2008), during childhood and adolescence in injurious
deaths, obesity and cognitive and socioemotional development (Shaw et al. 2005;
Starfield et al. 2002; Hertzman & Wiens 1996), during early adulthood in deaths from
injuries and mental health problems (Taylor et al. 2004; Fone & Dunstan 2006; Power
et al. 1997; Naess et al. 2004; Naess et al. 2007) and in middle age and late life, in
mortality and morbidity (e.g. chronic diseases) (House et al. 1994; Brooks-Gunn et al.
1999; Keating & Hertzman 1999).
A review of more than 200 Australian studies (Turrell et al. 1999) and numerous
international studies that focused on or included infants and children found that infants
from a lower socioeconomic position were more likely to have low birthweight for
gestational age (Bor et al. 1993; Morrison et al. 1989a; Beard et al. 2009; Kramer et al.
2000a), experience developmental delays (Najman et al. 1992; Carmichael & Williams
1983a; Sonnander & Claesson 1999), have prolonged periods of acute illness
(Carmichael & Williams 1983b; McConnochie et al. 1999; Margolis et al. 1992),
increased need for resuscitation after birth (Jonas et al. 1992), had higher rates of
preterm delivery (Fisher 1978; Kramer et al. 2000b; Morgen et al. 2008), have hospital
stays of 28 days or more after birth (Fisher 1978), experience more hospital
admissions (Taylor et al. 1992; Petrou et al. 2006; Marcin et al. 2003) and were less
likely to be breastfed, and when breastfeeding did occur, were more likely to be
breastfed for a shorter duration (Hitchcock et al. 1982; Lawson & Tulloch 1995; Scott et
al. 1997; Sweet & Power 2009). The findings for infant mortality were more
15
inconsistent, with some reporting that rates of mortality were higher in infants from a
lower socioeconomic position (Quine & Quine 1993; Quine 1991; Singh & Yu 1995;
Judge 2009; Collins & David 1992), and others finding no association (Morrison et al.
1989a; Smeeton et al. 2004). The Australian review also highlighted that no Australian
studies had examined the relationship between socioeconomic position and risk factors
among infants (Turrell et al. 1999).
Several theories have been proposed in an attempt to explain the relationship between
socioeconomic status and health and how inequalities due to socioeconomic position
occur. These theories each have a slightly different focus and include the materialist or
structuralist theory (Frohlich et al. 2001; Macintyre 1997), psychosocial theory (Marmot
& Wilkinson 1999; Siegrist & Marmot 2004; Evans & Stoddart 2003; Krieger 2001)
(Wilkinson & Pickett 2009; Kawachi & Kennedy 1997), neo-material theory (Lynch et al.
2000); (Dunn et al. 2006; Ross et al. 2000), social production of disease or political
economy of health theory (Doyal 1979; Conrad & Kern 1981; Link & Phelan 1995; Link
& Phelan 1996; Raphael 2002) and finally, the ecosocial theory (Krieger 1994). The
materialist theory essentially suggests that it is absolute social position and a lack of
resources that affects health, while the psychosocial theory proposes that it is not
absolute position or income that affects health, but rather how the psychosocial
aspects of relative social position and relative income is perceived. The neo-material
theory proposes that inequalities occur as the result of individuals negative experiences
and lack of resources, coupled with an underinvestment in a range of infrastructure,
including human, physical, health and social infrastructure (Lynch et al. 2000). The
social production of disease theory suggests that in the move towards capitalist and
neo-liberal societies, it is the accumulation of wealth and power and the importance of
material assets that result in inequalities, with the rich becoming richer and the bulk of
the wealth being controlled by few (Townsend 1986; Sanders 1985; Raphael 2002).
Finally, the ecosocial theory suggests a spatiotemporal, multilevel perspective that
16
recognises the interactions between the social, physical and biological environments to
create inequalities (Krieger 1994; Krieger 2001).
The social determinants of health approach draws attention to social and economic
policies. In Australia it is likely that these policies have contributed to the increase seen
in health and social inequalities (Conley 2004) and these are driven by the political
party of the day. Over the last few decades the political party in power has changed
very few times, with the Liberal Party in power between 1977 and 1983, and then again
in 1996-2007 and the Labor Party leading the nation in 1983-1996. Traditionally,
Labour party policies are targeted towards public ownership of key industries,
government intervention in the economy, redistribution of wealth and publicly funded
healthcare and education. Towards the end of their term in office, it was generally
believed that Labour party policies became slightly more market-oriented in response
to international globalisation, resulting in increased inequalities (Conley 2004; Taylor
1995). Liberal Party policies tend to be oriented towards more neo-liberal and market-
oriented policies, supporting free trade and minimal government intervention in the
economy. These policies shift the focus from society and on to the individual, where for
example, in regards to health, society’s responsibility for the health of the population is
reduced and the onus for improving health is transferred to the individual. This can
clearly be seen in countries where there has been a strong push for private insurance-
based health systems over public systems.
The importance of these policies in reducing inequalities was highlighted in the World
Health Organizations Commission on Social Determinants of Health (CSDH) Final
Report, Closing the Gap in a Generation (Commission on Social Determinants of
Health 2008). While the CSDH provides recommendations regarding social and
economic policies aimed at reducing inequalities, the authors highlight that for these
recommendations to be successful there will need to be a shift away from the current
neo-liberal policies, which they have attributed as the cause for the large increase in
17
inequalities (Coburn 2000); (Langille 2004) (Navarro 2009). They also suggest that
neo-liberal policies have been detrimental, and have led to increases in inequalities, as
they encourage governments to reduce their involvement in economic and social
activities, deregulate labour and financial markets and promote privatisation (Navarro
2009; Langille 2004; Coburn 2004). Furthermore, they maintain that neo-liberal policies
favour those closer to the top of the socioeconomic ladder, which Coburn (Coburn
2004) argues results in, ‘major increases in social inequality, poverty, income inequality
and social fragmentation,’ (Coburn 2004).
The relationship between income inequality and health has been well-established.
Although the underlying reason of why income distribution affects health is not
completely understood, it is apparent that the way in which society deals with emerging
social inequalities and reinforces social hierarchies plays an important role (Keating &
Hertzman 1999). Previous research has shown that life expectancy increases as per
capita income increases (Wilkinson 1996). However, as countries became wealthier
the relationship between health and wealth becomes more complex. The relationship
between life expectancy and per capita income changed in the 1990’s, as some
countries began to achieve higher levels of wealth. As these new levels of wealth were
achieved, the health-wealth curve, that had been previously observed, began to
plateau, so that increases in income per capita were no longer producing further
increases in life expectancy. Furthermore, while the wealthier countries have now
reached a plateau on the health-wealth curve, these countries have varying health
statuses that Wilkinson and Pickett (Wilkinson & Pickett 2009)) suggest, are related to
the level of income inequality within those countries.
Insight into the reasons behind the varying health statuses of wealthy countries was
provided by Wilkinson (1992; Wilkinson & Pickett 2009). Wilkinson (1992; Wilkinson &
Pickett 2009) showed that countries with relatively equal income distributions are
healthier than those with relatively unequal distributions. Many other studies also
18
support the notion that income inequality is detrimental to population health (Wagstaff
& van Doorslaer 2000; Emerson 2009; Lynch et al. 2004; Kennedy et al. 1996;
Kawachi 2002; Kawachi et al. 1997), including Waldman (1992) who found that infant
mortality is positively related to the income share of the top 5% of the income
distribution.
Figure 1: Correlation between income inequality and the UNICEF index of child wellbeing in 23 rich countries (Adapted from (Pickett & Wilkinson 2007))
CONTEXTUAL VS. COMPOSITIONAL DIFFERENCES
Just as Wilkinson identified health differences between countries, and just as research
has shown that socioeconomic position determines health, recent studies have
demonstrated that differences in health occur and can be identified at the community
level. Macintyre and Ellaway (2003) propose that the variations found between
locations can be explained by compositional differences, where there is a difference in
the kinds of people residing there, or contextual differences, which are attributable to
differences between places. Compositional differences seem to be a reasonable
explanation and if this was the case, ‘poor’ people everywhere would have the same
health status. However, research has shown that this is not entirely the case, with
some studies finding strong evidence that neighbourhood characteristics do impact on
19
health (Macintyre & Ellaway 1999). Therefore, it is reasonable to assume that
contextual differences must be playing some part in explaining health inequalities.
Macintyre, Maciver and Sooman (1993) propose that contextual differences operate
through mechanisms such as the availability and accessibility of health services,
infrastructure, attitudes towards health and health related behaviours and a lack of
social support.
Until recently however, there has been a tendency to examine compositional
differences and focus on individuals, instead of examining area-level or contextual
differences. The primary reason behind this reluctance is the fear of committing the
ecological fallacy, whereby inferences concerning individuals are drawn from
population data (Schwartz 1994). However, the importance of area-level effects has
been recognised and the focus of research has now shifted to area-level effects which
are now more often examined and accounted for than before (Macintyre & Ellaway
2003).
AREA VS. INDIVIDUAL-LEVEL SOCIOECONOMIC MEASURES
To measure and monitor socioeconomic inequalities appropriately, suitable and reliable
measures are critical. Traditionally, administrative health data has been collected for
purposes of health service utilisation and therefore little socioeconomic or demographic
data were collected. As individual-level socioeconomic measures are often not
available, particularly when data are obtained through state or national administrative
data collections, area-level socioeconomic measures are often used (Acheson 1998;
Marmot & McDowall 1986; Turrell & Mathers 2000). These area-level measures are
typically based on data collected during the National Census which, in Australia, is
conducted every five years and is completed by the total population. While these area-
based measures might be the best socioeconomic measure currently available to
researchers and are also relatively easy to access, it has been reported that when
20
individual-level socioeconomic measures are available and compared to area-level
measures, there can be considerable discrepancies between the two levels (Demissie
et al. 2000; Geronimus et al. 1996). Furthermore, when individual-level measures are
not available, the effects of the aggregated area-level measures tend to be
overestimated because we cannot separate individual effects from truly contextual
influences. Therefore caution needs to be taken when interpreting these findings as
contextual effects (Geronimus et al. 1996).
MEASURING SOCIOECONOMIC POSITION
Income, education and occupation are the traditional measures of socioeconomic
position and social gradients persist regardless of which one of these measures is used
(Li et al. 2009). However, these traditional individual-level measures are often
unavailable or difficult to obtain, particularly in routinely collected, population-level
administrative datasets. As an alternative to these traditional measures, composite
measures have now been produced and are often used in research. One set of
composite measures used by Australian researchers are the socioeconomic indexes
for areas (SEIFA), that have been produced by the Australian Bureau of Statistics
(ABS) since 1990 using Census data. Of the indices constructed, the index of relative
socioeconomic disadvantage is most commonly used by Australian researchers
(McCracken 2001). However, there are limitations associated with these composite
measures. One widely reported limitation of the SEIFA is that the indices are only
available at area level, and hence should not be used as proxy measures of individual-
level socioeconomic status (Leonard et al. 2005). An additional limitation of SEIFA is
that it does not provide a true representation of the Aboriginal and Torres Strait
Islander population, as the indices are based on a population that is 97% non-
Aboriginal. Gray & Auld (2000) identify this problem and suggest that these current
measures of socioeconomic disadvantage produced by the ABS are not culturally
appropriate measures. However, as there is currently no other measure available to
21
use for Aboriginal populations, researchers are often left with the decision to use
SEIFA and acknowledge its weakness, or use no measure at all. In an attempt to
overcome this problem the Aboriginal and Torres Strait Islander Commission (ATSIC)
selected four variables that they believe measure socioeconomic disadvantage in
Aboriginal populations. These measures are family income, housing, educational
attainment and level of non-employment (Gray & Auld 2000). Although ATSIC used
these variables to construct an index, and applied it to rank 36 ATSIC regions, Gray
and Auld (2000) recognise that this measure did not capture some important social
conditions and as such, an index of socioeconomic disadvantage that measures social
and economic circumstances more adequately in Aboriginal populations remains to be
refined and tested extensively (Gray & Auld 2000).
RACIAL INEQUALITIES
Racial inequalities in health have been reported in many countries around the world,
including the US, UK, Canada, Australia and New Zealand (Heaman et al. 2005; Craig
et al. 2002; Goldenberg et al. 2008; Ahern et al. 2003; Ananth et al. 2005; Messer et al.
2008; Jatrana & Blakely 2008; Blakely et al. 2005; Blakely et al. 2004; Hill et al. 2007).
The inequalities have been found in a variety of health and developmental outcomes
across the life span. In regards to early life, racial inequalities have been found in infant
mortality (David & Collins 1991; Collins & David 2009; Collins & David 1992;
Freemantle et al. 2006b; Freemantle et al. 2006c), low birthweight (Hancock 2007;
Collins & David 2009) and preterm birth (Gielchinsky et al. 2004; Panaretto et al.
2002a; Messer et al. 2008; Ahern et al. 2003; Heaman et al. 2005; O'Campo et al.
2008; Collins et al. 2007). While it has been well established that minority groups often
experience greater levels of social and economic disadvantage, the factors underlying
racial inequalities in health remain mostly unknown and it is still largely debatable
whether the social and economic inequalities at the individual-level are the sole
contributing factor (Wild & McKeigue 1997; Smaje 1996; Navarro 1990; Sheldon &
22
Parker 1992). It has been suggested that racial inequalities, at least in part, are due to
residential segregation (Massey & Denton 1989; Massey & Denton 1993; Sampson &
Wilson 1995; Hearst et al. 2008; Williams & Collins 2001), racial discrimination
(Dominguez et al. 2008; David & Collins 1991; Collins et al. 2000), neighbourhood
perception and neighbourhood stressors (Zapata et al. 1992; Collins et al. 1998), with
many minority groups living in areas that are considerably worse than where the
mainstream populations reside.
ABORIGINAL HEALTH AND INEQUALITIES
Just over two decades ago De Jonge (1986, as cited in (Choo 1990)) described the
Aboriginal population in Central Australia as being ‘a typical ‘fourth world’ setting where
the vanquished indigenous population looks like ‘third world’ people, but are
surrounded by the material wealth of the modern pioneers’ (p. 34).
While the health status of Aboriginal and Torres Strait Islander people has improved
over the last few decades, they still do not enjoy the same levels of health that non-
Aboriginals do. Furthermore, research suggests that the Aboriginal population does not
display the normal social gradients that are seen in the mainstream population, and
that the health status of Aboriginal children is much worse than that of non-Aboriginal
children (Choo 1990; Thomson 1995). A report examining the health differences
between Aboriginal and non-Aboriginal children showed that for the period, 1998-2002,
the perinatal mortality rate for Aboriginal children was twice that for non-Aboriginal
children, and the hospital separation rate in Western Australia for Aboriginal children
aged 0-3 years was four times higher than that for non-Aboriginal children (Trewin &
Madden 2005). Aboriginal children have higher rates of ear infections, such as otitis
media (Choo 1990; Coates 2002) and the hospital admission rate in 2002-2003 for
infectious diseases in Aboriginal children aged four years or less was more than twice
23
the rate for non-Aboriginal children (Steering Committee for the Review of Government
Service Provision 2003-).
Inequalities between Aboriginal and non-Aboriginal populations in Australia have been
widely reported across a range of health outcomes (Freemantle et al. 2006b;
Freemantle et al. 2006c; Zinn 1997). While some research has suggested that the
differences are, at least in part, due to social disadvantage (AIHW: Leeds KL 2007),
other research has suggested that it is likely to be a result of pathological factors and
greater exposure to other risk factors such as smoking, alcohol and drug use (Blair
1996; Schultz et al. 1991). Findings from the Western Australian Aboriginal Child
Health Survey indicate that high levels of socioeconomic disadvantage occurs within
Aboriginal families, regardless of whether such disadvantage is measured by
education, employment, occupational skill level or income (Zubrick et al. 2004).
THE IMPORTANCE OF INVESTIGATING AND MONITORING SOCIAL AND
RACIAL INEQUALITIES
Despite the fact that increasing social and racial inequalities are now a highly
discussed issue and there is an increasing acknowledgment of the importance of the
early years of life for shaping life-long health and well-being, the research regarding the
impact of inequalities on child health and development in Australia has been fairly
limited. Where this research exists it has tended to solely focus on individual-level
contributing factors for poor health outcomes and has been based on small sample
sizes or conducted in localised study areas. This thesis aims to improve on these
limitations by utilising the linkage of various government administrative databases to
enable the investigation of individual, family and area-level factors. Census data will
also be used to obtain additional area-level information. Given the current limitations in
monitoring changes in social and racial inequalities on an ongoing basis, this thesis will
demonstrate how inequalities in health and development can effectively be monitored
24
and better understood by using routinely collected, population-level data and by
examining contributing factors that are available at the individual and community level.
Therefore the overall aim of this thesis was to examine trends and patterns in birth
rates and investigate social and racial inequalities in poor fetal growth and preterm birth
in Western Australia.
The specific objectives of this thesis are:
1. To describe trends and patterns in birth rates and associated parental
demographic and socioeconomic characteristics in Western Australia between
1984 and 2006 (Chapter Three),
2. To investigate whether the introduction of the Baby Bonus in 2004 was
associated with increased birth rates and if so, determine who was contributing
to that increase (Chapter Four),
3. To develop a measure of individual-level socioeconomic status using routinely
collected parental occupation titles and investigate the differences between
individual and area-level socioeconomic status, as well as its influence on
perinatal outcomes (Chapter Five),
4. To examine the trends in low birthweight due to poor fetal growth and preterm
birth over the last two decades (Chapters Six and Seven),
5. To investigate the social and racial disparities in poor fetal growth and preterm
birth and identify the parental, social and biological factors that may contribute
to these disparities (Chapters Six and Seven), and
6. To investigate the social and racial inequalities in preterm birth among
Aboriginal and non-Aboriginal infants for high-risk (women with a pregnancy
complication and/or pre-existing medical condition) and low-risk women (no
pregnancy complications or pre-existing medical conditions) (Chapters Six and
Seven).
25
This thesis makes several unique contributions. Firstly, it bridges a significant
knowledge gap in understanding trends in social and racial inequalities in birth rates
and perinatal outcomes, over the last two decades in Western Australia. Secondly, it
sheds light on the contributing factors to these inequalities. Thirdly, it overcomes a
major limitation of previous research by developing and incorporating an individual-
level indicator of socioeconomic status in multi-level models. Furthermore, as all
analyses were performed on population-level data using multi-level information, the
results have greater generalisability and policy implications for intervention and
prevention.
26
CHAPTER TWO: GENERAL METHODS
INTRODUCTION
This PhD thesis is a study within the Developmental Pathways Project (DPP) (Stanley
et al. 2004; Glauert et al. 2008). It is unique as it uses population-level, longitudinal,
administrative datasets from the Western Australian Government Departments of
Health, Education, Child Protection and Corrective Services. Unit record data from
each dataset have been linked by the Data Linkage Branch, based within the Western
Australian Health Department. The aim of the PhD thesis was to utilise these
comprehensive linkages of the various longitudinal and population-level health datasets
to investigate social and racial inequalities in birth rates and infant outcomes in
Western Australia over the last 23 years. The study population for this project was
defined as all births in Western Australia from 1980 to 2006, representing a birth cohort
of 669,262 births. The primary datasets used for this study include the Midwives’
Notification System, Hospital Morbidity System and Birth Registrations.
DATA SOURCES
Midwives’ Notification System
The Midwives’ Notification System (MNS) is a register of all births in Western Australia
(WA) and it has been in use since 1975. However, complete birth information is only
available from 1980 onwards (Gee & Dawes 1994). The MNS is a statutory
surveillance system under the WA Health Act, and as such the attending registered
midwife is required by law to complete a Notification of Case Attended (NoCA) form,
providing all required information about the birth, infant and mother (Gee & Dawes
1994).
27
There are approximately 25,000 births annually in Western Australia and the NoCA
form is completed for each birth, live born or stillborn, that is at least 20 weeks
gestation or 400grams birthweight. Information collected on the NoCA form includes
demographic information about the mother, the pregnancy (e.g. maternal medical
conditions and pregnancy complications), the labour and delivery and the infant (e.g.
gender and birthweight). Midwives’ are provided with a copy of guidelines, which are
designed to assist them in the completion of the form. The completed forms are
checked for omissions and possible errors and measures are taken to ensure the
accuracy and completeness of records. Validation studies of the MNS were undertaken
in 1987 and 1994 to provide an indication of the reliability of the data (Gee & Dawes
1994; Hill 1987).
Birth Registrations
Registration of births is a statutory requirement for all births which are 20 weeks or
greater in gestation and is the responsibility of the Western Australian Registrar of
Births, Deaths and Marriages. Birth registrations are generally completed by the
parents and are available for all children born in WA from 1974 onwards. The
information collected on the form includes details on the child, such as year of birth,
sex, place of birth, gestational age, plurality and birthweight. It also includes parental
information, such as date of birth (prior to 1992 parent’s age was collected),
occupation, Aboriginality, place of birth, previous children born and date of marriage.
As registration forms are submitted to the Registrar General following the birth, there is
usually a delay between the actual birth and the submission of the forms, and
occasionally, a delay by the registry to process the form. As a result of this delay, births
at the end of the year can occasionally be registered in the following year, and in some
cases even later (Harper 2006). Hospitals and birthing centres also regularly notify the
Registrar General of births. This process ensures that all births are registered. If the
28
birth registration form is not submitted by the parents within 60 days of the birth, the
Registrar follows up accordingly.
Hospital Morbidity Data System
The hospital morbidity data system includes all admissions to public and private
hospitals in Western Australia from 1970. Each hospital admission/separation is
recorded and the information collected by this system includes: patient
sociodemographic characteristics, such as postcode, gender, age, marital status and
employment status; hospital admission information such as length of stay, source of
referral, admission status, care type, insurance status; and diagnosis and procedure
codes, which are recorded using the International Classification of Disease coding
system (ICD-9 and ICD-10), where up to 20 diagnostic and procedure codes may be
recorded per episode of care.
Census Data
Socio-economic Indexes For Areas
The Socioeconomic Indexes for Areas (SEIFA) are area-based indices produced by the
Australian Bureau of Statistics (ABS) using the National Census1
1 In Australia, the National Census is compulsory and is completed by the total population.
data collected every
five years. The indices have been created since 1990 (using the 1986 census data)
and while several indices have been created following each census, three of those
have been consistent across all census years. They are the Index of Relative
Socioeconomic Disadvantage (IRSD), Index of Economic Resources (IEO) and Index
of Education and Occupation (IEO). The IRSD only distinguishes between relative
disadvantage and a lack thereof, the IER provides an indication of access to economic
resources, and the IEO provides an indication of qualifications and occupations (Pink
2008c). All of these indices are available at a range of levels, with the smallest unit
29
available being a Collection District (CD), containing up to 250 households, with larger
aggregations including postcode, statistical local area, State and National (Pink 2008b).
Each CD is assigned a score for each SEIFA by the ABS using principal components
analysis. Information regarding variables included in each index, as well as the
loadings and weights obtained through principal components analysis have been
described elsewhere2
(McLennan 1990; Castles 1994; McLennan 1998; Trewin 2004;
Pink 2008c). However, as individual data are aggregated to obtain these area-level
estimates, under certain conditions some CDs are excluded from analyses and
therefore are not assigned a SEIFA score. The proportion of CDs not assigned a
SEIFA score varies with each census. These conditions include low populations (usual
resident populations of less than or equal to 10), high-levels of non-response, offshore
(e.g. oil rigs) or shipping areas, five or fewer people employed, or five or less occupied
private dwellings (Pink 2008c). As the raw scores are time-specific and based on
information collected during the census, they are not comparable across time (Pink
2008b). To make the SEIFA comparable across census years, scores are converted
into quintiles based on state or national cut-offs calculated by the ABS (<10%, 10-24%,
25-49%, 50-74%, 75-90%, >90%).
DATA LINKAGE IN WESTERN AUSTRALIA
Data linkage in WA was initially established in 1980, with the creation of the Maternal
and Child Health Research Database (MCHRD) (Stanley et al. 1994). This database
linked data from birth registrations, death registrations and hospital admissions for the
total WA population (Freemantle et al. 2006a). The idea of data linkage was not new,
with the Oxford Record Linkage Study already well underway (Acheson 1964) and
discussions in Australia having first occurred in the 1970’s (Hobbs & McCall 1970).
2 See Appendix A for a list of variables in each index and a brief summary regarding the construction of SEIFA.
30
However, WA was the first Australian state to successfully initiate health data linkage.
While the success has been attributed to several factors (Armstrong & Kricker 1999),
perhaps the most important factors was the collaboration between the WA Health
Department and the University of Western Australia, as well as the substantial capital
funding provided by the WA Lotteries Commission. As a result of that funding and the
ongoing support, the Data Linkage Branch (DLB) was established in 1995 to develop
and maintain the Western Australia Data Linkage System (WADLS), which is a system
of linkages connecting health data from separate sources for individuals in WA
(Holman et al. 1999; Department of Health 2009). When the WADLS was established,
it only included the linking of information from six health datasets. However, it has now
expanded to include more than 20 data sources, including cohort studies and surveys.
The DLB works in collaboration with the University of Western Australia, the Telethon
Institute for Child Health Research and Curtin University of Technology and provides
population-level linked data to researchers, which is usually ‘pseudonymised’3
,
following ethics approvals and a stringent data application process.
The Linkage Process
Data linkage involves bringing records together from different sources that relate to the
same individual. These linkages are created and maintained using rigorous
internationally accepted privacy-sensitive protocols, probabilistic matching and
extensive clerical review (Kelman et al. 2002a). As of May 2009, the WADLS contained
over 20 million records for almost 3.9 million individuals.
3 ‘Re-identifiable or pseudonymised data: Data that have had individual identifiers coded, encrypted or concatenated in a way that renders the identification of an individual difficult but not strictly impossible, are considered to be re-identifiable. The identifier ‘codes’ could conceivably be decoded to re-identify records or, importantly, other fields could be used to narrow down the possible choices. This level of protection is desirable for research as the data are not ‘degraded’ or ‘blurred’ in any way. De-identification is seldom required for research projects and is not a viable option if separate data files are required to be linked.’ Ref: Kelman, Bass & Holman 2002, pg255)
31
The WADLS consists of six core datasets with links within and between these
population health datasets (Birth Registrations, Death Registrations, Hospital Morbidity,
Mental Health, Cancer Notifications, Midwives’ Notifications), some of which span more
than 35 years, as well as marriage records and the electoral roll (Figure 1). Monthly
updates of morbidity, mortality, cancer, midwives’, births and mental health linkages
ensure that the linked information remains current. Each link within the WADLS is
associated with a record in one of the core data sets, and all links in a particular chain
have been associated with the same individual through the process of probabilistic
record linkage. This method of linkage relies on the availability of similar personal
demographic information in each data source, such as name, gender and date of birth
(Newcombe et al. 1992), with identification improving as the number of identifiers used
increases (Roos & Wajda 1991). While personal demographic information from each
of the contributing data sources is imperative to the linkage, to maximise the protection
of individual privacy, all other details from the contributing data source are stored and
managed separately with clinical information never being seen together with the
identifying information (Kelman et al. 2002a).
While the data linkage process is quite complex, there are essentially four steps
involved (Kelman et al. 2002a). Firstly, the data custodians of each dataset provide
electronic files containing personal demographic information from each record to the
DLB. Secondly, using the identifiers, ‘linkers’ at the DLB then use specialist
probabilistic matching software to create linkage keys, providing a unique identifier for
each individual that is consistent across all data sources. Following this, the data are
returned to the respective data custodians with the linkage keys included and the DLB
also retains a copy of the linkage keys (Figure 2). The final step of the linkage process
occurs as approved research projects apply for access to the linked data. Once the
application has been approved by the DLB and the relevant data custodians, the DLB
provides a subset of the linkage keys to the data custodians based on the cohort
requested by the researcher (e.g. all births in WA between 1980 and 2006). The subset
32
of linkage keys is then used by the data custodians to extract the requested clinical
data. Each data custodian provides the researcher with the data from their data source,
so researchers will ultimately receive several data files (i.e. births, deaths, midwives’)
that they link to create the final linked dataset (Figure 3 and 4). Importantly, prior to the
researcher receiving the data, the linkage keys are encrypted so that the each dataset
includes a common ‘nonsense’ variable that allows the researcher to easily link
individuals across the various datasets they receive, while ensuring that they never
receive the linkage key. Additionally, the encryption is unique for each research project.
33
Figure 2: WA Data Linkage System as of April 2009 (Modified (Department of Health) 2009)
34
Linkage without consent
Data linkage is most effective when the total relevant population is included and the
linkage is completed for everyone in that population. The National Health and Medical
Research Council’s (NHMRC) guidelines for researchers and Human Research Ethics
Committees on ethical conduct in research involving humans (National Health and
Medical Research Council 2000; National Health and Medical Research Council 2007),
specifies that it is ethically acceptable to conduct data linkage for epidemiological
research without obtaining individual consent if the researcher can demonstrate that
the public interest in the proposed research is balanced by the public right to privacy.
Additionally, it is generally impractical to obtain consent due to the extended time
period and large sample size (Holman 2001) and the requirement to obtain consent
could lead to participation bias, compromising the generalisability of findings (Dunn et
al. 2004; Young et al. 2001).
Figure 3: The Linking Process - Steps 1, 2 and 3 (Modified (Kelman et al. 2002b))
35
Figure 4: The Linking Process - Step 4 (Modified (Kelman et al. 2002b))
ETHICS APPROVALS
All health datasets (Midwives’ Notification System, Hospital Morbidity Data System,
Birth and Death Registrations and Mental Health Information System) were obtained
via the Data Linkage Branch, subsequent to obtaining ethics approvals from the
Human Research Ethics Committee at the University of Western Australia (#1080), the
Department of Health of Western Australia Human Research Ethics Committee
(#200613) and the Western Australian Aboriginal Health Information and Ethics
Committee (#122-01/06).
36
STATISTICAL METHODS
Statistical analyses used include negative binomial regression and unadjusted and
multilevel multivariate logistic regression models. Multilevel models were applied to
appropriately estimate individual and area-level socioeconomic measures on the
outcome variables. As area-level socioeconomic data are only available at five-year
intervals when the National Census is completed, analyses were conducted in five year
groupings with each group centred around a census year (i.e. 1986: 1984-1988, 1991:
1989-1993, 1996: 1994-1998, 2001: 1999-2003 and 2006: 2004-2006). Analyses were
performed using SAS 9.1 and MLwiN 2.10 and p-values <0.05 were considered
statistically significant. Specific statistical analyses are detailed within each chapter.
37
CHAPTER THREE: BIRTH RATES AND PARENTAL CHARACTERISTICS
INTRODUCTION
This chapter provides a brief overview of birth rates in Western Australia (WA) between
1984 and 2006. It also provides a summary of various maternal and paternal
demographic and social characteristics, highlighting the changes that have occurred
over the last 23 years.
DATA SOURCES
The study population used for this chapter comprised all births in WA between 1984
and 2006 (N=580 767). Record linked data were obtained from the Midwives’
Notification System (MNS) and the Registrar General’s Birth Registrations (BR).
Data analysis
Birth rates were examined by parental social and demographic characteristics, parental
Aboriginality and area-level socioeconomic disadvantage. Parental demographic and
social characteristics examined include age, median age at first birth, Aboriginality,
residential location at the time of birth, area-level level socioeconomic status (SES) at
the time of the birth, marital status (never married, married/de facto, and
separate/widowed/divorced), parity (0, 1, 2 or 3+) and plurality (1, 2 or 3+). Parental
age was divided into six age categories: 15-19, 20-24, 25-29, 30-34, 35-39, 40-44 and
45-49 years. Maternal Aboriginality was obtained from the MNS and verified by the
information contained in the Birth Registrations and Hospital Morbidity records, while
paternal Aboriginality was obtained from Birth Registrations and verified using Hospital
Morbidity records. Residential location was defined as urban or rural, and was based
38
on residential postal areas. Area-level SES for each mother was determined using the
Index of Relative Socioeconomic Disadvantage (IRSD) available at Collection District
(CD) level (Pink 2008b). The IRSD only distinguishes between relative disadvantage
and a lack thereof. As IRSD data are only available at five-year intervals when the
National Census is completed, birth rates and parental characteristics were examined
in five year periods with each period centred around a census year (i.e. 1986: 1984-
1988, 1991: 1989-1993, 1996: 1994-1998, 2001: 1999-2003 and 2006:2004-2006).
Calculation of birth rates
Birth rates were calculated based on the total number of births divided by the total
number of women of childbearing age (15 to 49 years) in the relevant census year.
The total number of childbearing women was obtained from the Australian Bureau of
Statistics (ABS).
39
BIRTH RATES IN WESTERN AUSTRALIA 1984 TO 2006
There was a steady decline in the birth rate between 1984-1988 and 1999-2003. The
greatest decrease occurred between 1984-1988 and 1989-1993, with a reduction of 4
births per 1000 women aged between 15 and 49 years (Figure 5). After 1999-2003 the
trend reversed and the birth rate slightly increased from 51.1 births per 1000 women to
52.5 births per 1000 women in 2004-2006. However, the birth rate in 2004-2006
remained considerably lower than that in 1984-1988 (62.0 births per 1000 women).
Figure 5: Birth rates in Western Australia
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Birt
h ra
te (p
er 1
000
wom
en)
40
BIRTH RATES BY AREA-LEVEL SOCIAL DISADVANTAGE
A typical social gradient can be seen and persisting across the 23 years, with the most
disadvantaged group consistently having the highest birth rate and the least
disadvantaged group consistently having the lowest birth rate (Figure 6). While the
lowest birth rate for the most disadvantaged group occurred in 1999-2003 with 60.9
births per 1000 women, the highest birth rate for the least disadvantaged group was
40.5 births per 1000 women in 2004-2006. The difference in birth rates between the
most and least disadvantaged groups peaked in 1994-1998 at 33 births per 1000
women, which was followed by the smallest difference of 20.9 births per 1000 women
in 1999-2003. The overall birth rate was lowest in 1999-2003.
Figure 6: Birth rates by area-level social disadvantage
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
1989 - 1993 1994 - 1998 1999 - 2003 2004 - 2006
Birt
h ra
te (p
er 1
000
wom
en)
1 (Least Disadv.) 2 3 4 5 6 (Most Disadv.)
41
BIRTH RATES BY MATERNAL AGE GROUP
There has been considerable change in birth rates by maternal age group, particularly
in the last decade. The most noticeable change has been the shift from 25-29 year olds
having the highest birth rate between 1984 and 1998, to 30-34 year olds from 1999 to
2006 (Figure 7). In 2004-2006, 35-39 year olds had the second highest birth rate,
surpassing 20-24 year olds. It is apparent that for the 20-24 year olds there is a
declining trend in birth rates, whereas for women aged 30-34 and 35-39, an increasing
trend can be seen. This reflects a delay of first births in young women aged 20-29.
While 45-49 year olds continue to have the lowest birth rate, it has doubled since 1984-
1988, increasing from 0.3 births per 1000 women to 0.6 births per 1000 women. The
birth rate has also more than doubled for 35-39 year olds and 40-44 year olds in the
last two decades, increasing from 29.9 births per 1000 women to 63.9 births and 5.5
births per 1000 women to 12.8 births, respectively. Additionally, while teenage birth
rates were relatively stable between 1984 and 1998, they have decreased by more
than 15% in recent years, from 20.0 births per 1000 women aged 15-19 years in 1994-
1998 to 16.3 births per 1000 women in 2004-2006.
Figure 7: Birth rates by maternal age group
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Birt
h ra
te (p
er 1
000
wom
en)
15-19 20-24 25-29 30-34 35-39 40-44 45-49
42
BIRTH RATES BY MATERNAL ABORIGINALITY
The birth rate was between 2 and 2.5 times higher for Aboriginal women than non-
Aboriginal women (Figure 8). However, birth rates for both Aboriginal and non-
Aboriginal women have decreased since 1984. Aboriginal women had the greatest
reduction, with the birth rate decreasing by 20% from 143.9 births per 1000 women in
1984-1988 to 111.0 births per 1000 women in 2004-2006.
Figure 8: Birth rates by maternal Aboriginality
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
1984 - 1988 1989 - 1993 1994 - 1998 1999 - 2003 2004 - 2006
Birt
h ra
te (p
er 1
000
wom
en)
Non-Aboriginal Aboriginal
43
PLURALITY
More than 97% of births are singletons, though the proportion of multiple pregnancies
has steadily increased by two-thirds since 1984-1988 (Table 1). Twins account for most
of the multiple births as reflected in both absolute numbers and proportions. This
shows an increasing trend in twin births.
Table 1: Plurality
Plurality 1984-1988 % (N)
1989-1993 % (N)
1994-1998 % (N)
1999-2003 % (N)
2004-2006 % (N)
1 98.8% (118081)
98.6% (125478)
98.5% (125498)
98.3% (123320)
97.9% (79466)
2 1.2% (1443)
1.3% (1690)
1.4% (1833)
1.6% (2023)
2.0% (1660)
3+ 0.0% (48)
0.0% (62)
0.0% (63)
0.0% (54)
0.1% (48)
MATERNAL MEDIAN AGE AT FIRST BIRTH
In 1980-1983 the median age at first birth for mothers was 28 years (Table 2). In 1984-
1988 the median age fell to 25 years and has steadily increased since then until 1999-
2003 when it reached 28 years again.
Table 2: Maternal median age at first birth Median Age at first birth
1980-1983 28
1984-1988 25
1989-1993 26
1994-1998 27
1999-2003 28
2004-2006 28
44
MATERNAL RESIDENTIAL LOCATION AT TIME OF BIRTH
Since 1984-1988, the proportion of births occurring in urban areas has increased from
more than two-thirds of births to almost three-quarters. However, this increase may be
as a result of outer areas previously not considered as urban. For example areas such
as Rockingham and Kwinana were reclassified as urban as the growing WA population
moved further out (Table 3).
PARITY
First births have increased over time, from 38.6% to 41.8%, while second births have
remained relatively stable at around 33%. Third births have continued to decline and
while fourth or higher births increased slightly in 1989-1993, they have slowly
decreased annually by 0.2% (Table 3).
MATERNAL SMOKING DURING PREGNANCY
Smoking data was not collected prior to 1998, though from the data available since
then, smoking during pregnancy has reduced with less than 18% of mothers reporting
smoking during pregnancy in 2004-2006 (Table 3).
45
Table 3: Maternal characteristics 1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Location Urban 68.1% 69.9% 71.8% 73.2% 74.7%
Rural 31.8% 30.0% 28.0% 26.7% 25.1%
Parity First birth 38.6% 39.1% 40.2% 40.8% 41.8%
Second birth 33.7% 32.8% 33.2% 33.4% 33.2%
Third birth 17.9% 17.7% 16.4% 15.7% 15.2%
Fourth or higher birth 9.7% 10.4% 10.2% 10.0% 9.8%
Smoking Missing4 74.1%
Yes 5.9% 20.9% 17.2%
No 20.1% 79.1% 82.8%
PATERNAL AGE
Similar to maternal age, there has also been a slight shift in the paternal age
distribution over time. The greatest decrease occurred in fathers aged 25-29 years,
reducing by more than 40%, while the greatest increase occurred among 40-44 year
olds (138%) (Table 4). Since 1989, most fathers were aged between 30-34 years and
the proportion of teenage fathers has remained relatively stable.
PATERNAL ABORIGINALITY
Births to Aboriginal fathers increased between 1984 and 2003, increasing from 3.5% to
5.3%, but decreased slightly in 2004-2006 to 4.5% (Table 4). While the increase seen
may be real, it is also possible that it was related to a change in social attitudes
towards identifying as Aboriginal.
4 Data collection commenced in 1998
46
Table 4: Paternal Characteristics
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Paternal age group
Unknown 5.4% 5.2% 5.4% 5.4% 6.5% 15-19 1.4% 1.5% 1.3% 1.4% 1.3% 20-24 12.7% 10.7% 9.2% 8.7% 8.1% 25-29 32.4% 27.7% 23.5% 20.7% 18.2% 30-34 29.0% 31.0% 31.3% 30.9% 29.9% 35-39 13.4% 16.2% 19.5% 21.1% 23.0% 40-44 3.9% 5.5% 6.8% 8.4% 9.3% 45-49 1.7% 2.2% 2.9% 3.4% 3.8% Aboriginality Non-Aboriginal 96.5% 95.9% 95.4% 94.7% 95.5% Aboriginal 3.5% 4.1% 4.6% 5.3% 4.5%
PATERNAL MEDIAN AGE AT FIRST BIRTH
In 1980-1983 the median age at first birth for fathers was 31 years (Table 5). In 1984-
1988 the median age fell slightly to 28 years but has steadily increased since then
reaching 31 years in 2004-2006, after temporarily plateauing at 30 years between 1994
and 2003.
Table 5: Paternal median age at first birth
Median Age at first birth
1980-1983 31
1984-1988 28
1989-1993 29
1994-1998 30
1999-2003 30
2004-2006 31
47
MARITAL STATUS AT TIME OF BIRTH
Since 1984 more than 88% of births have been to mothers that are married or de facto,
with that percentage continuing to increase slightly over time (Figure 9). While the
proportion of mothers reporting they were never married was relatively stable between
1984 and 1998, it declined by more than 2% by 2004-2006. The percentage of mothers
reporting they were widowed, divorced or separated has remained relatively stable.
Figure 9: Percentage of births by marital status
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
80.0%
90.0%
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Married (incl De facto) Never Married Widowed/Divorced/Separated Unknown
48
CONCLUSION
The declining birth rate seen in WA between 1984 and 2003 is typical of the trend seen
in many developed countries over the last few decades. In an attempt to reverse the
declining birth rate in Australia, the Federal Government introduced a monetary
incentive in July 2004. Following the introduction of that incentive the WA birth rate did
increase slightly, as can be seen in Figure 5. This raises some important questions:
which socio-demographic group of child bearing mothers contributed to the increase in
birth rates following the introduction of the incentive? Did the incentive attract socially
disadvantaged women at child bearing age and teenagers more than others? These
questions will be addressed in Chapter Four by focusing on WA.
From the social determinants of health perspective, social gradients and inequalities in
birth outcomes are expected. Socially disadvantaged parents/families have greater
exposure to a wide range of risk factors for adverse or suboptimal birth outcomes , as
well as having less resources to ensure a healthy pregnancy and hence a healthy start
to life. Results from this chapter highlight an important socio-demographic factor that
may also contribute to social inequalities in poor infant health, namely the patterning of
birth rates by maternal and paternal social and demographic characteristics, with
parents residing in disadvantaged areas and Aboriginal parents having significantly
higher birth rates. Based on this clear social patterning of birth rates by area-level
disadvantage and by Aboriginality and from the viewpoint of social determinants of
health, we expect to see significant social and racial inequalities in birth outcomes in
WA. Given the critical importance of early life for later health and development, Chapter
Six and Chapter Seven will focus on two key indicators of infant health, namely poor
fetal growth and preterm birth. The analyses in the chapters will investigate in detail
social and racial inequalities in these birth outcomes, contributing factors and possible
changes over time.
49
CHAPTER FOUR: DO MONETARY INCENTIVES INCREASE BIRTH RATES IN SOCIALLY
DISADVANTAGED GROUPS?
INTRODUCTION
Fertility rates are declining around the world and have been doing so for more than 25
years. A recent UNICEF report indicated that between 1970 and 2006 the crude birth
rate fell by 35% in industrialised countries, 39% in developing countries and 34% in the
least developed countries(United Nations Children’s Fund (UNICEF) 2007). While there
are many factors contributing to the decline, several that are often highlighted are
delayed childbearing (Heck et al. 1997), higher educational attainment among women
(Kravdal & Rindfuss 2008) and the increased costs associated with raising children
(Gray et al. 2008a). In response, there has been considerable discussion by
governments, particularly those of industrialised countries with aging populations,
regarding what policy initiatives and strategies could be implemented to increase birth
rates.
Governments around the world have implemented a range of policies and strategies in
attempts to boost fertility rates, including offering monetary incentives, parental leave,
workplace flexibility and more widely available and subsidized child care (Anonymous
2007; Gauthier 2000; Grant et al. 2006; Shaerraden 2001; Dorozynski 2003;
REUTERS 2007). However, the success of these policies has varied across countries
and by the type of strategy implemented (Grant et al. 2004; Duclos et al. 2001; Milligan
2002; Singapore Government & (Family Services Division) 2008). While government
policy can influence birth rates, evidence suggests that social, economic and
demographic factors also impact, either directly or indirectly, on fertility. These factors
include the economic climate, unemployment rates, tax offsets and increases in
women’s educational attainment and labour force participation (d'Addio & d'Ercole
50
2005; Heck et al. 1997; Kravdal & Rindfuss 2008; Rosenbluth & Iversen 2006; Ahn &
Mira 2002).
In May 2004, in an attempt to reverse Australia’s declining fertility rates, the Australian
Federal Government announced the introduction of a fertility policy in the form of a
Maternity Payment. This payment is now known as the Baby Bonus and is paid to the
primary carer, usually the mother, following the birth of a child (including stillbirths and
early neonatal deaths). The payment was provided to the primary carer for births from
July 1st 2004 in recognition of the extra costs associated with having a child, and as
such, was not means-tested. Eligibility requirements included the primary carer and
infant meeting Australian residency requirements, formal registration of the birth
(effective from the 1st July 2007) and the primary carer had to be legally responsible
for the health and welfare of the child (Department of Families Housing Community
Services and Indigenous Affairs 2009a). Additionally, all eligibility requirements had to
be met within the first 13 weeks of the birth and while age did not affect eligibility, in
July 2007, parents under the age of 18 and families deemed to be vulnerable, received
the payment in fortnightly instalments (Department of Families Housing Community
Services and Indigenous Affairs 2009b). The payment was introduced at AUD$3000
(€1500, USD$2000) per baby in July 2004, increased to AUD$4000 (€2000,
USD$2500) in July 2006 and increased again to AUD$5000 (€2500, USD$3200) in
July 2008. The payments were also indexed in line with the Consumer Price Index
twice a year, in March and September.
Following the announcement of the Baby Bonus, there was extensive discussion, and
concerns were raised that such monetary incentives would attract teenagers and
socially disadvantaged groups, hence increasing birth rates amongst high risk groups.
However, to date, there has been no assessment of the effect of the monetary
incentive on birth rates in Australia, and the impact on particular groups, such as
socially disadvantaged groups and teenagers.
51
With the availability of total population level birth data for Western Australia (WA) from
1995 to 2006, the aim of this study was to investigate whether the introduction of the
Baby Bonus was associated with increased birth rates among teenagers and socially
disadvantaged groups.
METHODS
Data Sources
The population used for this study comprised all births in WA between 1995 and 2006
(N = 308 544). For the multivariate regression analyses this was reduced to births
between 2004 and 2006 (N = 81 179), as the required denominator (stratified by
maternal age, Aboriginality, location and community level socioeconomic status) was
not available prior to 2004. Data were obtained from the Midwives’ Notification System
(MNS), a statutory collection of all births in WA, with complete and validated birth
information from 1980 (Gee & Dawes 1994). Data available from this collection
includes demographic information about the mother, pregnancy (e.g. complications and
medical conditions), labour and delivery, and infant (e.g. gender and birthweight).
Independent Variables
Birth Rates were examined in relation to pre and post Baby Bonus years, by various
maternal characteristics, comprising maternal age, Aboriginality, residential location at
the time of birth, and mother’s community level socioeconomic status (SES) at the time
of the birth. Maternal age group was divided into six age categories: 15-19, 20-24, 25-
29, 30-34, 35-39, 40-44 and 45-49 years. Age groups as opposed to a continuous
measure were used as we were particularly interested in whether the teenage birth rate
had increased compared to other age groups. Mother’s Aboriginality was obtained from
the MNS and verified by the information contained in the Birth Registrations, self-
52
reported by parents. Residential location was defined as: major city, inner regional,
outer regional, rural or remote or very remote, and was based on the Remoteness Area
Index (RAI) produced by the Australian Bureau of Statistics (ABS) (Trewin 2006). The
RAI assigned to each mother was based on the location of her household Collection
District (CD) at the time of the birth. CDs are currently the smallest area unit for which
socioeconomic data can be obtained, and contain approximately 250 households,
although this may vary, particularly in more sparsely populated areas (Pink 2008b).
Community level SES for each mother was determined using CD level data available
from the 2006 Socioeconomic Indexes for Areas (SEIFA) (Pink 2008b). SEIFA are four
area-level indices produced by the ABS, based on data obtained from the National
Census, which is conducted every five years. The four indices produced are the Index
of Relative Socioeconomic Advantage and Disadvantage (IRSAD), Index of Relative
Socioeconomic Disadvantage (IRSD), Index of Economic Resources (IER) and Index
of Education and Occupation (IEO). The IRSAD can be used to distinguish between
areas of relative advantage and relative disadvantage, while the IRSD only
distinguishes between relative disadvantage and a lack thereof. The two other indices,
the IER and IEO, provide an indication of the community level access to economic
resources, and the average educational qualifications and occupational status of the
individuals residing in a CD, respectively (Pink 2008c). As with the RAI, the community
level SES assigned to the mother was based on her CD at the time of the birth.
Analyses
The prevalence and trend in the annual birth rate in Western Australia from 1995 to
2006 was assessed based on the annual number of births divided by the total number
of women of childbearing age (15 to 49 years) in the relevant year. The total number of
childbearing women was determined using population data, obtained from the ABS.
The 2006 Census data was used to obtain a further breakdown of childbearing women
53
in WA by the various maternal characteristics, and by the mother’s community level
SES at the time of birth. This enabled birth rates to be assessed by age group, SES,
location and Aboriginality for each year. Additionally, median age at first birth by SES
was examined, and Mantel-Haenszel chi-square (Kuritz et al. 1988) tests were used to
examine trends in parity across years. For example, we examined whether there was
an increase in first births between 1995 and 2006, in second births, in third births and
in fourth or more births.
A segmented negative binomial regression model was used to examine changes in
overall birth rates between 1995 and 2006 (Wagner et al. 2002). This model fits
separate trend lines to the pre and post Baby Bonus years and provides a direct
estimate of the immediate effect of the Baby Bonus. Although introduced in July 2004,
as the Baby Bonus announcement was made in May 2004, so infants born after
February 2005 could have been conceived with the intention of receiving the monetary
incentive. Therefore, 2005 was defined as the first post Baby Bonus year.
A negative binomial regression model was used to investigate the associations
between maternal characteristics and changes in the birth rate after the introduction of
the Baby Bonus (McCullagh & Nelder 1989); (Hilbe 2007). Multivariate regression
analyses were applied to examine the interaction between various maternal
characteristics and the year indicators. As we were investigating the changes in birth
rates for different groups of mothers between the pre and post Baby Bonus years the
interaction effects were the main focus of interest. Likelihood ratio statistics for type 3
analyses were calculated for all interactions, with p-values<0.05 considered as
statistically significant. Goodness-of-fit for each model was assessed using the
deviance and Pearson chi-square statistics, with all models having values within the
expected range. All analyses were performed using SAS 9.1 (SAS Institute Inc. 2002-
2003).
54
All multivariate regression models were adjusted for maternal age group, location, year
and mother’s community SES as measured by the SEIFA. All four SEIFA indicators
were examined, however as the indices were moderately or highly correlated only one
index was included in the model at a time. The IEO was found to be the best predictor
of community level SES and was used in all subsequent analyses. The IEO is a
measure of community SES based on a score calculated from the National census,
and divided into six groups. In this study, group one represents a CD where most of the
people living within that area have higher educational qualifications or are in highly
skilled occupations, and there are few people who have no qualifications or are in low
skilled occupations. Conversely, group six is a CD where many people living within that
area have few qualifications, are unemployed or are in low skilled occupations and also
have less people with higher educational qualifications or in highly skilled occupations
(Trewin 2004).
Ethics
Ethics approval for this study was obtained from the Human Research Ethics
Committee at the University of Western Australia and the Western Australian Aboriginal
Health Information and Ethics Committee. Approval for the use of health data was
obtained from the Confidentiality of Health Information Committee at the Department of
Health, WA.
RESULTS
After an initial period of no change, there was a steady decline in the birth rate between
1998 and 2004 (Figure 10). However, this trend reversed in 2005 when the birth rate
increased from 52.2 births per 1000 women, aged between 15 and 49 years, in 2004 to
55.2 births per 1000 women in 2005 and 58.6 births per 1000 women in 2006. The
segmented regression model estimates that prior to the introduction of the Baby Bonus,
55
the birth rate was declining annually by around 1% (RR=0.991, 95%
CI=0.988,0.993)(Figure 1). However in 2005, following the introduction of the Baby
Bonus, there was a significant 7.4% (p<0.01) increase in births compared to what could
have been expected based on the trend between 1995 and 2004. Additionally, the
difference in births between the pre and post Baby Bonus trends was approximately
3.9%, which was also statistically significant (p<0.01).
Figure 10: Actual birth rates and fitted trend
Between 2004 and 2006 birth rates increased across most maternal characteristics,
and all levels of community SES (Table 6). Prior to adjusting for any maternal
characteristics, women aged 30 to 34 years had the highest birth rate with 131.9 births
per 1000 women in 2006, while women aged 35 to 39 years had the greatest increase
in birth rate, with an absolute increase of 14.7 births per 1000 women from 2004 to
2006. Women aged 15 to 19 years were the only age group to have a slight decline in
their birth rate from 2005 to 2006, and they had the second lowest increase in overall
birth rate, with an increase of 1.1 births per 1000 women between 2004 and 2006.
Aboriginal women had a birth rate twice as high as non-Aboriginal women, and birth
rates by maternal residential location were highest in the remote and very remote areas
and lowest in the inner regional area in 2004, and lowest in the major city area in 2005
46.0
48.0
50.0
52.0
54.0
56.0
58.0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Birt
h R
ate
(per
100
0 w
omen
age
d 15
-49
year
s)
Actual birth rate Fitted trend
56
and 2006. Examining birth rates by community SES showed a gradual increase in the
birth rate across socioeconomic groups, with group one (highest SES) having the
lowest birth rate and group six (lowest SES) having the highest.
Table 6: Socioeconomic and demographic characteristics of mothers who gave birth in 2004 to 2006 2004 Births
N (Birth Rate*)
2005 Births N (Birth Rate*)
2006 Births N (Birth Rate*)
Total Births 25528 26986 28665
Maternal age group
15-19 1119 (16.4) 1215 (17.8) 1193 (17.5)
20-24 3765 (56.9) 3933 (59.4) 4429 (66.9)
25-29 6586 (106.1) 6824 (109.9) 7158 (115.3)
30-34 8611 (124.3) 8913 (128.6) 9142 (131.9)
35-39 4486 (60.1) 5003 (67.0) 5582 (74.8)
40-44 919 (12.2) 1055 (14.0) 1105 (14.6)
45-49 42 (0.6) 43 (0.6) 56 (0.8)
Aboriginality
Non-Aboriginal 23953 (50.5) 25265 (53.3) 26857 (56.6)
Aboriginal 1575 (104.1) 1721 (113.8) 1808 (119.5)
Location^
Major City 17125 (48.5) 18221 (51.6) 19376 (54.9)
Inner Regional 2828 (46.6) 3161 (52.1) 3510 (57.9)
Outer Regional 2255 (54.2) 2454 (59.0) 2592 (62.3)
Remote and Very Remote 1901 (58.4) 2087 (64.1) 2285 (70.2)
Index of Education and Occupation^
1 – Highly educated, highly
skilled occupations 1450(37.6) 1536 (39.9) 1733 (45.0)
2 3367 (41.4) 3420 (42.1) 3739 (46.0)
3 6740 (47.0) 7226 (50.4) 7728 (53.9)
4 5937 (50.0) 6547 (55.1) 6861 (57.8)
5 3668 (53.5) 3929 (57.3) 4328 (63.2)
6 - Few qualifications, low
skilled occupations 2067 (59.8) 2229 (64.4) 2322 (67.1)
* Birth rate per 1000 women aged 15 to 49 years
^ Numbers do not add up to totals due to missing data
57
There was an overall increase in median age at first birth across all socioeconomic
groups during the study period, though there was no significant change in 2005 or 2006
following the introduction of the Baby Bonus. Examination of trends in parity between
1995 and 2006 revealed there was a steady increase in first births (p< 0.01), but a
declining trend in third births (p<0.01) and fourth or more births (p<0.05). Between
2004 and 2006, only second births (p=0.0437) decreased markedly, though there was
a small increase in the percentage of third (14.9% to 15.3%) and fourth or more births
(9.4% to 9.7%).
After adjusting for the various maternal characteristics, including mother’s community
level socioeconomic status, women aged 25 to 29 years had the highest average birth
rate across all years, increasing from 125.1 births per 1000 women in 2004, to 145.4
births per 1000 women in 2006 (Figure 11A). Between 2004 and 2005 women aged 25
to 29 years also had the greatest increase in birth rate (13.5 births per 1000 women),
though they were surpassed by women aged 20 to 24 years in 2005 and 2006 (28.1
births per 1000 women) (Figure 11A). Although there was an increase in birth rates
across all years and among all age groups there were no statistically significant
differences in the birth rates among the various maternal age groups (p=0.98).
Similarly, after adjusting for all factors of interest, Aboriginal women continued to have
a higher birth rate compared to non-Aboriginal women (Figure 11B). Birth rates for both
Aboriginal and non-Aboriginal women increased at a similar rate (Figure 11B) and
there was not a significant difference between the changes in the two groups (p=0.80).
Women living in the major city area had the highest birth rate across all years, with
32.4 births per 1000 women in 2004, increasing to 38.8 births per 1000 women in 2006
(Figure 11C). The greatest increase in birth rates occurred amongst women living in
the outer regional area between 2004 and 2005 (5.9 births per 1000 women), and in
the inner regional area in 2005 and 2006 (4.3 births per 1000 women). While birth rates
58
increased among all maternal residential locations there were no differences between
the groups (p=0.98).
The final model examined birth rates among mothers’ community level socioeconomic
groups (Figure 11D). The greatest increase in births were among women residing in
areas of the highest SES (group one) who had the lowest birth rate in 2004 with 21.5
births per 1000 women, but the highest in 2006 with 38.1 births per 1000 women. In
contrast, women living in areas of the lowest SES (group six) had the smallest increase
in birth rate between 2004 and 2006 with an increase of only 2.9 births per 1000
women. Additionally, group six was the only group whose birth rate declined, with the
birth rate falling slightly from 36.3 births per 1000 women in 2005 to 35.2 births per
1000 women in 2006. However, while there was a general increase in birth rates
across most socioeconomic groups there was no significant overall difference between
groups (p=0.68).
59
Figure 11: A - Adjusted^ birth rate for maternal age group by year; B - Adjusted^ birth rate for Aboriginality by year; C - Adjusted^ birth rate for location by year; D - Adjusted^ birth rate for index of education and occupation by year
^All analyses were adjusted for Aboriginality, maternal age group, location and index of education and occupation.
60
DISCUSSION
This is the first study to investigate the effect of monetary incentives on birth rates at
community level, using total population level data and adjusting for various maternal
characteristics. Compared with 2004 (prior to the Baby Bonus), the WA birth rate
increased significantly in 2005 and again in 2006. Importantly, this increase was not
exclusively among teenagers or socially disadvantaged groups. Rather, there was a
general increase among most maternal age groups, in both Aboriginal and non-
Aboriginal women, in all maternal residential locations and among all community level
socioeconomic groups. Furthermore, there were no changes in median age at first birth
by socioeconomic group, and no significant change in parity after 2004 except for a
slight decline in second births. The women contributing most to the increase were aged
20 to 29 years, and primarily living in areas of highest socioeconomic advantage,
characterised by individuals with higher educational qualifications or in highly skilled
occupations.
The introduction of the Baby Bonus resulted in considerable speculation and
discussion regarding its potential impact on birth rates among teenagers and socially
disadvantaged groups, with strong suggestions that birth rates would increase. This
study does not support those claims, with results showing that the largest increase in
birth rates occurred in the highest socioeconomic group. Evidence from previous
studies lend support to these findings, showing that teenagers and socially
disadvantaged groups often have different motivations for having children, such as
viewing teen motherhood as a meaningful life experience (Frost & Oslak 1999; Merrick
1995), and having limited expectations in regards to continuing education and their
future career (Frost & Oslak 1999). Additionally, while the Baby Bonus may not appear
to be an appealing incentive for many women living in the highest socioeconomic
groups, anecdotal evidence published on online forums indicates that for some women
and families, it contributed to their decision to have a child as it was used as a
61
substitute for paid maternity leave which is largely unavailable in Australia (Various
2008); (Various 2008 ).
A monetary incentive similar to the Baby Bonus, known as the Allowance for Newborn
Children (ANC), was implemented in Quebec in the late 1980’s in an attempt to boost
birth rates. Unlike the Baby Bonus the amount paid to parents varied depending on
birth order, with third and subsequent births receiving a substantially higher payment
(CAN$8000) than the first (CAN$500) or second birth (CAN$1000)(Milligan 2002). But,
as the total number of births in Quebec did not appear to have increased, the ANC
program was deemed ineffective and cancelled in 1997. However, had birth rates been
adjusted for the number of childbearing women in the population, the true effect of the
fertility policy on birth rates and a positive response to the policy initiative would have
been revealed (Duclos et al. 2001; Milligan 2005). Similar to the posthoc findings from
Quebec, our analyses demonstrate an increase in birth rates of women of childbearing
age following the introduction of the Baby Bonus in Australia.
Our finding is also similar to a recently published Australian study from New South
Wales (NSW) (Lain et al. 2009) that showed an increase in their birth rate following the
introduction of the Baby Bonus, though they also found a small increase in teenage
births and third births. One of the major issues with the NSW analysis however, was
the use of a larger statistical local area (SLA) to define geographical location and
broader categories to assign socioeconomic status. Compared with the smaller area
level measure used in our study (CD level), a single SLA can consist of up to 309 CDs.
Additionally, the NSW analysis differed with the use of only two locations (rural or
metropolitan) compared to four used in this study, and they had only three
socioeconomic groups (top 20%, middle 60% and bottom 20%) compared to our six.
Results may have differed due to differences in the demographic characteristics of the
state populations and the timing of the Baby Bonus corresponding to a booming
economic climate in WA, not experienced by NSW.
62
Our results are also validated by a recent survey of approximately 5,000 women
regarding their intentions of having a child following the introduction of the Baby Bonus.
Findings revealed that the intentions of the women surveyed did change following the
introduction of the policy, with most women indicating that they intended on having
another child (Drago et al. 2009). However, findings were limited as there was no
follow-up of behaviours, and intentions are often different to behaviours (Quesnel-
Vallee & Morgan 2003).
The extent to which the increase in the birth rate between 2004 and 2006 can be
attributed to the Baby Bonus is difficult to determine. While there has also been an
increase in the crude birth rate across Australia since the introduction of the Baby
Bonus (Pink 2008a), the Australian Productivity Commission (APC) has stated that it
believes the contribution of the policy to the increased birth rate is only moderate, and
that several other factors are likely to have contributed. These factors include the
improved economic climate, increased flexibility in the labour market and births to
women who previously delayed childbearing (Lattimore & Pobke 2008). While births to
women who previously delayed childbearing may partially explain higher birth rates
among 25 to 34 year old women, it does not explain the increase in the birth rate
observed among women aged 20 to 24. Possible explanations might be a change in
women’s attitudes about childbearing, with more tertiary educated women now
deciding to have children at an earlier age (Vere 2007), or alternatively that the Baby
Bonus policy has encouraged women to have children at slightly younger ages, while
the incentive is available.
The strengths of this study are that the study population comprised all births in WA
between 1995 and 2006, and small area level analyses were conducted at CD level.
Using population level data ensured that all births in WA were included in the analyses,
and using administrative data resulted in less than 8% of births having missing
63
information. Availability of smaller CD level data is a particular strength as SLA or other
larger area measures are often used to represent communities. These larger
aggregations may mask considerable SES variation among CDs and the individuals
living within them. There is also the potential for misclassification at higher
aggregations, with an Australian study finding almost 50% of residents are
misclassified at postal code level compared to CD level (Hyndman et al. 1995). In the
2006 Census, there were 4370 CDs in WA, compared to 249 SLAs, the next smallest
area of analysis. Access to CD level data allowed small area level analysis to examine
the influence of community SES on the Baby Bonus.
As data were obtained through administrative datasets, there was a lack of individual
level SES information available. Therefore, CD level information was used to determine
the mother’s community level SES, as it was the smallest area level measure of SES
available. A limitation of community level SES is that while the within-in group
differences for groups one and six tend to be quite small, there is much more diversity
in groups two, three, four and five (Trewin 2003). Additionally, while we were able to
obtain data on all mothers who gave birth in WA, we were not able to obtain any
information regarding which mothers applied for the Baby Bonus. Consequently, it was
not possible to examine whether there were any differences between women who
claimed the Baby Bonus, and those who did not. However given that information
pertaining to the Baby Bonus was provided at the hospital following the birth of the
child, it is highly likely that all eligible families had access to relevant information and
documentation. Additionally, as the announcement of the Baby Bonus was made in
May 2004, most births in January and early February 2005 could not have been
conceived with the intention of receiving the Baby Bonus. However, as year was the
only variable available for identifying the time of birth, some births have been included
that may not have been conceived with the intention of receiving the Baby Bonus.
Finally, it is possible that the improved economic climate in WA may partially account
for the increase in the birth rate over recent years. For almost a decade, WA has
64
experienced continued economic growth as a result of the mining boom that has
significantly reduced unemployment rates, and increased household disposable
income (Howard & King 2008; Garton 2008). This in turn may have contributed to the
increased birth rates in WA. Unfortunately, this study was unable to disentangle the
effects of the resources boom from that of the Baby Bonus on birth rates in WA.
In conclusion, results have shown that there has been a significant increase in the birth
rate in WA between 2004 and 2006 that appears to be associated with the introduction
of the Baby Bonus policy. However, our findings highlight that the increase in the birth
rate among teenage women and socially disadvantaged groups is not different to that
for other women in WA. Future studies with longer follow up periods are required to
investigate birth rates and confirm the trends and findings of this study. It would also be
of interest to investigate the differences between socioeconomic groups, and
encourage more Australian states to undertake similar analyses at the CD level to
compare findings. As the findings from our study are different to those from the study
conducted in NSW, it would suggest that the Government’s interventions to raise
fertility did not have a uniform impact in all states. Additionally, it indicates that the
impact of the policy was affected by other factors, including the economic climate and
demographic characteristics of the state’s population, which may be important factors
to consider for any future fertility policy. Finally, given the current global financial crisis
and the Australian Government’s announcement to introduce means testing for the
Baby Bonus policy from January 1st 2009 (Commonwealth of Australia 2008), it would
be of considerable interest to conduct a follow up study in the next few years to
examine the effect this has on the birth rate in the future.
65
CHAPTER FIVE: IMPROVING THE MONITORING OF HEALTH INEQUALITIES USING ROUTINELY
COLLECTED DATA TO OBTAIN AN INDIVIDUAL-LEVEL SOCIOECONOMIC INDICATOR
INTRODUCTION
To measure and monitor socioeconomic inequalities appropriately, suitable and reliable
indicators of socioeconomic status (SES) are required. Area-based socioeconomic
measures are often used as proxies for individual-level socioeconomic measures due
to a lack of individual-level information, particularly when data is obtained through state
or national data collections (Acheson 1998; Marmot & McDowall 1986; Turrell &
Mathers 2000). These area-based measures are typically derived from data collected
during the National Census which, in Australia, is conducted every five years and is
completed by the total population. In other countries, such as the UK, the census is
only conducted every ten years, and in the US and Canada, only one in seven and one
in five people, respectively, complete the full census. While these area-based
measures are the best socioeconomic measure currently available to researchers and
are also relatively easy to access, recent research from Canada and US has shown
considerable discrepancies between area and individual-level socioeconomic
measures (Demissie et al. 2000; Geronimus et al. 1996).
In Australia, Collection Districts (CDs) are currently the smallest area unit for which
socioeconomic data can be obtained, and contain approximately 250 households,
although this may vary, particularly in more sparsely populated areas (Pink 2008b). In
the 2006 Census, there were 4370 CDs in Western Australia, compared to 249
Statistical Local Areas (SLAs), the next smallest area of analysis. In using area-level
socioeconomic data only, results must be interpreted carefully to ensure that the
ecological fallacy is not committed, whereby inferences concerning individuals are
66
drawn from the aggregated data (Robinson 1950; Schwartz 1994). Additionally,
Geronimus et al (Geronimus et al. 1996) state that caution needs to be taken when
interpreting and referring to findings as contextual effects when individual-level
measures are not available, as the aggregated area-level measure may be confounded
with individual effects.
While previous studies have examined differences between individual and area-level
socioeconomic measures, most have used small sample sizes and have relied on
individual-level measures obtained from survey data (Demissie et al. 2000; Geronimus
et al. 1996; Walker & Becker 2005). This study overcomes these limitations by
investigating differences between individual and area-level measures using total-
population level data from administrative databases.
When individual-level socioeconomic measures are available for use, particularly in
child health research, they are often related to the parents. Traditionally, when
available, paternal socioeconomic status (SES) has been used to represent individual
and family-level SES (Goldthorpe 1983; Macintyre et al. 2003). However, with the
increasing number of women in paid employment and the increasing number of single-
parent families, the need to also examine maternal SES is becoming increasingly
important. Parental educational attainment and occupation are the most commonly
used measures to define individual-level SES (Genereux et al. 2008; Morrison et al.
1989a; Kalff et al. 2001), though there is still some debate as to whether maternal SES,
paternal SES or both should be used (Korupp et al. 2002; Baxter 1994; Dibben et al.
2006; Marks 2008).
Using population level data available for all births between 1980 and 2006 in Western
Australia, the aims of this study were: 1) Recode parental occupation titles and develop
a measure of individual-level socioeconomic status using the routinely collected
information from Birth Registrations; 2) Examine parental skill distribution over time,
67
stratifying by gender; and, 3) Investigate the influence of both individual-level and area-
level socioeconomic indicators on poor fetal growth, using the measure of individual-
level socioeconomic status developed from parental occupation titles.
METHODS
Data Sources
The population used for this study comprised all parents of infants born in Western
Australia (WA) between 1980 and 2006. This comprised 669 262 births to 326 579
mothers and 346 484 fathers. Data were primarily obtained from the Registrar
General’s Birth Registrations (BR) which is a statutory collection of all births and
includes limited information about the infant (birthweight, gestational age and gender),
as well as parental information such as maternal and paternal country of birth,
occupation, age and Aboriginality. Maternal residential location at the time of birth
(used to assign area-level socioeconomic status) was obtained from the Midwives’
Notification System (MNS). The MNS is also a statutory collection of all births in WA,
with complete birth information from 1980 onwards (Gee & Dawes 1994), and data
includes demographic information about the mother, the pregnancy (e.g. complications
and medical conditions), the labour and delivery, and the infant (e.g. gender and
birthweight). While the BR form is completed by the parents following the birth, the
MNS form is completed by the attending midwife at the time of birth. The BR and MNS
data were linked by Data Linkage WA, located within the Department of Health of WA.
Data are linked via probabilistic matching, using identifiers such as full name, address,
date of birth and gender that are common to all data sets (Kelman et al. 2002a 1033).
Only de-identified data are provided to researchers.
68
Genealogical Data
Genealogical data was provided through the Western Australian Family Connections
Genealogical project, which is also linked to the MNS and BR data by Data Linkage
WA. Genealogy is determined using information available from birth, death and
marriage registrations. Additional health datasets such as the MNS are also used to
identify and confirm relationships between parents and children (Glasson et al. 2008). It
has been estimated that approximately 6% of the population do not have a
genealogical link due to incorrect or a lack of information on birth registrations, or not
appearing in other health datasets. There is currently no information relating to
adoptions or divorce, and there is no way to confirm that the father listed on the birth
certificate is the biological father.
Parental Occupation
Maternal and paternal occupation titles are provided by parents on the BR form. These
titles are provided to researchers as string variables. There are no standard guidelines
provided to parents in the reporting of job titles, so there is little consistency in the
occupations recorded. As a result, some titles that are provided are generic (e.g.
public servant, unpaid worker in family business), while others are vague and as such,
uninformative (e.g. giver, watcher).
Between 1980 and 2005 there were 23 776 parental occupation titles, including titles
that contained various descriptions and spellings of the same occupation. The first
stage of this study involved the recoding of occupation titles to a standard measure
using the Australian and New Zealand Standard Classification of Occupations
(ANZSCO) (Trewin & Pink 2006). This classification system assigns a four-digit
occupation code based on skill level and skill specialisation, including formal education,
experience or on-the-job training that would typically be required to perform that job.
Several additional categories were created, including self-employed, unemployed and
69
volunteers, resulting in a total of 387 groups. Using these groupings an ANZSCO skill
level was then assigned to each parent (Table 7). All occupation titles that could not be
assigned a skill level because they were missing, had inadequate information (i.e.
giver) or were unpaid occupations not associated with a particular skill level (e.g.
housewife), were assigned to an unknown group and analysed separately. Maternal
and paternal occupations were also assessed and coded separately. A complete list of
occupation groups and the associated skill levels is provided in Appendix B.
Table 7: ANZSCO skill levels
ANZSCO skill level
Associated education or work experience
1
Bachelor degree or higher qualification. At least five years of relevant
experience may substitute for the formal qualification. In some
instances relevant experience and/or on-the-job training may be
required in addition to the formal qualification.
2
Associate Degree, Advanced Diploma or Diploma. At least three years
of relevant experience may substitute for the formal qualifications. In
some instances relevant experience and/or on-the-job training may be
required in addition to the formal qualification.
3
Certificate IV or Certificate III including at least two years of on-the-job
training. At least three years of relevant experience may substitute for
the formal qualifications. In some instances relevant experience and/or
on-the-job training may be required in addition to the formal
qualification.
4
Certificate II or III. At least one year of relevant experience may
substitute for the formal qualifications listed above. In some instances
relevant experience may be required in addition to the formal
qualification.
5
Certificate I or compulsory secondary education. For some occupations
a short period of on-the-job training may be required in addition to or
instead of the formal qualification. In some instances, no formal
qualification or on-the-job training may be required.
99 No assigned skill level due to missing or inadequate information
supplied
(Trewin & Pink 2006)
70
Validation of Occupation Codes
Recoding of occupation titles was validated by three independent reviewers. Each
reviewer was provided with a random sample of 1000 occupation titles and a list of
occupation categories. They were blinded to the original coding. Inconsistencies
between reviewers were minimal, with only 5% of occupation titles recoded differently.
Differences in recoding were resolved following discussions with all reviewers.
Area-level Socioeconomic Status
Area-level socioeconomic status (SES) for each mother was determined using
Collection District (CD) data available for the Socioeconomic Indexes for Areas
(SEIFA). SEIFA are four area-level indices produced by the ABS since 1986 and based
on data obtained from the National Census, conducted every five years. The index
used in this study was the Index of Relative Socioeconomic Disadvantage (IRSD),
which identifies relative disadvantage, or a lack thereof, in communities (Pink 2008b).
To make the IRSD comparable across census years, scores were converted into
quintiles based on state or national cut-offs calculated by the ABS (<10%: least
disadvantaged, 10-24%, 25-49%, 50-74%, 75-90%, >90%: most disadvantaged). As
9.5% of all births had missing CD information, this IRSD group was assigned as
unknown.
Poor fetal growth
The percentage of optimal birth weight (POBW) (Blair et al. 2005) was used to assess
fetal growth restriction. POBW was used as it is considered to be a more appropriate
measure of fetal growth, as it accounts for important factors that influence birthweight
such as gestational age, maternal height, parity and infant gender. POBW was
calculated by dividing the observed birthweight by the optimal birthweight, where the
optimal birthweight was determined using information available from the MNS,
71
adjusting for the factors listed above. For our analyses poor fetal growth was defined
as having a POBW of 87% or less, corresponding to less than the 10th
percentile for
the population (Blair et al. 2005).
Statistical Analyses
Initially, differences in maternal and paternal skill distribution by census year were
investigated. Comparisons were then made of the maternal/paternal skill level recorded
at first birth and the highest maternal/paternal skill level recorded across all births. As
genealogical data linking fathers with children was not available in 2006, the cross-
tabulations of skill level across births only included births from 1980 to 2005.
Frequencies were calculated for each census year and for maternal and paternal skill
level by IRSD group and maternal skill level by paternal skill level. For some
frequencies data was restricted to births from 1984 onwards due to IRSD data not
being available prior to this date.
Logistic regression was used to investigate whether parental skill level, as a measure
of individual-level SES, was a significant predictor of poor fetal growth independent of
area-level SES. These analyses included all live singleton, term (>=37 weeks) births in
WA between 1999 and 2006. As we were also interested in possible interactions
between individual and area-level SES, the interaction effects of maternal and paternal
skill level with area-level social disadvantage were also examined. Collinearity between
maternal and paternal skill level was examined prior to undertaking the regression
analyses. In the final model, parental age and marital status were included as
covariates to investigate whether the effects of individual or area-level SES changed
after adjusting for known risk factors. All variance inflation factors were less than 2.5
indicating that collinearity was not an issue (Allison 1999). The Likelihood Ratio Chi-
square test was used to assess improvements in model fit. All analyses were
72
performed using SAS 9.1 and MLwiN Beta 2.0, and p-values<0.05 were considered
statistically significant.
Ethics
Ethics approval for this study was obtained from the Human Research Ethics
Committee at the University of Western Australia and the Western Australian Aboriginal
Health Information and Ethics Committee. Approval for the use of health data was
obtained from the Confidentiality of Health Information Committee at the Department of
Health, WA.
RESULTS
There has been a noticeable change in the occupations reported by mothers on birth
registrations in WA over the last 26 years. Between 1980 and 1983 most mothers did
not report an occupation (99%), however by 1984 to 1988 the proportion had declined
to just over a third, with housewife being the second most recorded occupation (data
not shown). From 1989 housewife was the single most recorded maternal occupation
group, though it has reduced substantially from 44% in 1989-1993 to 14.6% in 2004-
2006. For fathers, the majority had a reported occupation, with only 5% missing from
1989 onwards. Prior to 1994, the most common occupation for fathers recorded was
miscellaneous labourer, but after 1994 ‘self-employed’ was the most frequently
reported occupation.
Figures 12 and 13 show the distribution of maternal and paternal skill level by census
year. It can be clearly seen that the proportion of mothers with no assigned skill level
has considerably decreased over time, from 80% in 1984-1988 to 24% in 2004-2006.
Over the same period, the proportion of mothers with a bachelor degree or higher
increased from 7% to 26%, which was the largest overall increase across all skill
levels. Similarly, the proportion of fathers with a bachelor degree or higher also
73
increased over time, from 23% in 1984-1988 to 27% in 2004-2006. While the
proportion of fathers with no assigned skill level varied only slightly over time (11% to
14%), the proportion of fathers with a Certificate II or III, or a Certificate IV or Certificate
III and 2 years of experience steadily declined between 1984-1988 and 2004-2006,
from 27% to 24% and 18% to 14%, respectively.
Figure 12: Maternal skill level distribution
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
80.0%
90.0%
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Bachelor degree or higher Assoc. degree, advanced diploma or diplomaCertificate IV or Certificate III & 2 yrs experience Certificate II or IIICertificate I or compulsary secondary education No assigned skill level
Figure 13: Paternal skill level distribution
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Bachelor degree or higher Assoc. degree, advanced diploma or diploma
Certificate IV or Certificate III & 2 yrs experience Certificate II or III
Certificate I or compulsary secondary education No assigned skill level
74
A cross-tabulation of maternal and paternal skill levels revealed that the proportion of
mothers with no assigned skill level increased with decreasing paternal skill level.
Almost 36% of mothers had no assigned skill level when the father had a bachelor
degree or higher, which increased to 70% when the father also had no assigned skill
level (Table 8). While there was not a strong association or correlation overall between
maternal and paternal skill levels (r=0.12), fathers with a bachelor degree or higher
were most likely to have a child with a mother with the same skill level (37%).
Table 8: Paternal skill level by maternal skill level
All births (1980-2006)
Paternal skill level Bachelor degree
or higher
Associate degree,
advanced diploma
or diploma,
Certificate IV, or
Certificate III and 2 years of
on-the-job training.
Certificate II or III.
Certificate I or
compulsory second
education.
No assigned skill level
Mat
erna
l ski
ll le
vel
Bachelor degree or
higher 36.9% 18.4% 12.3% 11.0% 8.9% 8.5%
Associate degree,
advanced diploma or diploma,
5.0% 15.0% 4.3% 3.7% 3.4% 2.3%
Certificate IV, or
Certificate III and 2 years of
on-the-job training.
6.2% 6.8% 8.9% 6.5% 5.8% 3.8%
Certificate II or III. 11.5% 15.0% 16.5% 17.8% 14.0% 9.2%
Certificate I or
compulsory second
education.
4.1% 7.0% 9.9% 9.8% 12.9% 6.5%
No assigned skill level
36.3% 37.8% 48.2% 51.3% 55.0% 69.8%
75
Examination of the skill level recorded at first birth by the highest skill level recorded
across all births highlighted the internal consistency of maternal and paternal skill level
(Table 9 and 10, respectively). Additionally, the skill level reported at the first birth was
the highest skill level reported across all births for more than 85% of mothers and 70%
of fathers. For parents with no assigned skill level at the first birth, 14% of mothers and
almost 18% of fathers had an assigned skill level in subsequent births. Further
investigation revealed that occasionally skill level decreased across births or changed
to an unassigned skill level. When this occurred, it was usually as a result of a change
in occupation to student or housewife (data not shown).
Table 9: Maternal skill level at first birth compared to highest skill level reported across all births
All births (1980-2005)
Maternal skill level recorded at first birth Bachelor degree
or higher
Associate degree,
advanced diploma
or diploma,
Certificate IV, or
Certificate III and 2 years of
on-the-job training.
Certificate II or III.
Certificate I or
compulsory second
education.
No assigned skill level
Hig
hest
mat
erna
l ski
ll le
vel
reco
rded
acr
oss
all b
irths
Bachelor degree or
higher 100.0% 5.4% 2.8% 3.4% 2.1% 4.3%
Associate degree,
advanced diploma or diploma,
0.0% 94.6% 2.9% 2.6% 2.3% 1.5%
Certificate IV, or
Certificate III and 2
years of on-the-job training.
0.0% 0.0% 94.4% 2.1% 1.6% 1.6%
Certificate II or III. 0.0% 0.0% 0.0% 91.9% 5.9% 3.7%
Certificate I or
compulsory second
education.
0.0% 0.0% 0.0% 0.0% 88.1% 2.8%
No assigned skill level
0.0% 0.0% 0.0% 0.0% 0.0% 86.0%
76
Table 10: Paternal skill level at first birth compared to highest skill level reported across all births
All births 1980-2005
Paternal skill level recorded at first birth Bachelor degree or
higher
Associate degree,
advanced diploma
or diploma,
Certificate IV, or
Certificate III and 2 years of
on-the-job training.
Certificate II or III.
Certificate I or
compulsory second
education.
No assigned skill level
Hig
hest
pat
erna
l ski
ll le
vel r
ecor
ded
acro
ss a
ll bi
rths
Bachelor degree or
higher 100.0% 10.5% 3.9% 6.7% 5.5% 4.6%
Associate degree,
advanced diploma or diploma,
0.0% 89.5% 3.4% 4.5% 4.4% 1.7%
Certificate IV, or
Certificate III and 2
years of on-the-job training.
0.0% 0.0% 92.7% 5.1% 8.8% 3.7%
Certificate II or III. 0.0% 0.0% 0.0% 83.7% 11.2% 3.7%
Certificate I or
compulsory second
education.
0.0% 0.0% 0.0% 0.0% 70.1% 4.0%
No assigned skill level 0.0% 0.0% 0.0% 0.0% 0.0% 82.2%
Investigation of parental skill level by area level SES showed similar results for mothers
and fathers (Table 11). There was a considerably higher proportion of parents with a
bachelor degree or higher living in the least disadvantaged areas (34% Mothers; 49%
Fathers), with the proportion decreasing with increasing level of disadvantage (Table
11). A similar trend was seen with skill level two, Associate degree, Advanced diploma
or Diploma, but the reverse trend was seen with skill level five, Certificate I or
Compulsory secondary education. The reverse trend was also seen quite strongly in
parents with no assigned skill level. The breakdown of the unknown area-level SES by
skill level revealed that the largest proportion of women with an unknown area-level
SES are those with no assigned skill level (55%), followed by women with a Bachelor
degree or higher (18%). For fathers in the unknown area-level SES group, the largest
77
proportion had a Bachelor degree or higher (36%), with Certificate IV or Certificate III
and 2 years of experience being the next highest (19%).
Table 11: Maternal and paternal skill level by area level social disadvantage
All births (1984-2006) IRSD 1 2 3 4 5 6 Unknown
Maternal skill level Bachelor degree or
higher 33.9 24.8 19.4 14.6 11.0 7.0 18.2
Associate degree, advanced diploma or
diploma, 6.1 6.1 5.7 5.1 4.2 2.7 4.5
Certificate IV, or Certificate III and 2 years of on-the-job
training.
6.8 7.4 7.6 6.9 6.3 4.6 5.0
Certificate II or III. 11.9 14.7 16.2 15.6 13.9 10.7 10.5 Certificate I or
compulsory second education.
3.3 5.8 7.9 9.7 10.6 9.8 6.9
No assigned skill level 37.9 41.2 43.2 48.2 54.0 65.3 54.9 Paternal skill level
Bachelor degree or higher 49.1 33.6 24.5 18.2 13.3 8.6 36.0
Associate degree, advanced diploma or
diploma, 12.7 12.7 10.8 9.2 7.6 5.2 6.9
Certificate IV, or Certificate III and 2 years of on-the-job
training.
16.1 24.3 28.7 29.4 28.7 23.6 19.2
Certificate II or III. 9.8 14.2 16.9 18.8 18.5 16.0 12.8 Certificate I or
compulsory second education.
5.4 7.8 10.2 12.7 15.2 18.6 12.5
No assigned skill level 6.8 7.5 8.9 11.8 16.8 28.0 12.7
Findings from the logistic regression showed a clear social gradient in poor fetal growth
for maternal and paternal skill level, as well as area-level SES. The odds of having an
infant with poor fetal growth increased with decreasing skill level and increasing area-
level disadvantage (Table 12, 13). The gradient for paternal skill level was noticeably
steeper than that for maternal skill level. Additionally, for both maternal and paternal
skill level, those with no assigned skill level were considerably more likely to have an
infant with poor fetal growth than any other skill level. For parents with an unknown
area-level SES the odds were more similar to group five. Type 3 analysis of effects
78
Table 12: Unadjusted Odds Ratios predicting poor fetal growth (Births: 1999 – 2006)
Variable Unadjusted OR (95% CI)
Index of Relative Socioeconomic Disadvantage (IRSD)
1 (Least Disadvantaged) Ref
2 1.08 (1.02, 1.15)**
3 1.16 (1.10, 1.23)***
4 1.27 (1.20, 1.35)***
5 1.49 (1.41, 1.58)***
6 (Most Disadvantaged) 2.00 (1.88, 2.13)***
Unknown 1.47 (1.38, 1.57)***
Maternal skill level
Bachelor degree or higher Ref
Associate degree, advanced diploma or diploma, 1.06 (1.01, 1.12)*
Certificate IV, or Certificate III and 2 years of on-the-job training. 1.19 (1.13, 1.25)***
Certificate II or III. 1.24 (1.19, 1.28)***
Certificate I or compulsory second education. 1.31 (1.25, 1.37)***
No assigned skill level 1.76 (1.70, 1.82)***
Paternal skill level
Bachelor degree or higher Ref
Associate degree, advanced diploma or diploma, 1.11 (1.05, 1.16)***
Certificate IV, or Certificate III and 2 years of on-the-job training. 1.26 (1.21, 1.30)***
Certificate II or III. 1.30 (1.25, 1.36)***
Certificate I or compulsory second education. 1.45 (1.38, 1.51)***
No assigned skill level 2.07 (1.99, 2.16)***
79
Table 13: Maternal and paternal skill level and area-level social disadvantage predicting poor fetal growth (Births: 1999 – 2006)
Variable Model 15
OR (95% CI) Model 26
OR (95% CI) Model 37
OR (95% CI) Model 48
OR (95% CI)
Index of Relative Socioeconomic Disadvantage (IRSD)
1 (Least Disadvantaged) Ref Ref Ref Ref
2 1.08 (1.02, 1.15)** 1.09 (1.02, 1.16)** 1.05 (0.98, 1.11)
3 1.16 (1.10, 1.23)*** 1.16 (1.09, 1.22)*** 1.08 (1.01, 1.14)*
4 1.27 (1.20, 1.35)*** 1.23 (1.16, 1.30)*** 1.11 (1.04, 1.17)**
5 1.49 (1.41, 1.58)*** 1.40 (1.32, 1.49)*** 1.22 (1.14, 1.29)***
6 (Most Disadvantaged) 2.00 (1.88, 2.13)*** 1.73 (1.62, 1.85)*** 1.44 (1.35, 1.55)***
Unknown 1.47 (1.38, 1.57)*** 1.33 (1.24, 1.42)*** 1.22 (1.14, 1.30)***
Maternal skill level
Bachelor degree or higher Ref Ref
Associate degree, advanced diploma or diploma, 1.03 (0.97, 1.09)
Certificate IV, or Certificate III and 2 years of on-the-job training. 1.12 (1.05, 1.18)***
Certificate II or III. 1.15 (1.10, 1.20)***
Certificate I or compulsory second education. 1.17 (1.11, 1.23)***
No assigned skill level 1.38 (1.32, 1.43)***
Type 3 Analysis of Effects –(Interaction with IRSD) 46.03*
Paternal skill level
Bachelor degree or higher Ref Ref
5 Model 1 – IRSD only 6 Model 2 – IRSD, maternal and paternal age and marital status (main effects only) 7 Model 3 – IRSD, maternal and paternal age, marital status and maternal and paternal skill level (main effects only) 8 Model 4 – IRSD, maternal and paternal age, marital status and maternal and paternal skill level (incl. interactions between skill level and IRSD)
80
Associate degree, advanced diploma or diploma, 1.07 (1.02, 1.13)**
Certificate IV, or Certificate III and 2 years of on-the-job training. 1.16 (1.12, 1.21)***
Certificate II or III. 1.18 (1.13, 1.24)***
Certificate I or compulsory second education. 1.28 (1.22, 1.34)***
No assigned skill level 1.36 (1.29, 1.44)***
Type 3 Analysis of Effects –(Interaction with IRSD) 71.87***
Log Likelihood 488556.00 487967.45 486525.51 486462.39
Log Likelihood Ratio Chi-square Test^9 592.55*** 1441.94*** 93.12***
9 Compares Model 2 with Model 1, Model 3 with Model 2 and Model 4 with Model 3
* p < 0.05 **p <0.01 ***p <0.001
81
indicated that the global interactions between IRSD and maternal and paternal skill
level were significant in Model 3 and Model 4, after accounting for several parental
demographic characteristics. Results from the Likelihood Ratio Chi Square test
indicated that Model 2 was better than Model 1, Model 3 better than Model 2 and
Model 4 was considerably better than Model 3 (Table 13).
Figures 14 and 15 illustrate the interactions between parental skill level and area-level
SES. After adjusting for parental age and marital status, the rate of poor fetal growth
per 1000 infants increased with increasing levels of area-level disadvantage across all
paternal skill levels (Figure 14). Infants born to fathers with a bachelor degree or higher
consistently had the lowest rate of poor fetal growth, with 192.2 growth restricted
infants per 1000 infants born in the least disadvantaged areas and 244.8 growth
restricted infants per 1000 infants born in the most disadvantaged areas. The highest
rate of poor fetal growth typically occurred to infants born to fathers with no assigned
skill level. In the least disadvantaged areas, the rate of poor fetal growth in infants born
to these fathers was 226.1 per 1000 infants, while in the most disadvantaged areas it
was 360.6 per 1000 infants. Infants born to fathers with no assigned skill level who also
had an unknown area-level SES had the highest rate of poor fetal growth with 383.4
per 1000 infants. For all other paternal skill levels, the highest rate of poor fetal growth
occurred in the most disadvantaged area-level group.
82
Figure 14: Adjusted^ rate of poor fetal growth for Father’s skill level by area-level disadvantage
^Adjusted for maternal age, paternal age and marital status
Similarly for mothers, the rate of poor fetal growth per 1000 infants generally increased
with increasing levels of area-level disadvantage for all maternal skill levels (Figure 15).
One noticeable difference however was that mothers with an associate degree,
advanced diploma or diploma, had the lowest rates of poor fetal growth when they also
resided in one of the three area-level groups with lower levels of disadvantage (least
disadvantaged: 182.5 per 1000 infants; Group 2: 191.4 per 1000 infants; Group 3:
204.5 per 1000 infants). In areas with higher levels of disadvantage, mothers with a
bachelor degree or higher had the lowest rates of poor fetal growth (Group 4: 200.4 per
1000 infants; Group 5: 212.2 per 1000 infants; most disadvantaged: 255.0 per 1000
infants). Mothers with no assigned skill level had the highest rate of poor fetal growth in
all area-level SES groups, except for the least disadvantaged group. The highest rate
of poor fetal growth for all mothers, regardless of their skill level, were in the most
disadvantaged area-level group.
83
In essence, Figures 14 and 15 reveal significant interaction effects. While the rate of
poor fetal growth increased with the level of disadvantage for all parental skill levels,
the rate at which it increased was different depending on which skill level was under
consideration.
Figure 15: Adjusted^ rate of poor fetal growth for Mother’s skill level by area-level disadvantage
^Adjusted for maternal age, paternal age and marital status
DISCUSSION
To our knowledge this is the first Australian study to utilise routinely collected health
data to develop an individual measure of socioeconomic status and to assess the
impact and robustness in an applied example. Parental occupation titles reported on
birth registration forms was utilised to obtain a measure of maternal and paternal SES,
84
to investigate how maternal, paternal and area-level independently predict poor fetal
growth. Results show that the recording of maternal occupation has significantly
improved over time, with less than 25% of mothers being unable to be assigned a skill
level in 2004-2006, compared with 80.5% in 1984-1988. While approximately 34% of
mothers and 49% of fathers with a bachelor degree or higher were living in the least
disadvantaged areas, 66% of mothers and 51% of fathers with the highest skill level
lived in more disadvantaged areas. Overall, there is considerable variation in individual
SES within area-level socioeconomic groups. Findings also showed that maternal,
paternal and area-level SES were all important predictors of poor fetal growth. Hence,
these measures should be considered in analyses when available.
These findings are consistent with several previous studies elsewhere which have also
concluded that when both individual and area-level SES information are included in
analyses, both have significant independent effects on the outcome variables of
interest (Luo et al. 2006; Kalff et al. 2001; Shouls et al. 1996; Van Os et al. 2000;
Steenland et al. 2004; Korupp et al. 2002). Findings from this study challenge the
common assumption that area-level SES is a good proxy for individual-level SES, and
highlight considerable variation in maternal and paternal skill level within each area-
level socioeconomic group. This study also reveals the significance of the interaction
between maternal and paternal skill level and area-level disadvantage (IRSD),
indicating that there is a somewhat more complex relationship between them and the
effects of individual and area-level SES are not independent. The relationship between
individual and area-level SES shown in Figures 14 and 15 highlight that regardless of
area-level SES, higher individual-level SES, as measured by skill level, is protective
against poor fetal growth. As the area-level disadvantage increases, the rate of poor
fetal growth also increases for all parental skill levels. This suggests that there is
something unique and detrimental about socioeconomic disadvantage at the area-
level. Therefore, policy and intervention programs need to be directed at both the
individual and the community, to reduce rates of poor fetal growth.
85
An examination of the distribution of maternal and paternal skill levels in the unknown
socioeconomic group also challenges the common belief that the parents in that
socioeconomic group are most likely to be disadvantaged. The results clearly show that
over a third of the fathers and almost a fifth of the mothers living in areas with unknown
area-level SES have a Bachelor degree or higher. An Australian study using data from
a disability survey demonstrated that the individual-level SES measures (family income
and family income adjusted for family size,) were considerably better predictors of
people with disability than the area-level SES measure (Walker & Becker 2005).
Furthermore it has been shown that when area-level SES measures are used, health
inequality estimates are often weaker compared to estimates based on individual-level
SES measures (Walker & Becker 2005; Steenland et al. 2004). However, this
difference is not found in this study, suggesting that whether or not the individual or
area-level SES measures capture the same or different degree of health inequality may
depend on the outcome or the region under investigation.
Findings from this study highlight the importance of including both maternal and
paternal socioeconomic measures, when available. This finding has also been
supported in a previous study by Parker and Schoendorf (Parker & Schoendorf 1992)
which found that separately, neither maternal nor paternal education, sufficiently
represented the full effect of parental education. Traditionally, only paternal SES has
been used to represent family SES due to a lack of maternal SES information, or
because socioeconomic classifications are based on male occupation groups and the
belief that regardless of the mothers occupational status, it is the fathers occupation
that is important for family-level SES (Dibben et al. 2006a; Macintyre et al. 2003;
Goldthorpe 1983). However, it has been acknowledged that paternal SES may not be
the most suitable measure, particularly for some outcomes such as pregnancy
outcomes (Pattenden, 1999). The importance of maternal SES as has been shown in
this study, has also been recognised and this variable is increasingly included in
86
studies examining various infant and child health and developmental outcomes
(Morgen et al. 2008; Parker & Schoendorf 1992; Korupp et al. 2002; Crook 1995). With
the increasing number of mothers in the paid labour force and the rising number of
single parent families (Linacre 2007), it is critically important to be able to adjust for
both maternal and paternal SES in research examining social gradients in child health.
The primary strength of this study is that it provides a relatively easy method for
obtaining maternal and paternal SES measures from routinely collected administrative
data. By using total population level data, a more reliable comparison of individual and
area-level SES was obtained, as well as how maternal and paternal SES compares.
Additionally, the availability of genealogical data was a considerable advantage as it
provided a rare opportunity to look across births to examine changes in maternal and
paternal skill level over time. The use of CDs to represent area-level SES measure is
also a strength, as it is based on small areas containing approximately 250
households, as opposed to the postcodes and larger areas (each containing up to 212
CDs) that are often used to represent communities, but may mask SES variation
among CDs. An additional strength is that the occupation recoding used in this study
has the wider applicability. As the ANZSCO was used to recode occupation titles, the
codes can easily be recoded to represent major occupation groups, such as Managers,
Professionals, Technicians and Trades Workers and Clerical and Administrative
Workers (Trewin & Pink 2006). Furthermore, the ANZSCO occupation codes can be
converted to the International Standard Classification of Occupations so that
international comparisons can be made (Australian Bureau of Statistics 2006).
While the BR data can be used to obtain maternal and paternal skill levels, there are a
few potential limitations. Firstly, the occupation information is only collected at the time
of each birth so any improvements in skill level made after that birth may not be
recorded. However, as our findings have shown over 70% of fathers and 86% of
mothers do not record a higher skill level after their first birth and with the availability of
87
genealogical data skill level for each mother and father can be checked across all
births. Secondly, as coding is based solely on the occupation title provided by the
parents, there is the possibility that the mother or father may actually have a higher skill
level than required for the job. Additionally, while the proportion of mothers reporting
housewife as their occupation has significantly decreased over time there is still almost
15% (data not shown), who therefore can not be assigned a skill level. Lastly, the
unknown skill level may contain a significant proportion of mothers who were not active
in the labour force around the time of the birth of the child, but who in fact had an
occupational skill. For example, if these mothers recorded their occupation as
housewife or mother, they will not be assigned a skill level.
In conclusion, this study demonstrates how routinely collected BR information can be
used to obtain individual-level SES for parents. Furthermore, it highlights the internal
consistency of the measures and clearly demonstrates the importance of including
maternal, paternal and area-level SES measures when examining health inequalities.
By including both individual and area-level SES measures, researchers will be better
positioned to separate contextual effects from those that are due to individual-level
SES. Additionally, with the availability of individual-level SES measures from
population-level data, area-level measures will rarely need to be used as proxies for
individual-level SES, thus minimising the risk of committing the ecological fallacy.
88
CHAPTER SIX: RISING SOCIAL AND RACIAL INEQUALITIES AMONGST INFANTS WITH POOR
FETAL GROWTH IN WESTERN AUSTRALIA, 1984 TO 2006
INTRODUCTION
Low birthweight is an important public health issue, and is a major determinant of
infant mortality and subsequent childhood morbidity. Low birthweight children are at
a greater risk of poorer physical (Boardman et al. 2002) and mental health
outcomes(Bohnert & Breslau), and are more likely to be diagnosed with disabilities
(Leonard et al. 2008). Research has also shown that intrauterine growth may be
important for cognitive and respiratory function (Dubois & Girard 2006; Tong et al.
2006). Additionally, children born with low birthweight have an increased risk of
behavioural problems and learning difficulties(Reichman 2005; Roberts et al. 2007),
and have poorer health outcomes as adults (Barker 1995; Ozanne et al. 2004).
Barker (Barker 1995) suggests that if a foetus receives a limited supply of nutrients,
then it will adapt, resulting in permanent changes to its structure and metabolism.
According to Barker’s hypothesis, it is these changes that are responsible for
increasing the risk of chronic disease later in life, such as cardiovascular disease,
diabetes, obesity and hypertension (Barker 1998; Rich-Edwards et al. 1997).
Birth weight is influenced by both social and biological factors including maternal
age, maternal ethnicity, parity, marital status, maternal infections, smoking during
pregnancy, pre-existing medical conditions and pregnancy complications (Corchia et
al. 1995; Mohsin et al. 2003; Kramer 1987; Manderbacka et al. 1992). Paternal
characteristics such as occupation, education and age have also recently been
89
linked with low birth weight (Parker & Schoendorf 1992; Chen et al. 2008), though
these studies have been limited due to small sample sizes and an inability to adjust
for multiple maternal, paternal and social factors. Additionally, community level
factors such as relative disadvantage are also known risk factors of low birth weight
(Spencer et al. 1999; Cerdá et al. 2008), with extensive research clearly
demonstrating consistent social gradients in low birth weight, as also found in many
other health and developmental outcomes (Kramer et al. 2000b; Parker JD et al.
1994).
Infants can be of low birthweight because they are born too soon (preterm) or too
small (growth restricted). It is obvious that the reasons for preterm delivery and poor
fetal growth are likely to be quite different and hence it is desirable, if accurate data
on gestational age are available, that these two categories of low birthweight are
studied separately. Using population level data available for all births in Western
Australia the aim of this study was to examine the trends in low birthweight due to
poor fetal growth over the last two decades, to investigate the disparities between
socioeconomic groups and between Aboriginal and non-Aboriginal populations and
to investigate the maternal, social and biological factors that may contribute to social
gradients in poor fetal growth.
METHODS
Data Sources
The population used for this study comprised all live singleton, term (>=37 weeks)
births in Western Australia (WA) between 1984 and 2006 (N = 567 468). Data were
obtained from the Midwives’ Notification System (MNS) and the Registrar General’s
90
Birth Registrations (BR). The MNS is a statutory collection of all births in WA, with
complete birth information from 1980 (Gee & Dawes 1994), and data includes
demographic information about the mother, the pregnancy (e.g. complications and
medical conditions), the labour and delivery, and the infant (e.g. gender and
birthweight). The BR is also a statutory collection of all births and includes additional
information about maternal and paternal country of birth, occupation, age and
Aboriginality. The MNS form is completed by the attending midwife at the time of
birth, while the BR form is completed by the parents following the birth. The MNS
and BR data were linked by Data Linkage WA, located within the WA Department of
Health. Data are linked via probabilistic matching, using identifiers such as full
name, address, date of birth and gender, that are common to all data sets (Kelman
et al. 2002b). Only de-identified data are provided to researchers.
Low birth weight
The percentage of optimal birth weight (POBW) (Blair et al. 2005) was used to
determine low birthweight due to poor fetal growth, instead of the more traditional
definition but less useful of <2500g. This measure is considered to be a more
appropriate measure, as it takes into account important factors that influence
birthweight such as gestational age, maternal height, parity and infant gender.
POBW was calculated by dividing the observed birthweight by the optimal
birthweight, where the optimal birthweight was determined using information
available from the MNS, adjusting for the measures listed. For our analyses poor
fetal growth was defined as having a POBW of 87% or less, corresponding to less
than the 10th percentile for the population (Blair et al. 2005).
91
Independent Variables
Inequalities in poor fetal growth were examined by various maternal characteristics,
including maternal age group (15-19, 20-24, 25-29, 30-34, 35-39, 40-44 and 45-49
years), parity (0, 1, 2 or 3+), marital status (never married, married/de facto and
separate/widowed/divorced), maternal skill level, any pregnancy complication, any
medical condition, maternal residential location at the time of birth and community
level socioeconomic status (SES) at the time of the birth. Maternal smoking during
pregnancy was only examined from 1999 to 2006, as collection only commenced in
1998. Father’s skill level and age group were also included in analyses. Aboriginality
of the infant was defined as having at least one parent being identified as Aboriginal
in any of the MNS or BR records. Residential location was defined as urban or rural,
using the Remoteness Area Index (RAI) produced by the Australian Bureau of
Statistics (ABS) (Trewin 2006). The RAI assigned to each mother was based on the
location of her household Collection District (CD) at the time of birth. CDs are
currently the smallest area unit for which socioeconomic data can be obtained, and
contains approximately 250 households, although this may vary, particularly in more
sparsely populated areas (Pink 2008b). Community level SES for each mother was
determined using CD level data available from the Socioeconomic Indexes for Areas
(SEIFA). SEIFA are four area-level indices produced by the ABS, based on data
obtained from the National Census, conducted every five years. The index used in
this study was the Index of Relative Socioeconomic Disadvantage (IRSD), which
identifies relative disadvantage, or a lack thereof, in communities (Pink 2008b). As
9.5% of all births had missing CD information, this IRSD group was assigned as
unknown.
92
A measure of parental skill and education level was used as an indicator of
individual level socioeconomic status based on the occupation title provided by
parents on the BR form. This occupation title was subsequently coded using the
Australian and New Zealand Standard Classification of Occupations (Trewin & Pink
2006)Langridge, A., Li, J., Nassar, N. & Stanley, F. 2009, unpublished) to assign
skill level ranging from 1, a bachelor degree or higher qualification, to 5, a Certificate
One or compulsory secondary education. Where an occupation title was not
provided or where inadequate information was supplied, parental skill level was
recoded and grouped into an unknown category.
Statistical Analyses
The prevalence and trend of poor fetal growth infants in WA between 1984 and
2006 was assessed based on the total number of term singleton live births. As
community SES data is only available at five-year intervals when the National
Census is completed, analyses were conducted in five year groupings with each
group centred around a census year (i.e. 1986: 1984-1988, 1991: 1989-1993, 1996:
1994-1998, 2001: 1999-2003 and 2006:2004-2006). As MNS and BR data beyond
2006 was not available at the time of analyses, the last year grouping contains only
three years (2004-2006). Only data for three time points will be presented: 1984-
1988, 1994-1998 and 2004-2006, as results in intervening years were similar.
Univariate and multilevel multivariate logistic regression was used to investigate
relative differences in poor fetal growth infants between socioeconomic groups,
taking into account both individual and area level characteristics. As there were
considerable differences between Aboriginal and non-Aboriginal infants, these were
analysed separately. Based on results from the univariate analyses and previous
research, all multivariate regression models were adjusted for parental age group,
93
skill level and Aboriginality (only in analyses examining differences between
Aboriginal and non-Aboriginal infants), residential location at time of birth,
community level socioeconomic disadvantage, parity, marital status, any medical
condition and any pregnancy complication. An additional multivariate model also
adjusted for smoking during pregnancy. The contribution of smoking to the
inequalities in poor fetal growth infants between the most and least disadvantaged
groups was calculated using the following formula from Singh-Manoux et al (Singh-
Manoux et al. 2008):
(RRSES, controlling for all parental and social factors including smoking – RRSES, controlling for all parental and social
factors)/(RRSES, controlling for all parental and social factors
– 1) x 100.
As the gradient reversed, the contribution of smoking for Aboriginal infants was not
estimatable.
Multilevel models were used to account for individual and community level SES
measures used in the same models. In all models, the least disadvantaged group
(group one) was used as the reference group. All analyses were performed using
MLwiN 2.10 and p-values<0.05 were considered statistically significant.
Ethics
Ethics approval for this study was obtained from the Human Research Ethics
Committee at the University of Western Australia and the Western Australian
Aboriginal Health Information and Ethics Committee. Approval for the use of health
data was obtained from the Confidentiality of Health Information Committee at the
Department of Health, WA.
94
RESULTS
There has been a steady decline of approximately 0.7% per annum in the
percentage of term singleton infants with poor fetal growth over the last 23 years,
from 18% in 1984-1988 to 16% in 1999-2003 and then plateauing from 2003
onwards (Figure 16). When examined by community level socioeconomic
disadvantage, the decreasing trend in the percentage of poor fetal growth infants
occurred in all groups except for the two most disadvantaged groups (group five and
six) where it remained steady over time (Figure 17). Overall the proportion of infants
with poor fetal growth rose with increasing levels of social disadvantage, with the
difference between the most and least disadvantaged groups increasing
considerably over time, from 6% in 1984-1988 (16%-22%) to 10% in 2004-2006 (13-
23%).
Figure 16: Trends in poor fetal growth (POBW) in Western Australia: 1984-2006
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
1984-1988 1989-1993 1994-1998 1999-2003 2004-2006
Perc
enta
ge o
f low
birt
hwei
ght b
irths
95
Figure 17: Inequalities in poor fetal growth (POBW) by area-level disadvantage in Western Australia: 1984-2006
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
1984-1988 1994-1998 2004-2006
Perc
enta
ge o
f low
birt
hwei
ght i
nfan
ts
1 (Least) 2 3 4 5 6 (Most) Unknown
Figure 18 highlights that across most years and socioeconomic groups, a higher
percentage of Aboriginal infants were born of poor fetal growth. In 1984-1988, 1.2%
of all births in the least disadvantaged communities were to Aboriginal parents but
20.9% of those infants had poor fetal growth, compared to 15.7% of non-Aboriginal
infants. By 2004-2006, 1.4% of all births in the least disadvantaged communities
were to Aboriginal parents and the percentage of Aboriginal infants with poor fetal
growth had declined slightly to 19.8%, though still higher than the 12.5% of non-
Aboriginal infants.
96
Figure 18: Poor fetal growth by Aboriginality and social disadvantage in Western Australia: 1984-2006
0%
5%
10%
15%
20%
25%
30%
35%
1 6 1 6 1 6
1984-1988 1994-1998 2004-2006
Perc
enta
ge o
f low
birt
hwei
ght i
nfan
ts
Non-Aboriginal Aboriginal^
Further investigation of the differences between Aboriginal and non-Aboriginal
infants revealed that across the study period, Aboriginal infants were significantly
more likely to have poor fetal growth and the difference between Aboriginal and non-
Aboriginal infants continued to increase overtime (Table 13). Prior to adjusting for
any factors, in 1984-1988 Aboriginal infants were 1.55 (95% CI=1.46, 1.64) times
more likely to have poor fetal growth than non-Aboriginal infants, and by 2004-2006
this had increased to a two-fold difference (OR=2.01, 95% CI=1.89, 2.15). The
greatest reduction in risk occurred after adjusting for parental and biological factors,
which reduced the odds by a third (OR=1.29, 95% CI=1.20, 1.39) for 1984-1988,
and 25% (OR=1.45, 95% CI=1.33, 1.58) for 2004-2006. Further adjustment for
social factors reduced the odds again, with the odds of an infant with poor fetal
growth occurring among Aboriginal parents of 1.18 (95% CI=1.10, 1.28) for 1984-
1988, and 1.27 (95% CI=1.16, 1.38) for 2004-2006. The final model, adjusted for
smoking, lowered the odds even further to 1.17 (95% CI=1.07, 1.28) for 2004-2006.
97
A crude analysis of the differences in poor fetal growth by socioeconomic status
revealed striking differences for Aboriginal and non-Aboriginal infants. For non-
Aboriginal infants, infants in all socioeconomic groups (two to six) were significantly
more likely to have poor fetal growth compared with the least disadvantaged group
(one) (Table 14). In 1984-1988, non-Aboriginal parents from the most
disadvantaged communities were 1.35 (95% CI=1.24, 1.47) times more likely to
have an infant with poor fetal growth than non-Aboriginal parents from the least
disadvantaged communities, and by 2004-2006, the odds increased to 1.70 (95%
CI=1.53, 1.89). After adjusting for parental characteristics, but prior to adjusting for
maternal smoking during pregnancy, the effect of community socioeconomic group
on poor fetal growth decreased, although the social gradient continued to persist.
The social gradient was further attenuated after adjusting for maternal smoking
during pregnancy though non-Aboriginal parents living in the most disadvantaged
communities (groups five and six) were still significantly more likely to have an infant
with poor fetal growth than non-Aboriginal parents living in the least disadvantaged
communities (Group 5: OR=1.12, 95% CI=1.01, 1.23; Group 6: OR=1.24; 95%
CI=1.10, 1.39). The addition of smoking to the model accounted for a third of the
differences between the most and least disadvantaged group.
The results for Aboriginal infants were considerably different compared to non-
Aboriginal infants, with no clear social gradient existing for the first three groups
(Table 15). Across all time periods examined there was no difference in groups two
to four. In more recent years, infants born to Aboriginal parents from the most
disadvantaged communities (group six) were significantly more likely to have an
infant with poor fetal growth (1994-1998: OR = 2.03, 95% CI=1.11, 3.70; 2004-2006:
OR=1.75, 95% CI=1.02, 3.00). After adjusting for parental and social factors there
98
was no longer a difference between the most and least disadvantaged groups for
any of the time periods studied and no effect of smoking was found.
99
Table 14: Racial disparities in poor fetal growth and contributing factors in Western Australia: 1984-2006
Years Aboriginality Unadjusted
OR (95% CI)
Adjusted
OR (95% CI)
1 Adjusted
OR (95% CI)
2 Adjusted
OR (95% CI)
3
1984-1988
Non-Aboriginal
Aboriginal
Ref
1.55 (1.46, 1.64)***
Ref
1.29 (1.20, 1.39)***
Ref
1.18 (1.10, 1.28)***
Ref
n/a
1994-1998
Non-Aboriginal
Aboriginal
Ref
1.69 (1.61, 1.78)***
Ref
1.35 (1.27, 1.44)***
Ref
1.21 (1.14, 1.29)***
Ref
n/a
2004-2006 Non-Aboriginal
Aboriginal
Ref
2.01 (1.89, 2.15)***
Ref
1.45 (1.33, 1.58)***
Ref
1.27 (1.16, 1.38)***
Ref
1.17 (1. 07, 1.28)***
* p < 0.05 **p <0.01 ***p <0.001 1 Adjusted for: Maternal age, paternal age, parity, any medical condition and any pregnancy complication 2 Adjusted for: Maternal age, paternal age, parity, any medical condition, any pregnancy complication, marital status, residential location at time
of birth, community level socioeconomic disadvantage, maternal skill level and paternal skill level 3
Adjusted for Maternal age, paternal age, parity, any medical condition, any pregnancy complication, marital status, residential location at time
of birth, community level socioeconomic disadvantage, maternal skill level, paternal skill level and smoking during pregnancy
100
Table 15: Absolute and relative differences in poor fetal growth births by area-level disadvantage in non-Aboriginal infants
Years Socioeconomic Group Low
Birthweight N (%)
Total Births
Absolute Difference^
Relative Difference Unadjusted OR (95% CI)
AdjustedOR (95% CI)
1 AdjustedOR (95% CI)
2
1984
-198
8 IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4 IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
1535 (15.7)
2806 (16.1)
3674 (17.0)
3982 (18.1)
2366 (18.3)
1200 (20.0)
1883 (16.9)
9807
17430
21561
21986
12917
6002
11148
Ref
0.45
1.39
2.46
2.66
4.34
1.24
Ref
1.03 (0.97, 1.11)
1.11 (1.04, 1.18)**
1.19 (1.12, 1.27)***
1.21 (1.13, 1.30)***
1.35 (1.24, 1.47)***
1.62 (0.65, 4.03)
Ref
0.99 (0.92, 1.06)
1.02 (0.95, 1.09)
1.06 (0.99, 1.14)
1.07 (0.99, 1.15)
1.13 (1.03, 1.24)**
1.53 (0.61, 3.82)
Ref
n/a
n/a
n/a
n/a
n/a
n/a
1994
-199
8
IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4 IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
1011 (12.8)
2010 (14.2)
3916 (14.9)
4425 (16.2)
2820 (18.2)
1636 (19.4)
955 (16.0)
7879
14175
26327
27275
15476
8423
5987
Ref
1.35
2.04
3.39
5.39
6.59
3.12
Ref
1.12 (1.04, 1.22)**
1.19 (1.10, 1.28)***
1.32 (1.22, 1.41)***
1.51 (1.40, 1.64)***
1.64 (1.50, 1.78)***
1.45 (1.01, 2.10)*
Ref
1.07 (0.99, 1.16)
1.08 (1.00, 1.16)
1.15 (1.07, 1.24)***
1.28 (1.18, 1.39)***
1.30 (1.19, 1.43)***
1.24 (0.84, 1.82)
Ref
n/a
n/a
n/a
n/a
n/a
n/a
101
2004
-200
6
IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4 IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
767 (12.5)
1440 (13.7)
2166 (13.9)
2315 (14.4
1599 (16.8)
841 (19.5)
724 (15.5)
6158
10483
15551
16130
9527
4322
4675
Ref
1.28
1.47
1.90
4.33
7.00
3.03
Ref
1.12 (1.02, 1.23)*
1.14 (1.04, 1.24)**
1.18 (1.08, 1.29)***
1.42 (1.29, 1.55)***
1.70 (1.53, 1.89)***
1.70 (1.05, 2.75)***
Ref
1.10 (1.00, 1.21)
1.07 (0.98, 1.18)
1.06 (0.97, 1.16)
1.21 (1.09, 1.33)***
1.36 (1.21, 1.53)***
1.51 (0.92, 2.48)
Ref
1.09 (0.99, 1.20)
1.05 (0.96, 1.13)
1.02 (0.93, 1.11)
1.12 (1.01, 1.23)*
1.24 (1.10, 1.39)***
1.40 (0.85, 2.31) ^per 100 births (live singleton births only)
*Unadjusted model only contains community level socioeconomic disadvantage 1 Adjusted for: Mother’s age group, Father’s age group, Mother’s skill level, Father’s skill level, residential location at time of birth, community level
socioeconomic disadvantage, parity, marital status, any medical condition and any pregnancy complication 2
* p < 0.05 **p <0.01 ***p <0.001
Adjusted for Mother’s age group, Father’s age group, Mother’s skill level, Father’s skill level, residential location at time of birth, community level
socioeconomic disadvantage, parity, marital status, any medical condition and any pregnancy complication, smoking during pregnancy
102
Table 16: Absolute and relative differences in poor fetal growth births by area-level disadvantage in Aboriginal infants
Years Socioeconomic Group Low
Birthweight N (%)
Total Births
Absolute Difference^
Relative Difference Unadjusted OR (95% CI)
AdjustedOR (95% CI)
1 AdjustedOR (95% CI)
2
1984
-198
8 IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4
IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
24 (20.9)
48 (18.5)
131 (18.7)
404 (24.3)
444 (25.6)
543 (26.9)
397 (27.5)
115
259
701
1665
1734
2016
1443
Ref
-2.34
-2.18
3.39
4.74
6.06
6.64
Ref
0.86 (0.50, 1.49)
0.87 (0.54, 1.42)
1.21 (0.76, 1.93)
1.30 (0.82, 2.07)
1.40 (0.88, 2.22) -
Ref
0.76 (0.43, 1.35)
0.69 (0.41, 1.15)
0.89 (0.55, 1.47)
0.86 (0.52, 1.42)
0.84 (0.52, 1.41) -
Ref
n/a
n/a
n/a
n/a
n/a
n/a
1994
-199
8
IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4 IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
13 (16.7)
57 (19.7)
232 (19.0)
495 (22.4)
493 (22.9)
847 (28.9)
301 (30.7)
78
290
1221
2207
2153
2934
981
Ref
2.99
2.33
5.76
6.23
12.20
14.02
Ref
1.22 (0.63, 2.37)
1.17 (0.64, 2.16)
1.44 (0.79, 2.64)
1.48 (0.81, 2.71)
2.03 (1.11, 3.70)*
3.27 (1.58, 6.78)**
Ref
1.06 (0.54, 2.09)
0.81 (0.43, 1.51)
0.93 (0.50, 1.73)
0.91 (0.49, 1.71)
1.16 (0.62, 2.17)
1.58 (0.62, 3.99)
Ref
n/a
n/a
n/a
n/a
n/a
n/a
103
2004
-200
6
IRSD 1 (Least Disadvantaged) IRSD 2 IRSD 3 IRSD 4 IRSD 5 IRSD 6 (Most Disadvantaged) IRSD Unknown
17 (19.8)
29 (12.7)
121 (20.8)
281 (23.4)
382 (26.6)
554 (30.1)
282 (28.7)
86
228
583
1200
1438
1841
981
Ref
-7.05
0.99
3.65
6.80
10.32
8.98
Ref
0.59 (0.31, 1.14)
1.06 (0.60, 1.87)
1.24 (0.72, 2.14)
1.47 (0.85, 2.53)
1.75 (1.02, 3.00)*
1.38 (0.72, 2.66)
Ref
0.62 (0.30, 1.28)
0.94 (0.50, 1.77)
0.98 (0.53, 1.82)
1.09 (0.59, 2.02)
1.03 (0.55, 1.91)
1.07 (0.47, 2.47)
Ref
0.58 (0.28, 1.21)
0.86 (0.46, 1.64)
0.89 (0.48, 1.65)
0.94 (0.51, 1.75)
0.89 (0.48, 1.67)
0.93 (0.40, 2.16) ^per 100 births (live singleton births only)
*Unadjusted model only contains community level socioeconomic disadvantage 1 Adjusted for: Mother’s age group, Father’s age group, Mother’s skill level, Father’s skill level, residential location at time of birth, community level
socioeconomic disadvantage, parity, marital status, any medical condition and any pregnancy complication 2
* p < 0.05 **p <0.01 ***p <0.001
Adjusted for Mother’s age group, Father’s age group, Mother’s skill level, Father’s skill level, residential location at time of birth, community level
socioeconomic disadvantage, parity, marital status, any medical condition and any pregnancy complication, smoking during pregnancy
104
DISCUSSION
This is the first Australian study to utilise population-level data to examine racial and
social inequalities in poor fetal growth for the last two decades. Results demonstrate
that while there has been an overall decline in the rate of infants with poor fetal
growth over the last 23 years and a plateauing in recent years, inequalities between
socioeconomic groups have increased, particularly between the most and least
disadvantaged groups. Findings highlight a strong social gradient, with a higher
proportion of infants with poor fetal growth occurring with increasing levels of
disadvantage. The findings also highlight that Aboriginal infants are significantly
more likely to have poor fetal growth than non-Aboriginal infants and social disparity
has appeared to increase over time, particularly among non-Aboriginal infants. A
significant portion of the disparity is attributable to parental socioeconomic and
demographic characteristics and biological factors, such as maternal medical
conditions and pregnancy complications. After stratifying by Aboriginality and
accounting for these important social, biological and parental factors, the social
gradient remained for non-Aboriginal infants, but completely disappeared for
Aboriginal infants.
Consistent with the findings from this study, there has been extensive interest
examining the increasing inequalities in health and developmental outcomes over
the last decade, and in 2002 the UN set a goal to ‘reduce the rate of low birthweight
infants by at least a third of the current rate’((UNICEF) 2007), resulting in the
implementation of a range of policies and programs (Davis 2000; National Public
Health Partnership 2004; Panaretto et al. 2005; Stamp et al. 2008). Despite these
attempts, inequalities continue to increase or remain unchanged (Moser et al. 2003;
Lesley 2005), with most reductions in the rate of infants with poor fetal growth
105
occurring in the least disadvantaged groups (Bundred et al. 2003). It has been
suggested that the unchanged rates in the most disadvantaged groups could be due
to limited access to health services (Hughes & Simpson 1995), a lack of resources
(Chomitz et al. 1995) and ineffective intervention programs, such as smoking
cessation programs that may not be a priority for women living in the most
disadvantaged communities as they are likely to already have many other issues
and stressors in their lives (Mullen 1999; Wood et al. 2008; DiGiacomo et al. 2007).
Inequalities between Aboriginal and non-Aboriginal populations in Australia have
been widely reported across a range of health outcomes (Freemantle et al. 2006b;
Freemantle et al. 2006c; Zinn 1997). While some research has suggested that the
differences are, at least in part, due to social disadvantage (AIHW: Leeds KL 2007),
other research has suggested that it is likely to be a result of pathological factors
and greater exposure to other risk factors such as smoking, alcohol and drug use
(Blair 1996; Schultz et al. 1991). While research has demonstrated that after 35
weeks gestation Aboriginal infants have a slightly lower median birthweight even
after adjusting for gestational age (Kliewer EV & FJ. 1989) additional research has
indicated that pathological factors account for most of the difference in birthweight
between Aboriginal and non-Aboriginal infants (Blair 1996). Our findings reveal that
even after taking into account parental, biological and social factors, including
smoking, an infant with at least one Aboriginal parent is still 17% more likely to have
an infant with poor fetal growth compared with a non-Aboriginal infant. Furthermore,
results show that these inequalities have been increasing over time despite the
introduction of programs specifically aimed to improve Aboriginal maternal and
infant health, such as the Aboriginal Primary Care Program (Health Reform
Implementation Taskforce 2007) and the Healthy for Life Program (Department of
Health and Ageing 2005). Previous research supports these findings and suggest
106
differences could be due to poor antenatal attendance, and a higher proportion of
Aboriginal people living in remote areas with very limited access to health services
(Panaretto et al. 2002a). Malnutrition and low BMI prior to pregnancy (Sayers &
Powers 1997), unsafe alcohol use (Panaretto et al. 2006) and unmeasured maternal
social and psychological factors (Jonas et al. 1992; Morrison et al. 1989b) have also
been proposed as potential factors contributing to the differences. Based on these
findings, further research should be conducted to examine whether culturally
appropriate programs aimed at increasing access to services could reduce the
disparity and improve outcomes.
A distinctive finding from our study is the difference in social gradients between
Aboriginal and non-Aboriginal infants after adjusting for the various factors, and
more specifically, the lack of a social gradient for Aboriginal infants. For non-
Aboriginal infants there was a distinct social gradient in poor fetal growth that
increased over time and did not disappear after accounting for important parental
and social factors, including maternal smoking. For Aboriginal infants the
disappearance of the social gradient from an up to two-fold increased risk to no
association highlights that parental and social factors contribute substantially to
infant outcomes.
The strengths of this study are that the study population comprised all live,
singleton, full-term births in WA between 1984 and 2006, analyses were conducted
at CD level, and both maternal and paternal variables were included in the analyses,
including parental education which was used as an individual-level SES measure.
Using total population level administrative data ensures that all eligible births were
included in the analyses, and with very little missing data (5%). Analysis of data at
the CD level is a particular strength as the community level SES measure is based
107
on small areas containing approximately 250 households, as opposed to the
postcodes and larger areas (containing up to 212 CDs) that are often used to
represent communities, but may mask SES variation among CDs. Inclusion of
parental skill level as a proxy for individual-level SES is also an advantage as most
administrative datasets do not contain socioeconomic indicators at an individual
level. Further, the addition of several paternal variables in analyses is also beneficial
as past research has indicated that paternal characteristics are important in
predicting low birthweight (Chen et al. 2008; Parker & Schoendorf 1992). However
due to limited data most studies have been unable to control for these in analyses.
While administrative datasets provides population-level data, it often includes
varying amounts of missing data which can potentially lead to under-ascertainment.
In Australia, it has been acknowledged that Aboriginality is often missing (Australian
Institute of Health and Welfare 2008) from administrative datasets. In this study
linked health data overcomes this issue and increases reliability by providing a
means of identification from multiple data sources, particularly when there is
inconsistent recording of variables across datasets (Mak & Watkins 2008). An
additional limitation of using administration data is the lack of detail available for
certain variables. In the case of skill level, although parental job titles recorded at
birth were used to code skill level, the job title provides no information on duties
performed or indication of qualification.
In conclusion, while the overall rates of poor fetal growth in Western Australia have
decreased over the last two decades, the inequalities between the most and least
disadvantaged groups, and between Aboriginal and non-Aboriginal infants have
increased. This study highlights the misleading nature of observing only overall rates
of poor fetal growth that mask the increase in inequalities. It also highlights the
108
importance of monitoring trends in poor fetal growth by social characteristics. Failure
to do this may result in incorrect or misguided implementation of policies and
interventions that do not target the very populations that need them, and lead to
poorer health outcomes and greater health disparities. While findings suggest that
the disparity may at least in part be due to access to culturally appropriate services
and resources, it would be of interest to investigate which aspects of disadvantage
are contributing the most to the increasing inequalities between Aboriginal and non-
Aboriginal populations so that policymakers can be provided with more direction in
their attempt to close the gap.
109
CHAPTER SEVEN: SOCIAL AND RACIAL INEQUALITIES IN PRETERM BIRTHS IN WESTERN
AUSTRALIA, 1984 TO 2006
INTRODUCTION
Preterm birth is the leading cause of infant mortality in industrialised societies
(Goldenberg et al. 2008; McCormick 1985) and is associated with a range of short and
long-term morbidities (Wang et al. 2004; Saigal & Doyle 2008; Escobar et al. 2006;
Huddy et al. 2001; Boardman et al. 2002; Bohnert & Breslau 2008). While it has been
estimated that the average immediate cost of a preterm infant in neonatal intensive
care is AUS$1,500 a day and each infant born before 26 weeks costs approximately
AUS$90,000 (Smith & Shen), these infants also place additional strain on health,
education and social services in the long term (Petrou et al. 2001).
Rates of preterm births have been reported to be increasing in some industrialised
countries (Joseph et al. 1998; Langhoff-Roos et al. 2006; Chalmers et al. 2006; Ananth
et al. 2005), though this increase has not been seen in Australia, with relatively stable
national rates throughout the 1990’s (Day et al. 1999). Currently, most countries report
that 5-9% of births are preterm, though the United States has a higher rate at 12-13%.
Goldenberg (2008) suggests that the higher proportion of preterm births to black
women (16-18%), which is two to three times higher than that for white women (5-9%),
might be a contributing factor to the elevated rate seen in the US (Goldenberg et al.
2008). It has been well-established that there are racial inequalities in the US, UK, New
Zealand and Canada with African-American and Aboriginal women respectively,
experiencing considerably higher rates of preterm births than Caucasians (Heaman et
al. 2005; Craig et al. 2002; Goldenberg et al. 2008; Ahern et al. 2003; Ananth et al.
2005; Messer et al. 2008). While there are several Australian studies that have
compared preterm births in Aboriginal and non-Aboriginal infants, these studies have
110
been limited by small sample sizes and localised study areas and have not examined
or explained the racial inequalities (Eades et al. 2008; Najman et al. 1994; Panaretto et
al. 2007; Panaretto et al. 2002b). Social inequalities in preterm births have also been
documented in the UK, Denmark, Finland, Norway, Sweden, US and NZ (O'Campo et
al. 2008; Gray et al. 2008b; Morgen et al. 2008; Craig et al. 2002) with rising rates of
preterm birth occurring with increasing levels of disadvantage. However, there have
been limited investigation of social inequalities in preterm births and how such
inequalities may differ between Aboriginal and non-Aboriginal infants in Australia.
Preterm birth is influenced by biological and socio-demographic factors including
maternal age, parity, marital status, maternal education, smoking during pregnancy,
pre-existing medical conditions (e.g. essential hypertension, asthma, pre-existing
diabetes mellitus), pregnancy complications (e.g. threatened abortion, pre-eclampsia,
gestational diabetes, urinary tract infection), low maternal BMI prior to pregnancy and
the use of IVF (Olsén et al. 1995; Goldenberg et al. 2008; Messer et al. 2008; Morgen
et al. 2008). Paternal characteristics such as ethnicity, occupation and age have also
recently been linked with preterm birth (Morgen et al. 2008; Palomar et al. 2007; Astolfi
et al. 2006; Simhan & Krohn 2008). An association between preterm birth and
community level income, unemployment and deprivation has also been reported (Luo
et al. 2006; Ahern et al. 2003; O'Campo et al. 2008; Craig et al. 2002; Messer et al.
2008). While many studies have been able to adjust for a variety of socio-demographic
factors, our literature review could not find any studies examining preterm birth that had
adjusted for individual (e.g. maternal, paternal demographic factors) and community
level factors concurrently, which would enable us to address the important issue of
whether or not accounting for parental SES would reduce the inequality observed at
area-level.
Using population level data available for all live singleton births in Western Australia the
aim of this study was to examine the trends in preterm birth over the last two decades,
111
investigate the racial inequalities in preterm birth among Aboriginal and non-Aboriginal
infants, determine whether social inequalities are different for Aboriginal and non-
Aboriginal infants and shed light on factors contributing to these inequalities.
METHODS
Data Sources
The population used for this study comprised all live singleton births in Western
Australia (WA) between 1984 and 2006 (N = 567 468). Data were obtained from the
Midwives’ Notification System (Midwives’) and the Registrar General’s Birth
Registrations. The Midwives’ is a statutory collection of all births in WA (>20 weeks
gestation and/or >400g birthweight), with complete birth information from 1980
onwards (Gee & Dawes 1994), and the data include demographic information about
the mother, pregnancy (e.g. complications and medical conditions), labour, delivery,
and infant (e.g. gender and birthweight). Birth Registrations are also a statutory
collection of all births in WA and includes additional information about maternal and
paternal country of birth, occupation, age and Aboriginality. The Midwives’ form is
completed by the attending midwife at the time of birth, while the Birth Registration
form is completed by the parents following the birth. Both of these data sources were
linked for each individual by Data Linkage WA, located within the Department of Health
of WA. Data are linked via probabilistic matching, using identifiers such as full name,
address, date of birth and gender, which are common to all data sets (Kelman et al.
2002b). Only de-identified data are provided to researchers following multiple ethics
committee reviews.
Variables
Inequalities in preterm births (<37 weeks) were assessed by accounting for various
maternal characteristics, including maternal age group (15-19, 20-24, 25-29, 30-34, 35-
112
39 and 40 plus years), parity (0, 1, 2 or 3+), marital status (never married, married/de
facto and separate/widowed/divorced), maternal skill level, maternal residential location
at the time of birth and community level socioeconomic status (SES) at the time of the
birth. Pre-existing medical conditions (e.g. essential hypertension, asthma, pre-existing
diabetes mellitus) and pregnancy complications (e.g. threatened abortion, pre-
eclampsia, gestational diabetes, urinary tract infection) were also adjusted for in
analyses as a binary variable (yes/no), to account for the increased risk in preterm birth
associated with those conditions. Maternal smoking during pregnancy was only
examined from 1999 to 2006, as collection of smoking data commenced in 1998.
Father’s skill level and age group were also included in analyses. Aboriginality of the
infant was defined as having at least one parent being identified as Aboriginal in any of
the data sources.
Residential location was defined as urban or rural, using the Remoteness Area Index
(RAI) produced by the Australian Bureau of Statistics (ABS) (Trewin 2006). The RAI
assigned to each mother was based on the location of her household Collection District
(CD) at the time of birth. CDs are currently the smallest area unit for which
socioeconomic data can be obtained and contains approximately 250 households,
although this may vary, particularly in more sparsely populated areas (Pink 2008b).
Community level SES for each mother was determined using data available from the
Socioeconomic Indexes for Areas (SEIFA) at the CD level. SEIFA are four area-level
indices produced by the ABS, based on data obtained from the National Census,
conducted every five years. The index used in this study was the Index of Relative
Socioeconomic Disadvantage (IRSD), which identifies relative disadvantage, or a lack
thereof, in CDs (Pink 2008b). As 9.5% of all births had missing CD information, this
IRSD group was assigned as “unknown” and included as such in all analyses.
Parental skill level was used as an indicator of individual level socioeconomic status.
The skill level was based on the occupation title provided by parents on the Birth
113
Registration form, using the Australian and New Zealand Standard Classification of
Occupations (Trewin & Pink 2006) (Langridge et al, 2009). Skill level ranged from 1 (a
bachelor degree or higher qualification) to 5 (a Certificate One or compulsory
secondary education). Where an occupation title was not provided or where inadequate
information was supplied, parental skill level was recorded as unknown and this was
included as a separate group in all analyses.
Statistical Analyses
The prevalence and trend of preterm births in WA between 1984 and 2006 were
assessed as a proportion of the annual number of singleton live births. As community
SES data is only available at five-year intervals when the National Census is
completed, descriptive analyses (Figures 1-2) were conducted in five year groupings
with each group centred around a census year (i.e. 1986: 1984-1988, 1991: 1989-
1993, 1996: 1994-1998, 2001: 1999-2003 and 2006: 2004-2006). As Midwives’ and
Birth Registrations data beyond 2006 were not available at the time of analyses, the
last year grouping contains only three years. Multivariable analyses included all data
from 1984 to 2006, with census year included to reflect the change in socioeconomic
conditions over time. Cochran-Armitage trend tests (Agresti 2002) were used to
examine the trend across socioeconomic groups.
Univariate and multilevel multivariable logistic regression models were applied to
investigate relative differences in preterm birth between Aboriginal and non-Aboriginal
infants. As these two groups produced quite different results, analyses were conducted
separately for the two groups. All multivariable regression models included a range of
potential contributing factors for racial and social inequalities in preterm birth and these
factors were entered into the model in various stages. Model one comprised parental
biological and demographic factors (parental age group, parity, any medical condition,
any pregnancy complication and census year), model two included parental biological,
demographic and social factors (marital status, parental education level and residential
114
location at time of birth and community level socioeconomic disadvantage) and model
three included maternal smoking in addition to the factors in model two.
Multilevel models were applied to appropriately account for variations attributable to
individual and community level SES measures. In all models, the least disadvantaged
group (group one) was used as the reference group. Analyses were performed using
SAS 9.1 and MLwiN 2.10 and p-values <0.05 were considered statistically significant.
Ethics
Ethics approval for this study was obtained from the Human Research Ethics
Committee (#1080) at the University of Western Australia and the Western Australian
Aboriginal Health Information and Ethics Committee (#122-01/06). Approval for the use
of health data was obtained from the Confidentiality of Health Information Committee
(#200613) at the Department of Health, WA.
RESULTS
The prevalence of preterm births increased from 7.1% in 1984-1988 to 7.5% in 1999-
2003 before decreasing to 7.2% in 2004-2006 (Figure 19). When examined by
community level socioeconomic disadvantage, the most noticeable change was the
continued increase in preterm births among infant from the least disadvantaged
communities until 2004-2006 when it decreased slightly (Figure 19).
115
Figure 19: Trends and inequalities in preterm births by area-level social disadvantage in Western Australia, 1984-2006
Figure 20 highlights that across most years, a higher percentage of Aboriginal infants
were born preterm compared with non-Aboriginal infants, regardless of community-
level social disadvantage. Among infants born to mothers from the least disadvantaged
communities, the difference in the percentage of preterm births among Aboriginal and
Non-Aboriginal infants varied over time from being fairly similar in 1984-1988 to
differing by 3.7% in 1994-1998 and remaining different by 2.6% in 2004-2006. For
infants born to mothers from the most disadvantaged communities, the percentage of
preterm births was almost two-fold higher for Aboriginal infants (14.8%), compared to
non-Aboriginal infants (7.6%) and remained so over time. Results from the trend tests
revealed that for non-Aboriginal infants there was a significant linear increase in
preterm births across all socioeconomic groups in 1984-1988 (P<0.01), 1989-1993
(P<0.01) and 2004-2006 (P<0.01), while for Aboriginal infants a significant linear
increase in preterm births only existed up to 1993 (P<0.01).
116
Figure 20: Preterm births by Aboriginality and area-level social disadvantage in Western Australia, 1984-2006
Further investigation of the differences between Aboriginal and non-Aboriginal infants
revealed that across the study period, Aboriginal infants were significantly more likely
to be preterm and the disparity continued to increase (Table 16). The greatest
reduction in risk of preterm births occurred after adjusting for parental and biological
factors, which reduced the odds by almost a quarter [OR=1.58, 95% CI=1.44, 1.74] for
1984-1988, and 20% [OR=1.60, 95% CI=1.44, 1.79] for 2004-2006 (Table 16). Further
adjustment for social factors slightly reduced the odds again, with the odds of a preterm
infant occurring among Aboriginal parents still elevated at 1.51 [95% CI=1.37, 1.67] for
1984-1988, and similarly for 2004-2006, with no difference after accounting for
smoking.
117
Table 17: Racial disparities in preterm births and contributing factors in Western Australia: 1984-2006
Years Aboriginality Unadjusted OR (95% CI)
Model 110
OR (95% CI) Model 211
OR (95% CI) Model 312
OR (95% CI)
(N = 567468)
1984-1988
Non-Aboriginal
Aboriginal Reference
2.06 (1.91,
2.21)***
Reference
1.58 (1.44,
1.74)***
Reference
1.51 (1.36,
1.67)***
1989-1993 Non-Aboriginal
Aboriginal Reference
1.73 (1.61,
1.86)***
Reference
1.46 (1.33,
1.59)***
Reference
1.41 (1.28,
1.55)***
1994-1998
Non-Aboriginal Aboriginal
Reference
1.84 (1.73,
1.97)***
Reference
1.46 (1.34,
1.58)***
Reference
1.36 (1.24,
1.48)***
1999-2003
Non-Aboriginal
Aboriginal Reference
1.78 (1.67,
1.90)***
Reference
1.53 (1.40,
1.66)***
Reference
1.43 (1.31,
1.57)***
Reference
1.40 (1.28,
1.53)***
2004-2006
Non-Aboriginal
Aboriginal Reference
1.98 (1.83,
2.14)***
Reference
1.60 (1.44,
1.79)***
Reference
1.49 (1.33,
1.68)***
Reference
1.48 (1.32,
1.66)***
* p < 0.05 **p <0.01 ***p <0.001
10 Adjusted for parental biological and demographic factors (Parental age, parity, medical conditions, pregnancy complications) 11 Adjusted for parental biological, demographic and socioeconomic factors (as above, plus parental skill level, residential location, marital status, community socioeconomic status) 12 Adjusted for parental biological, demographic and socioeconomic factors, including smoking. Only contains data from 1999 onwards.
118
Results in Tables 17 and 18 present social inequalities in preterm birth for non-
Aboriginal and Aboriginal infants, respectively. For non-Aboriginal infants, univariate
analyses showed there was a slight, but consistent social gradient in preterm births.
After adjusting for biological (Model 1) and social factors (Model 2) the gap and trend in
preterm births with increasing disadvantage persisted, but the effect and differences
between the groups disappeared after accounting for smoking (Model 3) (Table 17).
For Aboriginal infants, there was a consistent social gradient with increasing risk of
preterm birth corresponding to increasing social disadvantage. When parental
biological and demographic factors in Model 1 and social factor in Model 2 the odds
ratios were attenuated, but the trend in the risk of preterm birth remained the same
between the groups and the odds highest among the category where social
disadvantage was not recorded. After adjusting for smoking, the risk of preterm
birth did not change for most of the groups and remained elevated, but results were
imprecise with wide confidence intervals due to small numbers (Table 18).
119
Table 18: Social inequalities in preterm birth and contributing factors: non-Aboriginal infants
Socioeconomic Group Unadjusted
OR (95% CI)
Model 113
OR (95% CI)
Model 214
OR (95% CI)
Model 315
OR (95% CI)
( N= 517073)
IRSD 1 (Least
Disadvantaged)
IRSD 2
IRSD 3
IRSD 4
IRSD 5
IRSD 6 (Most
Disadvantaged)
IRSD Unknown
Reference
1.04 (0.99, 1.09)
1.04 (0.99, 1.09)
1.05 (1.01, 1.10)*
1.13 (1.08, 1.19)***
1.20 (1.14, 1.27)***
0.91 (0.67, 1.25)
Reference
1.05 (1.00, 1.10)
1.05 (1.00, 1.10)
1.04 (0.99, 1.09)
1.11 (1.06, 1.17)***
1.14 (1.07, 1.21)***
0.89 (0.65, 1.23)
Reference
1.04 (0.99, 1.09)
1.03 (0.98, 1.08)
1.02 (0.97, 1.07)
1.08 (1.02, 1.13)**
1.10 (1.03, 1.16)**
0.87 (0.63, 1.20)
Reference
1.04 (0.96, 1.14)
1.03 (0.95, 1.11)
0.96 (0.89, 1.04)
0.99 (0.91, 1.08)
1.02 (0.92, 1.14)
0.90 (0.60, 1.36)
* p < 0.05 **p <0.01 ***p <0.001
13 Adjusted for parental biological and demographic factors (Parental age, parity, medical conditions, pregnancy complications) 14 Adjusted for parental biological, demographic and socioeconomic factors (as above, plus parental skill level, residential location, marital status, community socioeconomic status) 15 Adjusted for parental biological, demographic and socioeconomic factors, including smoking. Only contains data from 1999 onwards.
120
Table 19: Social inequalities in preterm birth and contributing factors: Aboriginal infants
Socioeconomic Group Unadjusted
OR (95% CI)
Model 116
OR (95% CI)
Model 217
OR (95% CI)
Model 318
OR (95% CI)
( N= 50395)
IRSD 1 (Least
Disadvantaged)
IRSD 2
IRSD 3
IRSD 4
IRSD 5
IRSD 6 (Most
Disadvantaged)
IRSD Unknown
Reference
1.09 (0.76, 1.56)
1.28 (0.92, 1.78)
1.48 (1.08, 2.05)*
1.70 (1.23, 2.34)**
1.86 (1.35, 2.55)***
1.89 (1.27, 2.80)***
Reference
1.14 (0.76, 1.71)
1.27 (0.88, 1.84)
1.47 (1.02, 2.10)*
1.56 (1.09, 2.23)*
1.63 (1.14, 2.34)**
2.08 (1.26, 3.43)**
Reference
1.12 (0.74, 1.68)
1.21 (0.83, 1.75)
1.35 (0.93, 1.95)
1.40 (0.97, 2.01)
1.43 (0.99, 2.07)
1.77 (1.06, 2.94)*
Reference
1.45 (0.70, 3.00)
1.05 (0.52, 2.10)
1.41 (0.72, 2.79)
1.42 (0.72, 2.79)
1.43 (0.72, 2.82)
1.42 (0.64, 3.19)
* p < 0.05 **p <0.01 ***p <0.001
16 Adjusted for parental biological and demographic factors (Parental age, parity, medical conditions, pregnancy complications) 17 Adjusted for parental biological, demographic and socioeconomic factors (as above, plus parental skill level, residential location, marital status, community socioeconomic status) 18 Adjusted for parental biological, demographic and socioeconomic factors, including smoking. Only contains data from 1999 onwards.
121
DISCUSSION
This is the first Australian study to utilise population-level data to examine racial and
social inequalities, in preterm births among Aboriginal and non-Aboriginal infants in
Western Australia, after accounting for known biological and social factors. Results
demonstrate that while the overall rate of preterm births in WA has remained fairly
static over the last 23 years, inequalities in preterm births between Aboriginal and non-
Aboriginal infants have increased and are unacceptably high. The State wide trend also
masked significant variations amongst infants from different socioeconomic areas. A
significant portion of the disparity between Aboriginal and non-Aboriginal infants is
attributable to parental socioeconomic and demographic characteristics, though the
disparity continues to persist even after adjustment of these factors. After stratifying by
Aboriginality and accounting for these important social, biological and parental factors,
the social gradient disappeared for Aboriginal and non-Aboriginal infants.
Preterm births are a major public health concern, accounting for more than two-thirds
of all perinatal deaths (Lumley 2003) and placing additional strain on health, education
and social services as a result of the associated morbidities (Petrou et al. 2001). The
steady trend seen in the WA rate of preterm births over the last 23 years is consistent
with National data for Australia in the 1990s (Day et al. 1999). However, the overall
trend for WA is in contrast to the increasing trends reported in other countries (Joseph
et al. 1998; Langhoff-Roos et al. 2006; Chalmers et al. 2006; Ananth et al. 2005). Our
findings are also different to those reported by Tracy et al (Tracy et al. 2007) who found
that between 1994 and 2003, the proportion of Australian women having a preterm
birth increased by 12%. This difference may be due to the inclusion of multiple births in
the previous study which are known to be at higher risk of preterm birth (The Practice
Committee of the American Society for Reproductive Medicine 2006).
122
Inequalities between Aboriginal and non-Aboriginal populations in Australia have been
reported across a range of infant outcomes including mortality, poor fetal growth and
low birthweight (Freemantle et al. 2006b; Panaretto et al. 2002a; Najman et al. 1994;
Eades et al. 2008)(Langridge, Li, Nassar and Stanley, 2009). It has been suggested
that these inequalities exist for various reasons, including: poorer living conditions,
higher proportion of infants born to mothers residing in rural and remote areas, higher
rates of smoking and consumption of alcohol during pregnancy and greater exposure
to social and economic disadvantage (Najman et al. 1994; Freemantle et al. 2006b;
Panaretto et al. 2002a; Elder et al. 1999). It has also been suggested that poorer
attendance to medical services, due to a lack of access to services, or particularly
culturally appropriate services, may be a contributing factor (Najman et al. 1994). Our
findings reveal that after accounting for parental biological, demographic and
socioeconomic factors including smoking, an Aboriginal infant is still almost 50% more
likely to be preterm compared with a non-Aboriginal infant in 1999-2003 and 2004-
2006. The reduced odds ratios after accounting for social and demographic factors
highlight the importance of these factors on the rates of preterm birth. This suggests
that if these factors can be modified through intervention and social and economic
policy, then it is highly likely that the social and racial inequalities in the rates of preterm
birth could be reduced.
These findings are particularly alarming as programs specifically aimed to improve
Aboriginal maternal and infant health, such as the Aboriginal Primary Care Program
and the Healthy for Life Program (Health Reform Implementation Taskforce 2007;
Department of Health and Ageing 2005) have been implemented in an attempt to
improve outcomes and ultimately reduce inequalities. However, while this study has
adjusted for many parental demographic and socioeconomic factors, it is possible that
there are unmeasured factors contributing to the increased risk, such as maternal
weight, maternal stress and family and community characteristics.
123
Although there was a general increasing trend with increasing disadvantage for
unadjusted results for all infants, a distinctive finding from our study is the lack of a
clear social gradient among Aboriginal and non-Aboriginal infants after adjusting for all
parental biological, demographic and socioeconomic factors, including smoking during
pregnancy. In contrast, previous research has clearly demonstrated a social gradient in
preterm birth, regardless of whether socioeconomic status is at the individual or area
level (O'Campo et al. 2008; Messer et al. 2008; Petersen et al. 2009; Gray et al. 2008b;
Morgen et al. 2008; Luo et al. 2006). Furthermore O’Campo et al (2008) found social
gradients persisted among non-Hispanic white women and non-Hispanic black women
when analysed separately, though only a small number of maternal socioeconomic and
demographic characteristics were adjusted for. Interestingly, a NZ study has reported
that the social gradient seen in preterm births during the 1980’s had disappeared by
the late 1990’s (Craig et al. 2002).
The strengths of this study are that the study population comprised all live, singleton
births in WA between 1984 and 2006, analyses were conducted at CD level, and
maternal, paternal and community level variables were included in the analyses. Using
total population level administrative data ensures that all eligible births were included in
the analyses, and with minimal missing data (5%). Analysis at the CD level is a
particular strength as the community level SES measure is based on small areas
containing approximately 250 households, as opposed to the postcodes and larger
aggregations (containing up to 212 CDs) that are often used to represent communities,
but may mask SES variation among CDs and hence individual women as well. Further,
the inclusion of individual and community level socioeconomic variables in analyses is
also a strength, as a review of past research has shown that previous studies have not
controlled for SES measures at both levels in examining preterm births.
While administrative datasets provide population-level data, it often includes varying
amounts of missing data which can potentially lead to under-ascertainment. In
124
Australia, it has been acknowledged that Aboriginality is often missing or recorded
inconsistently across datasets (Australian Institute of Health and Welfare 2008; Mak &
Watkins 2008). The use of linked health data in this study overcomes this issue and
increases reliability by providing a means of identification from multiple data sources.
An additional limitation of using administration data is the lack of detailed information
available on certain variables. In the case of skill level, although parental job titles
recorded at birth were used to code skill level, the job title provides no information on
duties performed.
In conclusion, while the overall rates of preterm birth in Western Australia have
remained fairly static over the last two decades, the disparity between Aboriginal and
non-Aboriginal infants has increased and is now similar to inequalities seen 20 years
ago. These findings highlight a major public health issue that should be of great
concern, given the short and long-term morbidities and complications associated with
preterm birth. This study highlights the misleading nature of observing aggregated
rates of preterm births which mask inequalities, and may result in incorrect or
misguided implementation of policies and interventions that do not target the very
populations that need them.
125
CHAPTER EIGHT: SUMMARY OF FINDINGS AND FUTURE DIRECTIONS
The studies within this thesis were the first to examine the effects of a monetary
incentive (i.e. the Baby Bonus) on birth rates in Western Australia (WA) (Chapter Four),
demonstrate the importance of the inclusion of both individual and area-level indicators
of socioeconomic status (SES) in examining inequalities in perinatal outcomes
(Chapter Five) and to investigate social and racial inequalities in poor fetal growth
(Chapter Six) and preterm birth in WA (Chapter Seven). The unique contributions that
this thesis has made includes bridging a significant knowledge gap in understanding
trends in social and racial inequalities in birth rates and perinatal outcomes over the
last two decades in Western Australia. It has also contributed towards bridging
knowledge gaps in understanding the contributing factors to these inequalities.
Furthermore, it overcomes a major limitation of previous research by developing and
incorporating an individual-level indicator of socioeconomic status in multi-level models.
Finally, as all analyses were performed on population-level data using multi-level
information, the results have greater generalisability and policy implications for
intervention and prevention.
Findings revealed that in WA the birth rate increased significantly following the
introduction of the monetary incentive. Despite speculation that the increase would
occur primarily among teenagers and socially disadvantaged groups, the results
showed that birth rates increased among most maternal age groups, in both Aboriginal
and non-Aboriginal women, in all maternal residential locations and among all
community level socioeconomic groups.
126
Recoding of occupation titles to develop an individual-level (parental) measure of SES
and the inclusion of this variable in the analysis showed that there is considerable
variation in individual SES within area-level socioeconomic groups. This highlighted the
importance of considering both individual-level (parental) and area-level SES when
examining inequalities in birth outcomes.
Investigation of the trends in poor fetal growth revealed that there has been a steady
decline of approximately 0.7% per annum in the percentage of term singleton infants
with poor fetal growth over the last 23 years, before it plateaued at 15.7% from 2003
onwards. However, this trend masks significant variations among socioeconomic and
racial groups, as well as considerable increases in the disparities between Aboriginal
and non-Aboriginal infants. Results showed that the proportion of infants with poor fetal
growth rose with increasing levels of social disadvantage, with the difference between
the most and least disadvantaged groups increasing considerably over time, from 6%
in 1984-1988 (16%-22%) to 10% in 2004-2006 (13-23%). In addition, Aboriginal infants
were significantly more likely to be growth restricted than non-Aboriginal infants.
Investigation of the social disparity within Aboriginal and non-Aboriginal infants
revealed that inequalities between the least and most disadvantaged group increased
over time, particularly among non-Aboriginal infants.
Examination of inequalities in preterm birth showed significantly different results. While
the overall rate of preterm births in Western Australia has remained fairly static over the
last 23 years, the inequalities between Aboriginal and non-Aboriginal infants have
increased. Although a significant portion of the disparity between Aboriginal and non-
Aboriginal infants is attributable to parental socioeconomic and demographic
characteristics, the disparity persists even after adjustment of these factors.
Findings from the assessment of the Baby Bonus incentive, showing an increase in
overall birth rates across the State, highlight the usefulness of routinely collected data
127
to provide an insight into the impact of policy at a population-level. Apart from this study
and a recent analysis in NSW (Lain et al. 2009), there appears to be no information
available or comparable study undertaken by the Australian Government to assess the
impact of the Baby Bonus policy prior to or during its implementation since July 2004.
Despite the lack of data, the Government made a number of amendments in the way
the payment was made to the recipients, following a public outcry regarding the
potential increased birth rates among teenagers and socially disadvantaged groups.
These amendments included a change in July 2007 from a one-off payment to 13
fortnightly payments for parents under the age of 18 and families deemed to be
vulnerable19
, as well as further changes in January 2009, with the introduction of
means testing and compulsory fortnightly payments for all parents (Family Assistance
Office 2009). These changes appeared to be made solely on the basis of the
speculation that particular social groups were having children for the sole purpose of
receiving the monetary incentive, despite a lack of evidence. While the results from this
study clearly showed that there was no dramatic increase in birth rates among
teenagers or socially disadvantaged groups in WA, the stigmatisation of these groups
that occurred when the policy was first announced may have had negative
consequences for these groups of women.
One of the primary strengths of this study was the use of Collection District data to
measure area-level SES. Previously, most studies have used postcode or Statistical
Local Area (SLA) measures of SES (Lain et al. 2009). This is problematic as previous
Australian research has shown that almost 50% of residents are misclassified at
postcode level compared to CD level (Hyndman et al. 1995). Additionally, in the 2006
Census, there were 4370 CDs in WA, compared to only 249 SLAs. Hence, CD level
data allow a much more detailed analysis. It would also be of interest to encourage
more Australian states to undertake similar analyses at the CD level to compare
19 Vulnerable families included those who are known to have a history of domestic or family violence, substance addiction, intellectual disability, mental illness, are known to child protection, or have a known history of difficulties managing competing financial requirements.
128
findings so that the Government can be informed about similarities or differences of the
outcome of the Baby Bonus policy implementation. Furthermore, given the Australian
Government’s announcement to introduce means testing for the Baby Bonus policy
from January 1st
2009 (Commonwealth of Australia 2008) and the availability of only
two years data at the time of this study, it would be of considerable interest to conduct
a follow up study in the next few years to examine the effect this policy change may
have on the birth rate over the next few years.
Findings presented in Chapter Five provide evidence that when individual-level
socioeconomic measures are cross-classified with area-level SES measures there is
considerable variation within area-level SES groups, a finding which has also been
supported by several other smaller studies in other countries (Demissie et al. 2000;
Geronimus et al. 1996). Previous research conducted elsewhere has shown that
compared to estimates obtained using individual-level SES measures, when area-level
SES measures are used, health inequality estimates are often weaker (Walker &
Becker 2005; Steenland et al. 2004). The absence of individual-level SES data in
Australian administrative datasets has previously been acknowledged and the use of
area-level SES as a proxy has often been applied (Glover et al. 2004). While these
area-based measures are the best socioeconomic measures currently available and
are relatively easy to access, they understate socioeconomic disadvantage, particularly
when used at postcode or SLA level (Glover et al. 2004).
While in most research it is the SES of the individual with the outcome of interest that is
sought after, in child health research it is the SES of one or both parents that are of
greatest interest. To date, in Australia and WA, most information on birth outcomes is
obtained from routinely collected data from the Midwives’ Notification System and Birth
Registrations and the majority of researchers using these databases rely on the use of
SES measures at the area-level (based on CD, postcode or SLA), which are also often
used as a proxy for individual-level SES. The development of parental skill level based
129
on parental occupation titles in this thesis demonstrates how routinely collected Birth
Registration information can be used to obtain individual-level SES measures for
parents. The benefit of this is that by including both the individual-level (parental) and
area-level SES measures, researchers will be better positioned to separate contextual
effects from those that are due to individual-level (parental) socioeconomic factors.
This can potentially assist in identifying the level at which intervention or prevention
programmes should be targeted. Further, the findings from Chapter Five have shown
there are important interactions between individual and area-level SES. Results
revealed that while the rate of poor fetal growth increased with increasing levels of
disadvantage for all parental skill levels, the rate at which it increased was different
depending on which skill level was under consideration. Therefore, to improve perinatal
outcomes, it is imperative that policy and intervention programs are directed at both the
individual and community.
There are also several other key benefits of constructing an individual-level measure of
SES using the occupation titles from Birth Registrations. Firstly, the individual-level
SES obtained from birth registrations can be linked to the parent’s records using the
Western Australian Family Connections Genealogical project (Glasson et al. 2008).
This will provide individual-level SES measures for those parents, which in turn would
enable researchers to address a wider range of research questions on both child and
adult health. Secondly, the contributions of both maternal and paternal SES can be
examined separately, so that the impact of each can be assessed. Finally, the
availability of individual-level SES measures from routinely collected population-level
data will allow researchers to include both individual and area-level measures of SES
in studies which measure and monitor social and racial inequalities and evaluate the
effectiveness of future interventions to reduce health inequalities. The World Health
Organisations Commission on Social Determinants of Health final report calls for
actions to close the gap in a generation, and evaluation of interventions to reduce
health inequity (Commission on Social Determinants of Health 2008). Yet most recent
130
research shows that one major hurdle in monitoring and evaluating health equity
focused interventions is the lack of adequate information about socioeconomic
indicators (Judge 2009; Kavanagh et al. 2009; Petticrew et al. 2009), particularly at the
individual level in most routinely collected population databases. Chapter Five
demonstrates that this significant problem can be resolved in the case of the WA Data
Linkage System by utilising information on parental occupation titles, available in Birth
Registrations.
Chapters Six and Seven include the maternal and paternal SES measures obtained
from the occupation recoding in the examination of racial and social inequalities among
infants with poor fetal growth or born preterm. Inequalities between Aboriginal and non-
Aboriginal populations in Australia have been widely reported across a range of infant
outcomes (Freemantle et al. 2006b; Panaretto et al. 2002a; Najman et al. 1994; Eades
et al. 2008; Sayers & Powers 1997; Jonas et al. 1992). The inequalities are, at least in
part, due to social disadvantage (Morrison et al. 1989b), pathological factors and other
lifestyle risk factors such as smoking, alcohol and drug use during pregnancy (Blair
1996; Schultz et al. 1991) (Najman et al. 1994; Freemantle et al. 2006b; Panaretto et
al. 2002a; Elder et al. 1999). It has also been suggested that poorer attendance to
medical services, due to a lack of access to culturally appropriate services, may be a
contributing factor (Najman et al. 1994). Findings from this thesis reveal that even after
accounting for parental demographic and socioeconomic factors, including maternal
smoking, an Aboriginal infant is still 17% more likely to be growth restricted compared
with a non-Aboriginal infant. Further, an Aboriginal infant born to a low risk woman (i.e.
has no medical conditions and/or pregnancy complications) is about 50% more likely to
be born preterm compared with a non-Aboriginal infant born to a low risk woman. For
Aboriginal infants born to high risk women (i.e. has medical conditions and/or
pregnancy complications), they are 30% more likely to be born preterm than a non-
Aboriginal infant born to a high risk woman.
131
Investigation of racial inequalities between Aboriginal and non-Aboriginal infants in WA
since 1984 showed that disparities have persisted over time. In fact, in 2004-2006, the
odds of an Aboriginal infant being growth restricted or preterm, even when the mother
was considered to be low risk, was higher than they were 20 years ago. It is concerning
that these racial inequalities not only persist but also increased in the recent years.
Perhaps what is more alarming is that these trends occur despite the introduction of
programs specifically aimed to improve Aboriginal maternal and infant health, such as
the Aboriginal Primary Care Program (Health Reform Implementation Taskforce 2007)
and the Healthy for Life Program (Department of Health and Ageing 2005).
Furthermore, findings presented in Chapters Six and Seven demonstrate quite clearly
that while the adjustment for important parental demographic and social factors did not
completely eliminate the disparity, the inequality was significantly reduced. This
highlights the potential importance of future social and economic interventions in
reducing the racial inequality. Future studies should investigate which aspects of social
disadvantage are contributing the most to the increasing inequalities between
Aboriginal and non-Aboriginal populations so that policymakers can be provided with
more direction in their attempt to close the gap.
Social inequalities were first demonstrated in adult mortality in the Black Report
(Townsend et al. 1988) and Whitehall study (Marmot et al. 1984) and since then, have
been reported by many studies among a number of health and developmental
outcomes in early life (Morrison et al. 1989b; Carmichael & Williams 1983a; Jonas et
al. 1992; Fisher 1978; Quine 1991). For the purpose of the studies within this thesis,
social inequalities were examined separately for Aboriginal and non-Aboriginal infants
for three primary reasons. Firstly, findings regarding racial inequalities confirmed that
there were already considerable differences in birth outcomes between Aboriginal and
non-Aboriginal infants that persisted even after adjustment for a wide range of factors.
Secondly, births to Aboriginal women account for only 7-8% of all births annually in
WA. Therefore if the births were not analysed separately, important information
132
regarding social inequalities within Aboriginal infants would have been masked. Finally,
factors that contribute to the social gradient in health within the Aboriginal population
may be different from those for the non-Aboriginal population.
Indeed, examination of social inequalities of adverse birth outcomes, including poor
fetal growth and preterm birth, by Aboriginality showed that there were some important
differences in the social gradients between Aboriginal and non-Aboriginal infants. For
non-Aboriginal infants there was a distinct social gradient that persisted after
accounting for the same parental demographic and social factors. Investigation of
preterm birth revealed similar findings among low risk women. For high-risk women
there was a J-shaped curve that remained even after adjustment for all factors. This
suggests that other unmeasured factors could be contributing to the remaining social
inequalities in fetal growth, including access to and use of critical antenatal care which
is affected by social and economic policies of the Federal and State Governments and
maternal weight, which is inversely related to SES (Yogev & Catalano 2009; Linne
2004).
However, the results for Aboriginal infants were considerably different to those for non-
Aboriginal infants, with a much clearer and consistent social gradient in preterm birth,
particularly for low risk women, while there was a lack of a social gradient for poor fetal
growth. The disappearance of the social gradient for Aboriginal infants after adjusting
for parental demographic and social factors highlights that these factors contribute
substantially to social disparity within Aboriginal infants. Therefore, by improving these
potentially modifiable social factors, including social and economic policies, birth
outcomes for Aboriginal infants can be improved.
While there are many factors that are likely to have contributed to the increase in
inequalities demonstrated in this thesis, it is plausible that the neo-liberal policies
implemented by previous governments have also played a role. Since it has been well-
133
established that social and economic policies increase inequalities (Commission on
Social Determinants of Health 2008), it is reasonable to assume that they must also be
part of the solution. The rising trends in social and racial inequalities in perinatal health
outcomes over the last two decades in WA, as clearly demonstrated in this thesis, are
parallel to a global trend of increasing poor child health and wellbeing outcomes and
rising social inequalities, as described by Li, Murray and Stanley (Li et al. 2008). They
pose a serious question about the complacency of the neo-liberal governments with the
level of economic development solely measure in Gross Domestic Product.
A significant benefit of seeking policy change and implementation of intervention and
prevention programs is that change can occur simultaneously on all levels. At a
National level, changes can be made to social welfare, taxation and family and work
policy, including child care and paid parental leave. At a state level, changes can be
implemented for education and health services, including improving access to services
and ensuring culturally appropriate services are available. Additionally, for more
individual-level intervention or prevention programs, clinical guidelines could be
implemented for the recommendation of screening for smoking and alcohol in
pregnancy, with a focus on providing support and counselling for prevention.
In conclusion, this thesis highlights several important issues. Firstly, birth rates across
all maternal groups in WA increased significantly following the introduction of the
monetary incentive, in the form of the Baby Bonus policy. When compared to results
from the Baby Bonus study from NSW, the WA findings suggest that the Government’s
interventions to raise fertility did not have a uniform impact in all states. This indicates
that the impact of the policy may be modified by other factors, including the economic
climate and demographic characteristics of the state’s population, which are important
factors to consider for any relevant future social or economic policy. Secondly, this
research, particularly the results describing poor fetal growth and preterm birth,
highlights the misleading nature of aggregated rates of birth outcomes that is often the
134
only information provided to policymakers. The aggregated rates mask inequalities and
fail to shed light on which groups in society have improved and which are failing to do
so. Furthermore, the aggregated rates often fail to show that the improvement made in
birth outcomes mostly occurred in the least disadvantaged group, with more
disadvantaged groups making small or no improvements.
The findings provide strong evidence of the importance of examining and monitoring
trends in birth outcomes by social characteristics and appropriate subgroups. This
research shows that disparities between Aboriginal and non-Aboriginal infants have
increased and in some birth outcomes the disparity is greater now than it was 20 years
ago. These findings highlight a major public health issue that should be of great
concern, given the short and long-term adverse health and developmental outcomes
associated with both preterm birth and poor fetal growth. The reduction of inequalities
in these poor birth outcomes can only be prevented or reduced through adequate
monitoring and effective intervention.
This body of work has demonstrated how adequate monitoring can be achieved using
routinely collected administrative datasets, which can also be used to evaluate policy
implementations. Furthermore the reduction of racial inequalities will need to involve
active consultation with Aboriginal community groups and organisations so that
culturally appropriate services can be implemented and suitable and effective policy
recommendations can be made. Based on the findings detailed within this thesis and
the recommendations that emanate from these findings for future research and policy
intervention, vulnerable populations within our society can be reached, and the ultimate
goal of improving perinatal health and ensuring a healthy start to life for all infants can
be achieved.
135
REFERENCES
(UNICEF), U. N. C. s. F., Children and the Millennium Development Goals: Progress
towards a world fit for children [Online], Available:
http://www.unicef.org/worldfitforchildren/files/Children_and_the_MDGs_Final_EN.pdf
[20 March].
Acheson, D. S. 1998, Independent inquiry into inequalities in health report, Stationery
Office, London.
Acheson, E. D. 1964, 'The Oxford Record Linkage Study: a Review of the Method with
Some Preliminary Results', Proc R Soc Med, vol. 57, pp. 269-74.
Agresti, A. 2002, Categorical data analysis, 2nd edn, Wiley-Interscience, New York.
Ahern, J., Pickett, K. E., Selvin, S. & Abrams, B. 2003, 'Preterm birth among African
American and white women: A multilevel analysis of socioeconomic characteristics and
cigarette smoking', Journal of Epidemiology and Community Health, vol. 57, no. 8, p.
606.
Ahn, N. & Mira, P. 2002, 'A note on the changing relationship between fertility and
female employment rates in developed countries', Journal of Population Economics,
vol. 15, no. 4, p. 667.
AIHW: Leeds KL, G. M., Laws PJ, Zhang J, Al-Yaman F & Sullivan EA 2007,
Indigenous mothers and their babies, Australia 2001–2004. , PER 38, AIHW, Canberra.
Allison, P. 1999, Logistic regression using the SAS system : theory and application,
SAS Institute, c1999., Cary, N.C.
136
Ananth, C. V. P. M. P. H., Joseph, K. S. M. D. P., Oyelese, Y. M. D., Demissie, K. M.
D. P. & Vintzileos, A. M. M. D. 2005, 'Trends in Preterm Birth and Perinatal Mortality
Among Singletons: United States, 1989 Through 2000. [Article]', Obstetrics &
Gynecology, vol. 105, no. (5, Part 1), pp. 1084-1091.
Anonymous 2007, 'Birthrates in Europe tick upward after a generation of decline', Work
& Family Life, vol. 21, no. 10, p. 1.
Armstrong, B. K. & Kricker, A. 1999, 'Record linkage--a vision renewed', Australian and
New Zealand Journal of Public Health, vol. 23, no. 5, p. 451.
Astolfi, P., De Pasquale, A. & Zonta, L. A. 2006, 'Paternal Age and Preterm Birth in
Italy, 1990 to 1998. [Report]', Epidemiology, vol. 17, no. 2, pp. 218-221.
Australian Bureau of Statistics, ANZSCO First Edition to ISCO-88 Correspondence
Tables [Online], Available:
http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/1220.02006?OpenDocumen
t [18 June].
Australian Institute of Health and Welfare 2008, Aboriginal and Torres Strait Islander
Health Performance Framework 2008 report: Detailed analyses, IHW 22, AIHW,
Canberra.
Barker, D. J. 1998, 'In utero programming of chronic disease', Clin Sci (Lond), vol. 95,
no. 2, pp. 115-28.
Barker, D. J. P. 1995, 'The Wellcome Foundation Lecture, 1994. The Fetal Origins of
Adult Disease', Proceedings: Biological Sciences, vol. 262, no. 1363, pp. 37-43.
137
Baxter, J. 1994, 'Is husband's class enough? Class location and class identity in the
United States, Sweden, Norway, and Australia', American Sociological Review, vol. 59,
no. 2, p. 220.
Beard, J. R., Lincoln, D., Donoghue, D., Taylor, D., Summerhayes, R., Dunn, T. M.,
Earnest, A. & Morgan, G. 2009, 'Socioeconomic and maternal determinants of small-
for-gestational age births: patterns of increasing disparity', Acta Obstet Gynecol Scand,
vol. 88, no. 5, pp. 575-83.
Becker, G. S. 1964, Human capital : a theoretical and empirical analysis, with special
reference to education, National Bureau of Economic Research distributed by
Columbia University Press, New York.
Blair, E. 1996, 'Why do Aboriginal newborns weigh less? Determinants of birthweight
for gestation', Journal of Paediatrics and Child Health, vol. 32, no. 6, pp. 498-503.
Blair, E., Liu, Y., De Klerk, N. & Lawrence, D. 2005, 'Optimal growth for the Caucasian
singleton and assessment of appropriateness of fetal growth: an analysis of a total
population perinatal database', BMC Pediatrics, vol. 5.
Blakely, T., Ajwani, S., Robson, B., Tobias, M. & Bonne, M. 2004, 'Decades of
disparity: widening ethnic mortality gaps from 1980 to 1999', N Z Med J, vol. 117, no.
1199, p. U995.
Blakely, T., Tobias, M., Robson, B., Ajwani, S., Bonne, M. & Woodward, A. 2005,
'Widening ethnic mortality disparities in New Zealand 1981-99', Soc Sci Med, vol. 61,
no. 10, pp. 2233-51.
138
Boardman, J., D., Powers, D., A. , Padilla, Y., C. & Hummer, R., A. 2002, 'Low birth
weight, social factors, and developmental outcomes among children in the United
States', Demography, vol. 39, no. 2, p. 353.
Bohnert, K. M. B. A. & Breslau, N. P. 2008, 'Stability of Psychiatric Outcomes of Low
Birth Weight: A Longitudinal Investigation.' Archives of General Psychiatry, vol. 65, no.
9, pp. 1080-1086.
Bor, W., Najman, J. M., Andersen, M., Morrison, J. & Williams, G. 1993,
'Socioeconomic disadvantage and child morbidity: An Australian longitudinal study',
Social Science & Medicine, vol. 36, no. 8, pp. 1053-1061.
Bronfenbrenner, U. 1977, 'Toward an experimental ecology of human development',
American Psychologist, vol. 32, pp. 513-531.
Bronfenbrenner, U. & Ceci, S. J. 1994, 'Nature-Nurture Reconceptualized in
Developmental Perspective: A Bioecological Model,' Psychological Review, vol. 101,
no. 4, pp. 568-586.
Brooks-Gunn, J. 1995, 'Children in families in communities: Risk and intervention in the
Bronfenbrenner tradition', in Examining lives in context : perspectives on the ecology of
human development, 1st ed edn, eds P. Moen, G. H. Elder Jr & K. Luscher, American
Psychological Association,, Washington, DC :.
Brooks-Gunn, J., Brown, B., Duncan, G. J. & Anderson Moore, K. 1995, Child
development in the context of family and community resources: an agenda for national
data collections, The National Academy of Sciences, Washington DC.
139
Brooks-Gunn, J., Duncan, G. J. & Britto, P. B. 1999, 'Are socioeconomic gradients for
children similar to those for adults? Achievement and health of children in the United
States', in Developmental health and the wealth of nations : social, biological, and
educational dynamics, eds D. P. Keating & C. Hertzman, Guilford Press,, New York :,
p. 394.
Bundred, P., Manning, D., Brewster, B., Buchan, I. & Spencer, N. 2003, 'Social trends
in singleton births and birth weight in Wirral residents, 1990-2001 * Commentary', Arch.
Dis. Child. Fetal Neonatal Ed., vol. 88, no. 5, pp. F421-425.
Carmichael, A. & Williams, H. E. 1983a, 'Infant health and family functioning in a poor
socioeconomic status multi-ethnic municipality', Australian Paediatric Journal, vol. 19,
pp. 61-67.
Carmichael, A. & Williams, H. E. 1983b, 'Patterns of illness in an infant population in a
poor socioeconomic status multi-ethnic municipality', Australian Paediatric Journal, vol.
19, pp. 68-72.
Castles, I. 1994, Information Paper 1991 Census Socio-economic Indexes for Areas,
Australian Bureau of Statistics, Canberra.
Cerdá, M., Buka, S. L. & Rich-Edwards, J. W. 2008, 'Neighborhood influences on the
association between maternal age and birthweight: A multilevel investigation of age-
related disparities in health', Social Science & Medicine, vol. 66, no. 9, pp. 2048-2060.
Chalmers, J., Shanks, E., Stott, A. & Paton, C., Preterm deliveries in Scotland [Online],
Available: http://www.bmj.com/cgi/eletters/332/7547/937#132366 [May 7].
140
Chen, X.-K., Wen, S. W., Krewski, D., Fleming, N., Yang, Q. & Walker, M. C. 2008,
'Paternal age and adverse birth outcomes: teenager or 40+, who is at risk?' Hum.
Reprod., p. dem403.
Chomitz, V. R., Cheung, L. W. Y. & Lieberman, E. 1995, 'The Role of Lifestyle in
Preventing Low Birth Weight', The Future of Children, vol. 5, no. 1, pp. 121-138.
Choo, C. 1990, Aboriginal child poverty, Brotherhood of St. Laurence,, Melbourne :.
Coates, H. 2002, 'Chronic suppurative otitis media in indigenous populations: The
Australian Aborigine', ENT: Ear, Nose & Throat Journal, vol. 81, no. 8, pp. 11-12.
Coburn, D. 2000, 'Income inequality, social cohesion and the health status of
populations: the role of neo-liberalism', Social Science & Medicine, vol. 51, no. 1, pp.
135-146.
Coburn, D. 2004, 'Beyond the income inequality hypothesis: class, neo-liberalism, and
health inequalities', Social Science & Medicine, vol. 58, no. 1, pp. 41-56.
Coleman, J. S. 1988, 'Social Capital in the Creation of Human Capital', The American
Journal of Sociology, vol. 94, p. S95.
Coleman, J. S. 1990, Foundations of social theory / James S. Coleman, Belknap Press
of Harvard University Press, Cambridge, Mass :.
Collins, J. W., Jr. & David, R. J. 1992, 'Differences in neonatal mortality by race,
income, and prenatal care', Ethn Dis, vol. 2, no. 1, pp. 18-26.
141
Collins, J. W., Jr. & David, R. J. 2009, 'Racial disparity in low birth weight and infant
mortality', Clin Perinatol, vol. 36, no. 1, pp. 63-73.
Collins, J. W., Jr., David, R. J., Simon, D. M. & Prachand, N. G. 2007, 'Preterm birth
among African American and white women with a lifelong residence in high-income
Chicago neighborhoods: an exploratory study', Ethn Dis, vol. 17, no. 1, pp. 113-7.
Collins, J. W., Jr., David, R. J., Symons, R., Handler, A., Wall, S. & Andes, S. 1998,
'African-American Mothers' Perception of Their Residential Environment, Stressful Life
Events, and Very Low Birthweight', Epidemiology, vol. 9, no. 3, pp. 286-289.
Collins, J. W., Jr., David, R. J., Symons, R., Handler, A., Wall, S. N. & Dwyer, L. 2000,
'Low-Income African-American Mothers' Perception of Exposure to Racial
Discrimination and Infant Birth Weight', Epidemiology, vol. 11, no. 3, pp. 337-339.
Commission on Social Determinants of Health 2008, Closing the gap in a generation:
Health equity through action on the social determinants of health, World Health
Organisation, Geneva.
Commonwealth of Australia, Budget Measures Budget Paper No. 2 2008-09 [Online],
CanPrint Communications Pty Ltd, Available: www.budget.gov.au [9 December].
Conley, T. J., Globalisation and Rising Inequality in Australia: Is Increasing Inequality
Inevitable in Australia? [Online], University of Adelaide, [10 August].
Conrad, P. & Kern, R. (eds) 1981, The Sociology of Health and Illness: Critical
Perspectives, St Martin's Press, New York.
142
Corchia, C., Bertollini, R., Forastiere, F., Pistelli, R. & Perucci, C. 1995, 'Is Maternal
Asthma a Risk Factor for Low Birth Weight? Results of an Epidemiologic Survey',
European Journal of Epidemiology, vol. 11, no. 6, pp. 627-631.
Craig, E. D., Thompson, J. M. D. & Mitchell, E. A. 2002, 'Socioeconomic status and
preterm birth: New Zealand trends, 1980 to 1999', Archives of Disease in Childhood -
Fetal and Neonatal Edition, vol. 86, pp. F142-F146.
Crook, C. J. 1995, 'The role of mothers in the educational and status attainment of
Australian men and women', Australian and New Zealand Journal of Sociology, vol. 31,
no. 2, pp. 45-73.
Cynader, M. S. & Frost, B. J. 1999, 'Mechanisms of Brain Development: Neuronal
sculpting by the physical and social environment', in Developmental Health and the
Wealth of Nations: Social, biological, and education dynamics, eds D. P. Keating & C.
Hertzman, The Guildford Press, New York.
d'Addio, A. C. & d'Ercole, M. M. 2005, Trends and Determinants of Fertility Rates in
OECD Countries: The Role of Policies, Organisation for Economic Co-operation and
Development, Paris.
David, R. J. & Collins, J. W., Jr. 1991, 'Bad outcomes in black babies: race or racism?'
Ethn Dis, vol. 1, no. 3, pp. 236-44.
Davis, R. M. 2000, 'Healthy People 2010: Objectives for the United States: Impressive,
but unwieldy', British Medical Journal, vol. 320, no. 7238, p. 818.
143
Day, P., Sullivan, E., Ford, J. B. & Lancaster, P. 1999, Australia's mothers and babies
1997, PER 12, Australian Institute of Health and Welfare National Periatal Statistics
Unit, Sydney.
Demissie, K., Hanley, J. A., Menzies, D., Joseph, L. & Ernst, P. 2000, Agreement in
measuring socio-economic status: Area-based versus individual measures, vol. 21,
Chronic diseases in Canada, Public Health Agency of Canada.
Department of Families Housing Community Services and Indigenous Affairs, (2
January 2009), 2.4.1 Eligibility for Baby Bonus [Online], Available:
http://www.facsia.gov.au/Guides_Acts/fag/faguide-2/faguide-2.4/faguide-2.4.1.html [27
February 2009].
Department of Families Housing Community Services and Indigenous Affairs, (2
January 2009), 2.4.3 Baby Bonus Instalments [Online], Available:
http://www.facsia.gov.au/guides_acts/fag/faguide-2/faguide-2.4/faguide-2.4.3.html [27
February 2009].
Department of Health, (12 May 2009), Western Australian Data Linkage Information:
Statistics Summary [Online], Available: http://www.datalinkage-wa.org/go/data-
linkage/statistics-summary.
Department of Health, Western Australian Data Linkage Information [Online],
Available: http://www.datalinkage-wa.org/ [1 July].
Department of Health and Ageing, Health Budget 2005-2006: Healthy for Life program
[Online], Commonwealth of Australia, Available:
http://www.health.gov.au/internet/budget/publishing.nsf/Content/health-budget2005-
hbudget-hfact5.htm [14 April].
144
Dibben, C., Sigala, M. & Macfarlane, A. 2006, 'Area deprivation, individual factors and
low birth weight in England: is there evidence of an "area effect"?
10.1136/jech.2005.042853', J Epidemiol Community Health, vol. 60, no. 12, pp. 1053-
1059.
DiGiacomo, M., Davidson, P. M., Davison, J., Moore, L. & Abbott, P. 2007, 'Stressful
life events, resources, and access: key considerations in quitting smoking at an
Aboriginal Medical Service', Australian and New Zealand Journal of Public Health, vol.
31, no. 2, pp. 174-176.
Dominguez, T. P., Dunkel-Schetter, C., Glynn, L. M., Hobel, C. & Sandman, C. A.
2008, 'Racial Differences in Birth Outcomes: The Role of General, Pregnancy, and
Racism Stress', Health Psychology, vol. 27, no. 2, pp. 194-203.
Dorozynski, A. 2003, 'France offers E800 reward for each new baby', British Medical
Journal, vol. 326, p. 1002.
Doyal, L. 1979, The Political Economy of Health, Pluto Press, London.
Drago, R., Sawyer, K., Sheffler, K., Warren, D. & Wooden, M. 2009, Did Australia's
Baby Bonus increase the fertility rate? Working Paper No. 1/09, Melbourne Institute of
Applied Economic and Social Research, Melbourne.
Dubois, L. & Girard, M. 2006, 'Determinants of birthweight inequalities: Population-
based study', Pediatrics International, vol. 48, no. 5, pp. 470-478.
Duclos, É., Lefebvre, P. & Merrigan, P., A ‘Natural Experiment’ on the Economics of
Storks: Evidence on the Impact of Differential Family Policy on Fertility Rates in
145
Canada. CREFE Working Paper No. 136 [Online], Montreal: Université du Québec à
Montréal, Centre for Research on Economic Fluctuations and Employment., Available:
http://www.iussp.org/Brazil2001/s30/S30_02_Duclos.pdf [8 December].
Dunn, J. R., Veenstra, G. & Ross, N. 2006, 'Psychosocial and neo-material dimensions
of SES and health revisited: Predictors of self-rated health in a Canadian national
survey', Social Science & Medicine, vol. 62, no. 6, pp. 1465-1473.
Dunn, K. M., Jordan, K., Lacey, R. J., Shapley, M. & Jinks, C. 2004, 'Patterns of
Consent in Epidemiologic Research: Evidence from Over 25,000 Responders',
American Journal of Epidemiology, vol. 159, no. 11, p. 1087.
Eades, S., Read, A. W., Stanley, F. J., Eades, F. N., McCaullay, D. & Williamson, A.
2008, 'Bibbulung Gnarneep ('solid kid'): Causal pathways to poor birth outcomes in an
urban Aboriginal birth cohort', Journal of Paediatrics and Child Health, vol. 44, no. 6,
pp. 342-346.
Elder, D., Hagan, R., Evans, S. F., Benninger, H. R. & French, N. P. 1999, 'Hospital
admissions in the first year of life in very preterm infants', Journal of Paediatrics and
Child Health, vol. 35, no. 2, pp. 145-150.
Emerson, E. P. 2009, 'Relative Child Poverty, Income Inequality, Wealth, and Health.
[Editorial]', Journal of the American Medical Association, vol. 301, no. 4, pp. 425-6.
Escobar, G. J., McCormick, M. C., Zupancic, J. A. F., Coleman-Phox, K., Armstrong,
M. A., Greene, J. D., Eichenwald, E. C. & Richardson, D. K. 2006, 'Unstudied infants:
outcomes of moderately premature infants in the neonatal intensive care unit', Arch.
Dis. Child. Fetal Neonatal Ed., vol. 91, no. 4, pp. F238-244.
146
Evans, R. G. & Stoddart, G. L. 2003, 'Consuming research, producing policy?'
American Journal of Public Health, vol. 93, no. 3, p. 371.
Family Assistance Office, Changes to Baby Bonus from 1 January 2009 [Online],
Australian Government, Available:
http://www.facs.gov.au/internet/FAO/fao1.nsf/content/payments-maternity_payment-
baby_bonus_1jan2009.htm [11 August].
Fisher, S. 1978, 'Relationship of mortality to socioeconomic status and some other
factors in Sydney in 1971', Journal of Epidemiology & Community Health, vol. 32, pp.
41-46.
Fone, D. L. & Dunstan, F. 2006, 'Mental health, places and people: A multilevel
analysis of economic inactivity and social deprivation', Health & Place, vol. 12, no. 3,
pp. 332-344.
Freemantle, C. J., Read, A., deKlerk, N. & McAullay, D. 2006a, 'Constructing a
mortality profile of Western Australian born Aboriginal infants, children and young
people using total population linked data', Australasian Epidemiologist, vol. 13, no. 1,
pp. 9-14.
Freemantle, C. J., Read, A. W., de Klerk, N. H., McAullay, D., Anderson, I. P. &
Stanley, F. J. 2006b, 'Patterns, trends, and increasing disparities in mortality for
Aboriginal and non-Aboriginal infants born in Western Australia, 1980-2001: population
database study', Lancet, vol. 367, no. 9524, pp. 1758-1766.
Freemantle, C. J., Read, A. W., de Klerk, N. H., McAullay, D., Anderson, I. P. &
Stanley, F. J. 2006c, 'Sudden infant death syndrome and unascertainable deaths:
Trends and disparities among Aboriginal and non-Aboriginal infants born in Western
147
Australia from 1980 to 2001 inclusive', Journal of Paediatrics & Child Health, vol. 42,
no. 7/8, pp. 445-451.
Frohlich, K. L., Corin, E. & Potvin, L. 2001, 'A theoretical proposal for the relationship
between context and disease', Sociology of Health & Illness, vol. 23, no. 6, p. 776.
Frost, J. & Oslak, S. 1999, Teenagers’ pregnancy intentions and decisions: A study of
young women in California choosing to give birth, 2, Alan Guttmacher Institute New
York.
Garton, P. 2008, 'The resources boom and the two-speed economy', Economic
Roundup, no. 3.
Gauthier, A. H. 2000, Public policies affecting fertility and families in Europe: A survey
of the 15 member states in European Observatory on Family Matters; Annual Seminar
2000: 'Low Fertility, Families and Public Policies', Sevilla (Spain).
Gee, V. & Dawes, V. 1994, Validation study of the Western Australian Midwives'
Notification System 1992, Health Dept. of Western Australia, Perth [W.A.] :.
Genereux, M., Auger, N., Goneau, M. & Daniel, M. 2008, 'Neighbourhood
socioeconomic status, maternal education and adverse birth outcomes among mothers
living near highways', J Epidemiol Community Health, vol. 62, no. 8, pp. 695-700.
Geronimus, A. T., Bound, J. & Neidert, L. J. 1996, 'On the validity of using census
geocode characteristics to proxy individual socioeconomic characteristics', Journal of
the American Statistical Association, vol. 91, no. 434, p. 529.
148
Gielchinsky, Y. M. D., Mankuta, D. M. D., Rojansky, N. M. D., Laufer, N. M. D.,
Gielchinsky, I. B. & Ezra, Y. M. D. 2004, 'Perinatal Outcome of Pregnancies
Complicated by Placenta Accreta', Obstetrics & Gynecology September, vol. 104, no.
3, pp. 527-530.
Glasson, E. J., de Klerk, N. H., Bass, A. J., Rosman, D. L., Palmer, L. J. & Holman, C.
D. J. 2008, 'Cohort Profile: The Western Australian Family Connections Genealogical
Project', Int. J. Epidemiol., vol. 37, no. 1, pp. 30-35.
Glauert, R., Li, J., Hagemann, E., De Klerk, N., The Developmental Pathways Project
Team, Team;, T. D. L. U. & Stanley, F. 2008, Novel data linkages for child and youth
developmental pathways: Using agency data from Health, Education, Child Protection,
Disability Services and Justice in Western Australia, Telethon Institute for Child Health
Research.
Glover, J., Rosman, D. & Tennant, S. 2004, 'Unpacking analyses relying on area-
based data: are the assumptions supportable?' International Journal of Health
Geographics, vol. 3, no. 1, p. 30.
Goldenberg, R. L., Culhane, J. F., Iams, J. D. & Romero, R. 2008, 'Epidemiology and
causes of preterm birth', The Lancet, vol. 371, no. 9606, pp. 75-84.
Goldthorpe, J. H. 1983, 'Women and Class Analysis: In Defence of the Conventional
View', Sociology, vol. 17, no. 4, pp. 465-488.
Grant, J., Hoorens, S., Sivadasan, S., Van Het Loo, M., Davanzo, J., Hale, L. & Butz,
W. 2006, 'Trends in European fertility: should Europe try to increase its fertility
rate…or just manage the consequences?<sup>1</sup>', International Journal
of Andrology, vol. 29, no. 1, pp. 17-24.
149
Grant, J., Hoorens, S., Sivadasan, S., van het Loo, M., DaVanzo, J., Hale, L., Gibson,
S. & Butz, W., Low Fertility and Population Ageing: Causes, Consequences, and Policy
Options [Online], RAND Corporation, Available:
http://www.rand.org/pubs/monographs/2004/RAND_MG206.pdf [8 December ].
Gray, M. & Auld, A. 2000, Towards an Index of Relative Indigenous Socioeconomic
Disadvantage, No. 196/2000, Centre for Aboriginal Economic Policy Research
Publications, The Australian National University, Canberra.
Gray, M., Qu, L. & Weston, R. 2008a, Fertility and family policy in Australia, 41,
Australian Institute of Family Studies, Melbourne.
Gray, R., Bonellie, S. R., Chalmers, J., Greer, I., Jarvis, S. & Williams, C. 2008b,
'Social inequalities in preterm birth in Scotland 1980-2003: findings from an area-based
measure of deprivation', BJOG: An International Journal of Obstetrics & Gynaecology,
vol. 115, no. 1, pp. 82-90.
Hancock, H. 2007, 'Low birth weight in Aboriginal babies--A need for rethinking
Aboriginal women's pregnancies and birthing', Women and Birth, vol. 20, no. 2, pp. 77-
80.
Harper, P. 2006, Births, 2005, Australian Bureau of Statistics, Canberra.
Haveman, R. & Wolfe, B. 1994, Succeeding generations : on the effects of investments
in children, Russell Sage Foundation, New York
150
Health Reform Implementation Taskforce, Western Australian Aboriginal Primary Care
Resource Kit [Online], Office of Aboriginal Health, Department of Health Available:
http://www.aboriginal.health.wa.gov.au/programs/apcap.cfm [14 April].
Heaman, M. I., Blanchard, J. F., Gupton, A. L., Moffatt, M. E. K. & Currie, R. F. 2005,
'Risk factors for spontaneous preterm birth among Aboriginal and non-Aboriginal
women in Manitoba', Paediatric & Perinatal Epidemiology, vol. 19, no. 3, pp. 181-193.
Hearst, M. O., Oakes, J. M. & Johnson, P. J. 2008, 'The Effect of Racial Residential
Segregation on Black Infant Mortality', Am. J. Epidemiol., vol. 168, no. 11, pp. 1247-
1254.
Heck, K., E. , Schoendorf, K., C. , Ventura, S., J. & Kiely, J., L. 1997, 'Delayed
Childbearing by Education Level in the United States, 1969-1994', Maternal and Child
Health Journal, vol. 1, no. 2, p. 81.
Hertzman, C. 1999, 'Population health and human development', in Developmental
health and the wealth of nations : social, biological, and educational dynamics, eds D.
P. Keating & C. Hertzman, Guilford Press,, New York :, p. 394.
Hertzman, C. 2002, Leave no child behind! : social exclusion and child development,
The Laidlaw Foundation, Toronto.
Hertzman, C. & Wiens, M. 1996, 'Child development and long-term outcomes: A
population health perspective and summary of successful interventions', Social
Science & Medicine, vol. 43, no. 7, pp. 1083-1095.
Hilbe, J. M. 2007, Negative Binomial Regression, Cambridge University Press,
Cambridge.
151
Hill, C. 1987, Validation study of the Western Australian Midwives' Notification System
1986 Health Dept. of Western Australia, Perth.
Hill, K., Barker, B. & Vos, T. 2007, 'Excess Indigenous mortality: are Indigenous
Australians more severely disadvantaged than other Indigenous populations?' Int. J.
Epidemiol., vol. 36, no. 3, pp. 580-589.
Hitchcock, N., McGuiness, D. & Gracey, M. 1982, 'Growth and feeding practices of
Western Australian infants.' The Medical Journal of Australia, vol. 1, pp. 372-376.
Hobbs, M. S. T. & McCall, M. G. 1970, 'Health statistics and record linkage in
Australia', Journal of Chronic Diseases, vol. 23, no. 5-6, pp. 375-381.
Holman, C. D., Bass, A. J., Rouse, I. L. & Hobbs, M. S. 1999, 'Population-based
linkage of health records in Western Australia: development of a health services
research linked database', Aust N Z J Public Health, vol. 23, pp. 453 - 459.
Holman, C. D. A. J. 2001, 'The impracticable nature of consent for research use of
linked administrative health records', Australian and New Zealand Journal of Public
Health, vol. 25, no. 5, pp. 421-422.
House, J. S., Lepkowski, J. M., Kinney, A. M., Mero, R. P., Kessler, R. C. & Herzog, A.
R. 1994, 'The Social Stratification of Aging and Health', Journal of Health and Social
Behavior, vol. 35, no. 3, pp. 213-234.
Howard, M. C. & King, J. E. 2008, 'Reflections on the long Australian boom [Paper in
special issue: The Australian Economic Boom: 1992-?.]', Journal of Australian Political
Economy, no. 61, pp. 56-70.
152
Huddy, C. L. J., Johnson, A. & Hope, P. L. 2001, 'Educational and behavioural
problems in babies of 32-35 weeks gestation', Arch. Dis. Child. Fetal Neonatal Ed., vol.
85, no. 1, pp. F23-28.
Hughes, D. & Simpson, L. 1995, 'The Role of Social Change in Preventing Low Birth
Weight', The Future of Children, vol. 5, no. 1, pp. 87-102.
Hyndman, J., Holman, C., Hockey, R., Donovan, R., Corti, B. & Rivera, J. 1995,
'Misclassification of social disadvantage based on geographical areas: comparison of
postcode and collector's district analyses', Int. J. Epidemiol., vol. 24, no. 1, pp. 165-
176.
Jatrana, S. & Blakely, T. 2008, 'Ethnic inequalities in mortality among the elderly in
New Zealand', Aust N Z J Public Health, vol. 32, no. 5, pp. 437-43.
Jolliffe, I. T. 1986, Principal component analysis, Springer-Verlag, New York ; Berlin :.
Jonas, O., Roder, D. & Chan, A. 1992, 'The association of maternal and socioeconomic
characteristics in metropolitan Adelaide with medical, obstetric and labour
complications and pregnancy outcomes', The Australian and New Zealand Journal of
Obstetrics and Gynaecology, vol. 32, pp. 1-5.
Joseph, K. S., Kramer, M. S., Marcoux, S., Ohlsson, A., Wen, S. W., Allen, A. & Platt,
R. 1998, 'Determinants of Preterm Birth Rates in Canada from 1981 through 1983 and
from 1992 through 1994', N Engl J Med, vol. 339, no. 20, pp. 1434-1439.
153
Judge, K. 2009, 'Inequalities in infant mortality: Patterns, trends, policy responses and
emerging issues in Canada, Chile, Sweden and the United Kingdom', Health Sociology
Review, vol. 18, no. 1, pp. 12-24.
Kalff, A., Kroes, M., Vles, J., Hendriksen, J., Feron, F., Steyaert, J., van Zeben, Jolles,
J. & van, O. 2001, 'Neighbourhood level and individual level SES effects on child
problem behaviour: a multilevel analysis.' Journal of Epidemiology & Community Health
[NLM - MEDLINE], vol. 55, no. 4, p. 246.
Kavanagh, J., Oliver, S., Lorenc, T., Caird, J., Tucker, H., Harden, A., Greaves, A.,
Thomas, J. & Oakley, A. 2009, 'School-based cognitive-behavioural interventions: A
systematic review of effects and inequalities', Health Sociology Review, vol. 18, no. 1,
pp. 61-78.
Kawachi, I. 2002, 'Income inequality and economic residential segregation', Journal of
Epidemiology and Community Health, vol. 56, no. 3, p. 165.
Kawachi, I. & Kennedy, B. P. 1997, 'Socioeconomic determinants of health : Health
and social cohesion: why care about income inequality?' BMJ, vol. 314, no. 7086, pp.
1037-.
Kawachi, I., Kennedy, B. P., Lochner, K. & Prothrow-Stith, D. 1997, 'Social capital,
income inequality, and mortality', American Journal of Public Health, vol. 87, no. 9, p.
1491.
Keating, D. P. & Hertzman, C. (eds) 1999, Developmental health and the wealth of
nations : social, biological, and educational dynamics, Guilford Press,, New York :.
154
Keating, D. P. & Miller, F. K. 1999a, 'Individual pathways in competence and coping', in
Developmental health and the wealth of nations : social, biological, and educational
dynamics, eds D. P. Keating & C. Hertzman, Guilford Press,, New York :, p. 394.
Keating, D. P. & Miller, F. K. 1999b, 'Individual Pathways in Competence and Coping:
From regulatory systems to habits of mind', in Developmental Health and the Wealth of
Nations: Social, biological, and education dynamics, eds D. P. Keating & C. Hertzman,
The Guildford Press, New York.
Kelman, C. W., Bass, A. J. & Holman, C. D. A. J. 2002a, 'Research use of linked health
data: a best practice protocol', Australian and New Zealand Journal of Public Health.
Kelman, C. W., Bass, A. J. & Holman, C. D. J. 2002b, 'Research use of linked health
data - a best practice protocol', Australian and New Zealand Journal of Public Health,
vol. 26, no. 3, p. 251.
Kennedy, B. P., Kawachi, I. & Prothrow-Stith, D. 1996, 'Income distribution and
mortality: Cross sectional ecological study of the Robin Hood index in the United
States', British Medical Journal, vol. 312, no. 7037, p. 1004.
Kliewer EV & FJ., S. 1989, 'Aboriginal and white births in Western Australia, 1980-
1986. Part I: Birthweight and gestational age.' Medical Journal of Australia, vol. 151,
no. 9, pp. 493-502.
Korupp, S. E., Ganzeboom, H. B. G. & van der Lippe, T. 2002, 'Do Mothers Matter? A
Comparison of Models of the Influence of Mothers' and Fathers' Educational and
Occupational Status on Children's Educational Attainment', Quality & Quantity, vol. 36,
no. 1, pp. 17-42.
155
Kramer, M. S. 1987, 'Determinants of low birth weight: methodological assessment and
meta-analysis', Bulletin of the World Health Organization, vol. 65, no. 5, pp. 663-737.
Kramer, M. S., Morin, I., Yang, H., Platt, R. W., Usher, R., McNamara, H., Joseph, K.
S. & Wen, S. W. 2002, 'Why are babies getting bigger? Temporal trends in fetal growth
and its determinants', The Journal of Pediatrics, vol. 141, no. 4, pp. 538-542.
Kramer, M. S., Seguin, L., Lydon, J. & Goulet, L. 2000a, 'Socio-economic disparities in
pregnancy outcome: why do the poor fare so poorly?' Paediatr Perinat Epidemiol, vol.
14, no. 3, pp. 194-210.
Kramer, M. S., Séguin, L., Lydon, J. & Goulet, L. 2000b, 'Socio-economic disparities in
pregnancy outcome: why do the poor fare so poorly?' Paediatric & Perinatal
Epidemiology, vol. 14, no. 3, pp. 194-210.
Kravdal, O. & Rindfuss, R. R. 2008, 'Changing Relationships between Education and
Fertility: A Study of Women and Men Born 1940 to 1964', American Sociological
Review, vol. 73, pp. 854-873.
Krieger, N. 1994, 'Epidemiology and the web of causation: Has anyone seen the
spider?' Social Science & Medicine, vol. 39, no. 7, pp. 887-903.
Krieger, N. 2001, 'Theories for social epidemiology in the 21st century: an ecosocial
perspective', Int. J. Epidemiol., vol. 30, no. 4, pp. 668-677.
Kuritz, S. J., Landis, J. R. & Koch, G. G. 1988, 'A General Overview of Mantel-
Haenszel Methods: Applications and Recent Developments', Annual Review of Public
Health, vol. 9, no. 1, pp. 123-160.
156
Lain, S. J., Ford, J. B., Raynes-Greenow, C. H., Hadfield, R. M., Simpson, J. M.,
Morris, J. M. & Roberts, C. L. 2009, 'The impact of the Baby Bonus payment in New
South Wales: who is having "one of the country"?' Medical Journal of Australia, vol.
190, no. 5, pp. 238-241.
Langhoff-Roos, J., Kesmodel, U., Jacobsson, B., Rasmussen, S. & Vogel, I. 2006,
'Spontaneous preterm delivery in primiparous women at low risk in Denmark:
population based study', BMJ, vol. 332, no. 7547, pp. 937-939.
Langille, D. 2004, 'The political determinants of health', in Social determinants of
health: Canadian perspectives, ed. D. Raphael, Canadian Scholar's Press,, Toronto :.
Lattimore, R. & Pobke, C. 2008, Recent Trends in Australian Fertility, Productivity
Commission Staff Working Paper, Canberra.
Lawson, K. & Tulloch, M. 1995, 'Breastfeeding duration: Prenatal intentions and
postnatal practices', Journal of Advanced Nursing, vol. 22, pp. 841-849.
Leonard, H., Nassar, N., Bourke, J., Blair, E., Mulroy, S., De Klerk, N. & Bower, C.
2008, 'Relation between Intrauterine Growth and Subsequent Intellectual Disability in a
Ten-year Population Cohort of Children in Western Australia', American Journal of
Epidemiology, vol. 167, no. 1, p. 103.
Leonard, H., Petterson, B., De Klerk, N., Zubrick, S. R., Glasson, E., Sanders, R. &
Bower, C. 2005, 'Association of sociodemographic characteristics of children with
intellectual disability in Western Australia', Social Science & Medicine, vol. 60, no. 7,
pp. 1499-1513.
157
Lesley, F. 2005, 'Changing patterns of inequality in birthweight and its determinants: a
population-based study, Scotland 1980-2000', Paediatric & Perinatal Epidemiology, vol.
19, no. 5, pp. 342-351.
Li, J., Mattes, E., Stanley, F., McMurray, A. & Hertzman, C. 2009, 'Editorial: Social
determinants of child health and well-being', Health Sociology Review, vol. 18, no. 1,
pp. 3-11.
Li, J., McMurray, A. & Stanley, F. 2008, 'Modernity's paradox and the structural
determinants of child health and well-being', Health Sociology Review, vol. 17, no. 1,
pp. 64-77.
Linacre, S. 2007, Article: One-parent families, Catalogue No. 4102.0, Australian
Bureau of Statistics, Canberra.
Link, B. G. & Phelan, J. 1995, 'Social conditions as fundamental causes of disease',
Journal of Health and Social Behavior, p. 80.
Link, B. G. & Phelan, J. C. 1996, 'Editorial: Understanding sociodemographic
differences in health--the role of fundamental social causes', American Journal of
Public Health, vol. 86, no. 4, p. 471.
Linne, Y. 2004, 'Effects of obesity on women's reproduction and complications during
pregnancy', Obes Rev, vol. 5, no. 3, pp. 137-43.
Lumley, J. 2003, 'Defining the problem: the epidemiology of preterm birth', BJOG: An
International Journal of Obstetrics & Gynaecology, vol. 110, no. s20, pp. 3-7.
158
Luo, Z.-C., Wilkins, R. & Kramer, M. S. 2006, 'Effect of neighbourhood income and
maternal education on birth outcomes: a population-based study', Canadian Medical
Association. Journal, vol. 174, no. 10, p. 1415.
Lynch, J., Davey Smith, G., Harper, S. & Hillemeier, M. 2004, 'Is Income Inequality a
Determinant of Population Health? Part 2. U.S. National and Regional Trends in
Income Inequality and Age- and Cause-Specific Mortality.' Milbank Quarterly, vol. 82,
no. 2, pp. 355-400.
Lynch, J. W., Davey Smith, G., Kaplan, G. A. & House, J. S. 2000, 'Income inequality
and mortality: Importance to health of individual income, psychosocial environment, or
material conditions', British Medical Journal, vol. 320, no. 7243, p. 1200.
Macintyre, S. 1997, 'The black report and beyond what are the issues?' Social Science
& Medicine, vol. 44, no. 6, pp. 723-745.
Macintyre, S. & Ellaway, A. 1999, 'Local opportunity structures, social capital and social
inequalities in health: what can central and local governments do?' Health Promotion
Journal of Australia, vol. 9, no. 3, pp. 165-70.
Macintyre, S. & Ellaway, A. 2003, 'Neighborhoods and health: An overview', in
Neighborhoods and health, eds I. Kawachi & L. F. Berkman, Oxford University Press,
New York.
Macintyre, S., Maciver, S. & Sooman, A. 1993, 'Area, class and health: Should we be
focusing on places or people?' Journal of Social Policy, vol. 22, pp. 213-233.
159
Macintyre, S., McKay, L. & Der, G. 2003, 'Socio-economic position and health: what
you observe depends on how you measure it', Journal of Public Health Medicine, vol.
25, no. 4, p. 288.
Mak, D. & Watkins, R. 2008, 'Improving the accuracy of Aboriginal and non-Aboriginal
disease notification rates using data linkage', BMC Health Services Research, vol. 8,
no. 1, p. 118.
Manderbacka, K., Teperi, J. & et al. 1992, 'Marital Status as a Predictor of Perinatal
Outcome in Finland', Journal of Marriage and the Family, vol. 54, no. 3, p. 508.
Marcin, J. P., Schembri, M. S., He, J. & Romano, P. S. 2003, 'A population-based
analysis of socioeconomic status and insurance status and their relationship with
pediatric trauma hospitalization and mortality rates', American Journal of Public Health,
vol. 93, no. 3, p. 461.
Margolis, P. A., Greenberg, R. A., Keyes, L. L., LaVange, L. M., Chapman, R. S.,
Denny, F. W., Bauman, K. E. & Boat, B. W. 1992, 'Lower respiratory illness in infants
and low socioeconomic status', Am J Public Health, vol. 82, no. 8, pp. 1119-26.
Marks, G. 2008, 'Are Father’s or Mother’s Socioeconomic Characteristics More
Important Influences on Student Performance? Recent International Evidence', Social
Indicators Research, vol. 85, no. 2, pp. 293-309.
Marmot, M. & Wilkinson, R. G. (eds) 1999, Social Determinants of Health, Oxford
University Press, Oxford.
Marmot, M. G., Davey Smith, G. & al, e. 1991, 'Health Inequalities Among British Civil
Servants: The Whitehall II Study', The Lancet, vol. 337, no. 8754, p. 1387.
160
Marmot, M. G. & McDowall, M. E. 1986, 'Mortality decline and widening social
inequalities', The Lancet, vol. 328, no. 8501, pp. 274-276.
Marmot, M. G., Shipley, M. J. & Rose, G. 1984, 'Inequalities in Death--Specific
Explanations of a General Pattern?' The Lancet, vol. 323, no. 8384, pp. 1003-1006.
Massey, D. S. & Denton, N. A. 1989, 'Hypersegregation in U.S. metropolitan areas:
black and Hispanic segregation along five dimensions', Demography, vol. 26, no. 3, pp.
373-91.
Massey, D. S. & Denton, N. A. 1993, American apartheid : segregation and the making
of the underclass, Harvard University Press, Cambridge, Mass. :.
McConnochie, K. M., Russo, M. J., McBride, J. T., Szilagyi, P. G., Brooks, A. M. &
Roghmann, K. J. 1999, 'Socioeconomic variation in asthma hospitalization: excess
utilization or greater need?' Pediatrics, vol. 103, no. 6, p. e75.
McCormick, M. C. 1985, 'The contribution of low birth weight to infant mortality and
childhood morbidity', N Engl J Med, vol. 312, no. 2, pp. 82-90.
McCracken, K. W. J. 2001, 'Into a SEIFA [socio-economic indexes for areas] SES
[socio-economic status] cul-de-sac?' Australian and New Zealand Journal of Public
Health, vol. 25, no. 4, pp. 305-306.
McCullagh, P. & Nelder, J. A. 1989, Generalized Linear Models, Chapman & Hall,
London.
161
McLennan, W. 1990, Socio-economic Indexes for Areas Information Paper 1986,
Australian Bureau of Statistics, Canberra.
McLennan, W. 1998, Information Paper 1996 Census of Population and Housing
Socio-economic Indexes for Areas, Australian Bureau of Statistics, Canberra.
Merrick, E. N. 1995, 'Adolescent childbearing as career "choice": Perspective from an
ecological context', Journal of Counseling and Development : JCD, vol. 73, no. 3, p.
288.
Messer, L. C., Kaufman, J. S., Mendola, P. & Laraia, B. A. 2008, 'Black-White Preterm
Birth Disparity: A Marker of Inequality', Annals of Epidemiology, vol. 18, no. 11, pp.
851-858.
Milligan, K. 2002, Quebec's Baby Bonus: Can Public Policy Raise Fertility?, C.D. Howe
Institute, Toronto, Ontario.
Milligan, K. 2005, 'Subsidizing the Stork: New Evidence on Tax Incentives and Fertility',
Review of Economics & Statistics, vol. 87, no. 3, pp. 539-555.
Mohsin, M., Wong, F., Bauman, A. & Bai, J. U. N. 2003, 'Maternal and Neonatal
Factors Influencing Premature Birth and Low Birth Weight in Australia', Journal of
Biosocial Science, vol. 35, no. 02, pp. 161-174.
Morgen, C. S., Bjork, C., Andersen, P. K., Mortensen, L. H. & Nybo Andersen, A.-M.
2008, 'Socioeconomic position and the risk of preterm birth--a study within the Danish
National Birth Cohort', Int. J. Epidemiol., vol. 37, no. 5, pp. 1109-1120.
162
Morrison, J., Najman, J. M., Williams, G. M., Keeping, J. D. & Andersen, M. J. 1989a,
'Socio-economic Status and pregnancy outcome. An Australian study', BJOG: An
International Journal of Obstetrics & Gynaecology, vol. 96, no. 3, pp. 298-307.
Morrison, J., Najman, J. M., Williams, G. M., Keeping, J. D. & Anderson, M. J. 1989b,
'Socio-economic status and pregnancy outcome. An Australian study.' British Journal of
Obstetrics and Gynaecology, vol. 96, no. 3, pp. 298-307.
Mortensen, L. H., Diderichsen, F., Arntzen, A., Gissler, M., Cnattingius, S., Schnor, O.,
Davey-Smith, G. & Nybo Andersen, A. M. 2008, 'Social inequality in fetal growth: a
comparative study of Denmark, Finland, Norway and Sweden in the period 1981-2000',
J Epidemiol Community Health, vol. 62, no. 4, pp. 325-331.
Moser, K., Li, L. & Power, C. 2003, 'Social inequalities in low birth weight in England
and Wales: trends and implications for future population health', J Epidemiol
Community Health, vol. 57, no. 9, pp. 687-691.
Mullen, P. D. 1999, 'Maternal smoking during pregnancy and evidence-based
intervention to promote cessation', Primary Care: Clinics in Office Practice, vol. 26, no.
3, pp. 577-589.
Naess, O., Claussen, B. & Davey Smith, G. 2004, 'Relative impact of childhood and
adulthood socioeconomic conditions on cause specific mortality in men', J Epidemiol
Community Health, vol. 58, no. 7, pp. 597-598.
Naess, O., Strand, B. H. & Smith, G. D. 2007, 'Childhood and adulthood
socioeconomic position across 20 causes of death: a prospective cohort study of 800
000 Norwegian men and women', J Epidemiol Community Health, vol. 61, no. 11, pp.
1004-1009.
163
Najman, J. M., Bor, W., Morrison, J., Andersen, M. & Williams, G. 1992, 'Child
developmental delay and socio-economic disadvantage in Australia: A longitudinal
study', Social Science & Medicine, vol. 34, no. 8, pp. 829-835.
Najman, J. M., Williams, G. M., Bor, W., Andersen, M. J. & Morrison, J. 1994,
'Obstetrical outcomes of Aboriginal pregnancies at a major urban hospital', Australian
Journal Of Public Health, vol. 18, no. 2, pp. 185-189.
National Health and Medical Research Council 2000, Guidelines Under Section 95 of
the Privacy Act 1988, Commonwealth of Australia, AusInfo.
National Health and Medical Research Council 2007, National Statement on Ethical
Conduct in Research Involving Humans, Commonwealth of Australia, AusInfo.
National Public Health Partnership 2004, Healthy Children - Strengthening promotion
and prevention across Australia: Developing a National Public Health Action Plan for
Children 2005-2008 - Consultation Paper, NPHP, Melbourne.
Navarro, V. 1990, 'Race or Class versus Race and Class: Mortality Differentials in the
United States', The Lancet, vol. 336, no. 8725, p. 1238.
Navarro, V. 2009, 'What we mean by the social determinants of health', Interational
Journal of Health Services, vol. 39, no. 3, pp. 423-441.
Newcombe, H. B., Fair, M. E. & Lalonde, P. 1992, 'The Use of Names for Linking
Personal Records', Journal of the American Statistical Association, vol. 87, no. 420, pp.
1193-1204.
164
O'Campo, P., Burke, J., G. , Culhane, J., Elo, I., T. , Eyster, J., Holzman, C., Messer,
L., C. , Kaufman, J., S. & Laraia, B., A. 2008, 'Neighborhood Deprivation and Preterm
Birth among Non-Hispanic Black and White Women in Eight Geographic Areas in the
United States', American Journal of Epidemiology, vol. 167, no. 2, p. 155.
Olsén, P., Läärä, E., Rantakallio, P., Järvelin, M.-R., Sarpola, A. & Hartikainen, A.-L.
1995, 'Epidemiology of Preterm Delivery in Two Birth Cohorts with an Interval of 20
Years', Am. J. Epidemiol., vol. 142, no. 11, pp. 1184-1193.
Ozanne, S. E., Fernandez-Twinn, D. & Hales, C. N. 2004, 'Fetal growth and adult
diseases', Seminars in Perinatology, vol. 28, no. 1, pp. 81-87.
Palomar, L., DeFranco, E. A., Lee, K. A., Allsworth, J. E. & Muglia, L. J. 2007, 'Paternal
race is a risk factor for preterm birth', American Journal of Obstetrics and Gynecology,
vol. 197, no. 2, pp. 152.e1-152.e7.
Panaretto, K., Lee, H., Mitchell, M., Larkins, S., Manessis, V., Buettner, P. G. &
Watson, D. 2006, 'Risk factors for preterm, low birth weight and small for gestational
age birth in urban Aboriginal and Torres Strait Islander women in Townsville',
Australian and New Zealand Journal of Public Health, vol. 30, no. 2, pp. 163-170.
Panaretto, K., Muller, R., Patole, S., Watson, D. & Whitehall, J. S. 2002a, 'Is being
Aboriginal or Torres Strait Islander a risk factor for poor neonatal outcome in a tertiary
referral unit in north Queensland?' Journal of Paediatrics and Child Health, vol. 38, no.
1, pp. 16-22.
Panaretto, K., S. , Mitchell, M., R. , Anderson, L., Larkins, S., L. , Buettner, P. G. &
Watson, D. 2007, 'Sustainable antenatal care services in an urban Indigenous
165
community: the Townsville experience', Medical Journal of Australia, vol. 187, no. 1, p.
18.
Panaretto, K. S., Lee, H. M., Mitchell, M., R. , Larkins, S., L. , Manessis, V., Buettner,
P. G. & Watson, D. 2005, 'Impact of a collaborative shared antenatal care program for
urban Indigenous women: a prospective cohort study', Medical Journal of Australia, vol.
182, no. 10, p. 514.
Panaretto, K. S., Muller, R., Patole, S., Watson, D. & Whitehall, J. S. 2002b, 'Is being
Aboriginal or Torres Strait Islander a risk factor for poor neonatal outcome in a tertiary
referral unit in north Queensland?' Journal of Paediatrics & Child Health, vol. 38, no. 1,
pp. 16-22.
Parker JD, Schoendorf KC & JL., K. 1994, 'Associations between measures of
socioeconomic status and low birth weight, small for gestational age, and premature
delivery in the United States.' Annals of Epidemiology, vol. 4, no. 4, pp. 271-8.
Parker, J. D. & Schoendorf, K. 1992, 'Influence of Paternal Characteristics on the Risk
of Low Birth Weight', Am. J. Epidemiol., vol. 136, no. 4, pp. 399-407.
Perry, I. J. & Lumey, L. H. 2004, 'Fetal growth and development: the role of nutrition
and other factors', in A life course approach to chronic disease epidemiology, eds D.
Kuh & Y. Ben-Shlomo, Oxford University Press, Oxford.
Petersen, C. B., Mortensen, L. H., Morgen, C. S., Madsen, M., Schnor, O., Arntzen, A.,
Gissler, M., Cnattingius, S. & Nybo Andersen, A. M. 2009, 'Socio-economic inequality
in preterm birth: a comparative study of the Nordic countries from 1981 to 2000',
Paediatric and Perinatal Epidemiology, vol. 23, no. 1, pp. 66-75.
166
Petrou, S., Kupek, E., Hockley, C. & Goldacre, M. 2006, 'Social class inequalities in
childhood mortality and morbidity in an English population', Paediatric and Perinatal
Epidemiology, vol. 20, no. 1, pp. 14-23.
Petrou, S., Sach, T. & Davidson, L. 2001, 'The long-term costs of preterm birth and low
birth weight: results of a systematic review', Child: Care, Health and Development, vol.
27, no. 2, pp. 97-115.
Petticrew, M., Tugwell, P., Welch, V., Ueffing, E., Kristjansson, E., Armstrong, R.,
Doyle, J. & Waters, E. 2009, 'Better evidence about wicked issues in tackling health
inequities', J Public Health, p. fdp076.
Pickett, K. E. & Wilkinson, R. G. 2007, 'Child wellbeing and income inequality in rich
societies: ecological cross sectional study', BMJ, vol. 335, no. 7629, pp. 1080-.
Pink, B. 2008a, Births, 2007, Australian Bureau of Statistics, Canberra.
Pink, B. 2008b, Information Paper An Introduction to Socio-Economic Indexes for
Areas (SEIFA) 2006, Australian Bureau of Statistics (ABS), Canberra.
Pink, B. 2008c, Socio-Economic Indexes for Areas (SEIFA) - Technical Paper 2006,
Australian Bureau of Statistics (ABS), Canberra.
Power, C., Hertzman, C., Matthews, S. & Manor, O. 1997, 'Social differences in health:
Life-cycle effects between ages 23 and 33 in the 1958 British birth cohort', American
Journal of Public Health, vol. 87, no. 9, p. 1499.
Quesnel-Vallee, A. & Morgan, S. P. 2003, 'Missing the Target? Correspondence of
Fertility Intentions and Behavior in the U.S', Population Research and Policy Review,
vol. 22, no. 5-6, p. 497.
167
Quine, M. P. & Quine, S. 1993, 'Estimation of Infant Mortality Rates Categorized by
Social Class for an Australian Population', Journal of the Royal Statistical Society:
Series C (Applied Statistics), vol. 42, no. 2, p. 333.
Quine, S. 1991, 'Social class as a risk factor for infant mortality in an Australian
population', Journal of Biosocial Science, vol. 23, no. 1, pp. 65-72.
Raphael, D. 2002, Social Justice is Good for Our hearts: Why Societal Factors – Not
Lifestyles – are Major Causes of Heart Disease in Canada and Elsewhere, Centre for
Social Justice Foundation and Research and Education, Toronto, Canada.
Reichman, N. E. 2005, 'Low Birth Weight and School Readiness', The Future of
Children, vol. 15, no. 1, pp. 91-116.
REUTERS, Spain adopts baby bonus scheme [Online], Available:
http://geographyfieldwork.com/SpainBabyBonus.htm [4 March].
Rich-Edwards, J. W., Stampfer, M. J., Manson, J. E., Rosner, B., Hankinson, S. E.,
Colditz, G. A., Willett, W. C. & Hennekens, C. H. 1997, 'Birth weight and risk of
cardiovascular disease in a cohort of women followed up since 1976', British Medical
Journal, vol. 315, no. 7105, p. 396.
Roberts, G., Bellinger, D. & McCormick, M. 2007, 'A Cumulative Risk Factor Model for
Early Identification of Academic Difficulties in Premature and Low Birth Weight Infants',
Maternal & Child Health Journal, vol. 11, no. 2, pp. 161-172.
Robinson, W. S. 1950, 'Ecological correlations and the behavior of individuals',
American Sociological Review, vol. 15, no. 3, pp. 351-357.
168
Roos, L. L. & Wajda, A. 1991, 'Record linkage strategies. Part I: Estimating information
and evaluating approaches', Methods Inf Med, vol. 30, no. 2, pp. 117-23.
Rosenbluth, F. M. & Iversen, T. 2006, The Political Economy of Fertility in Annual
Meeting of the American Political Science Association,
http://www.allacademic.com/meta/p151279_index.html, Philadelphia.
Ross, N. A., Wolfson, M. C., Dunn, J. R., Berthelot, J.-M., Kaplan, G. A. & Lynch, J. W.
2000, 'Relation between income inequality and mortality in Canada and in the United
States: cross sectional assessment using census data and vital statistics', BMJ, vol.
320, no. 7239, pp. 898-902.
Rutter, M. 2002, 'The interplay of nature, nurture, and developmental influences: the
challenge ahead for mental health.' Archives of General Psychiatry [NLM - MEDLINE],
vol. 59, no. 11, p. 996.
Saigal, S. & Doyle, L. W. 2008, 'An overview of mortality and sequelae of preterm birth
from infancy to adulthood', The Lancet, vol. 371, no. 9608, pp. 261-269.
Sampson, R. J. & Wilson, J. W. 1995, 'Toward a theory of race, crime, and urban
inequality', in Crime and inequality, eds J. Hagan & R. D. Peterson, Stanford University
Press, Stanford, Calif. :, pp. x, 372 p. :.
Sanders, D. 1985, The Struggle for Health: Medicine and the Politics of
Underdevelopment. , Macmillan, London.
SAS Institute Inc. 2002-2003, SAS, ed.^eds 9.1, Cary, NC, USA.
169
Sayers, S. & Powers, J. 1997, 'Risk factors for Aboriginal low birthweight, intrauterine
growth retardation and preterm birth in the Darwin Health Region', Australian and New
Zealand Journal of Public Health, vol. 21, no. 5, pp. 524-530.
Schultz, R., Read, A. W., Straton, J. A., Stanley, F. J. & Morich, P. 1991, 'Genitourinary
tract infections in pregnancy and low birth weight: case-control study in Australian
aboriginal women', BMJ, vol. 303, no. 6814, pp. 1369-1373.
Schwartz, S. 1994, 'The fallacy of the ecological fallacy: The potential misuse of a
concept and the consequences', American Journal of Public Health, vol. 84, no. 5, p.
819.
Scott, J., Binns, C. & Aroni, R. 1997, 'The influence of reported paternal attitudes on
the decision to breast-feed', Journal of Paediatrics and Child Health, vol. 33, no. 4, pp.
305-307.
Shaerraden, M. 2001, Singapore announces "Baby Bonus" and children's development
accounts, Center for Social Developments, Washingotn University, St Louis.
Shaw, C., Blakely, T., Crampton, P. & Atkinson, J. 2005, 'The contribution of causes of
death to socioeconomic inequalities in child mortality: New Zealand 1981-1999', The
New Zealand Medical Journal, vol. 118, no. 1227.
Sheldon, T. A. & Parker, H. 1992, 'Race and ethnicity in health research', J Public
Health, vol. 14, no. 2, pp. 104-110.
Shouls, S., Congdon, P. & Curtis, S. 1996, 'Geographic variation in illness and
mortality: the development of a relevant area typology for SAR districts', Health &
Place, vol. 2, no. 3, pp. 139-155.
170
Siegrist, J. & Marmot, M. 2004, 'Health inequalities and the psychosocial environment--
two scientific challenges', Social Science & Medicine, vol. 58, no. 8, pp. 1463-1473.
Simhan, H. N. & Krohn, M. A. 2008, 'Paternal race and preterm birth', American Journal
of Obstetrics and Gynecology, vol. 198, no. 6, pp. 644.e1-644.e6.
Singapore Government & (Family Services Division), Children Development Co-
Savings (Baby Bonus) Scheme [Online], Available:
http://www.babybonus.gov.sg/bbss/html/index.html [4 March ].
Singh-Manoux, A., Nabi, H., Shipley, M., Gueguen, A., Sabia, S., Dugravot, A.,
Marmot, M. & Kivimaki, M. 2008, 'The Role of Conventional Risk Factors in Explaining
Social Inequalities in Coronary Heart Disease: The Relative and Absolute Approaches
to Risk. [Article]', Epidemiology, vol. 19, no. 4, pp. 599-605.
Singh, G. K. & Yu, S. M. 1995, 'Infant mortality in the United States: Trends,
differentials, and projections, 1950 through 2010', American Journal of Public Health,
vol. 85, no. 7, p. 957.
Smaje, C. 1996, 'The Ethnic Patterning of Health: New Directions for Theory and
Research', Sociology of Health & Illness, vol. 18, no. 2, pp. 139-171.
Smeeton, N. C., Rona, R. J., Dobson, P., Cochrane, R. & Wolfe, C. 2004, 'Assessing
the determinants of stillbirths and early neonatal deaths using routinely collected data
in an inner city area', BMC Med, vol. 2, p. 27.
Smith, D. & Shen, X., Estimating the Risk of Preterm Birth [Online], Available:
http://biomed.csse.uwa.edu.au/grants_2.php [May 11].
171
Sonnander, K. & Claesson, M. 1999, 'Predictors of developmental delay at 18 months
and later school achievement problems', Dev Med Child Neurol, vol. 41, no. 3, pp. 195-
202.
Spencer, N., Bambang, S., Logan, S. & Gill, L. 1999, 'Socioeconomic status and birth
weight: Comparison of an area-based measure with the Registrar General's social
class', Journal of Epidemiology and Community Health, vol. 53, no. 8, p. 495.
Stamp, G., Champion, S., Anderson, G., Warren, B., Stuart-Butler, D., Doolan, J.,
Boles, C., Callaghan, L., Foale, A. & Muyambi, A. 2008, 'Aboriginal maternal and infant
care workers: partners in caring for Aboriginal mothers and babies', Rural and Remote
Health, vol. 8, no. 883.
Stanley, F., De Klerk, N., Bower, C., Li, J., Ferrante, A., Leonard, H. & Kendall, G.
2004, Developmental pathways to health, education and delinquent outcomes in
Western Australian children: A holistic approach to inform early intervention strategies,
Telethon Institute for Child Health Research, University of Western Australia.
Stanley, F., J. , Croft, M., L. , Gibbins, J. & Read, A., W. 1994, 'A population database
for maternal and child health research in Western Australia using record linkage',
Paediatric and Perinatal Epidemiology, vol. 8, no. 4, pp. 433-447.
Starfield, B., Robertson J & AW., R. 2002, 'Social class gradients and health in
childhood', Ambulatory pediatrics, vol. 2, no. 4, pp. 238-46.
Steenland, K., Henley, J., Calle, E. & Thun, M. 2004, 'Individual- and Area-Level
Socioeconomic Status Variables as Predictors of Mortality in a Cohort of 179,383
Persons', Am. J. Epidemiol., vol. 159, no. 11, pp. 1047-1056.
172
Steering Committee for the Review of Government Service Provision 2003-,
Overcoming indigenous disadvantage. Key indicators, Steering Committee for the
Review of Government Service Provision,, Melbourne :.
Sweet, L. & Power, C. 2009, 'Family Law as a determinant of child health and welfare:
Shared parenting, breastfeeding and the best interests of the child', Health Sociology
Review, vol. 18, no. 1, pp. 108-118.
Taylor, L. 1995, 'PM admits gap between rich and poor has widened', The Australian, 6
October, p. 2.
Taylor, R., Page, A., Morrell, S., Carter, G. & Harrison, J. 2004, 'Socio-economic
differentials in mental disorders and suicide attempts in Australia', Br J Psychiatry, vol.
185, no. 6, pp. 486-493.
Taylor, R., Quine, S., Lyle, D. & Bilton, A. 1992, 'Socioeconomic correlates of mortality
and hospital morbidity differentials by Local Government Area in Sydney 1985-1988',
Australian Journal Of Public Health, vol. 16, pp. 305-314.
The Practice Committee of the American Society for Reproductive Medicine 2006,
'Multiple pregnancy associated with infertility therapy', Fertility and Sterility, vol. 86, no.
5, Supplement 1, pp. S106-S110.
Thomson, N. 1995, 'The health of Australian Aborigines and Torres Strait Islanders: A
sociocultural perspective', in Sociology of health and illness : Australian readings, 2nd
ed edn, eds G. M. Lupton & J. M. Najman, Macmillan Education,, South Melbourne :.
173
Tong, S., Baghurst, P. & McMichael, A. 2006, 'Birthweight and cognitive development
during childhood', Journal of Paediatrics & Child Health, vol. 42, no. 3, pp. 98-103.
Townsend, P. 1986, 'Why are the many poor?' International Journal of Health Services,
vol. 16, pp. 1-32.
Townsend, P., Davidson, N. & Whitehead, M. 1988, Inequalities in health: The Black
report and the health divide, Penguin,, London :.
Tracy, S., Tracy, M. B., Dean, J., Laws, P. & Sullivan, E. 2007, 'Spontaneous preterm
birth of liveborn infants in women at low risk in Australia over 10 years: a population-
based study', BJOG: An International Journal of Obstetrics & Gynaecology, vol. 114,
no. 6, pp. 731-735.
Trewin, D. 2003, Information Paper Census of Population and Housing Socio-
Economic Indexes for Areas Australia 2001, Australian Bureau of Statistics, Canberra.
Trewin, D. 2004, Technical Paper Census of Population and Housing: Socio-Economic
Indexes for Area's (SEIFA) Australia 2001, Australian Bureau of Statistics, Canberra.
Trewin, D. 2006, Statistical Geography Volume 1 - Australian Standard Geographical
Classification (ASGC), Australian Bureau of Statistics (ABS), Canberra.
Trewin, D. & Madden, R. 2005, The Health and Welfare of Australia's Aboriginal and
Torres Strait Islander Peoples, Australian Institute of Health and Welfare
Australian Bureau of Statistics, Canberra.
174
Trewin, D. & Pink, B. 2006, ANZSCO - Australian and New Zealand Standard
Classification of Occupations, Australian Bureau of Statistics/ Statistics New Zealand
Canberra.
Turrell, G. & Mathers, C. D. 2000, 'Socioeconomic status and health in Australia',
Medical Journal of Australia., vol. 172, no. 9, pp. 434-438.
Turrell, G., Oldenburg, B., McGuffog, I. D. & Dent, R. 1999, Socioeconomic
determinants of health : towards a national research program and a policy and
intervention agenda, Queensland University of Technology, School of Public Health,
Centre for Public Health Research,, Brisbane :.
United Nations Children’s Fund (UNICEF), State of the world's children 2008 [Online],
UNICEF, Available: http://www.unicef.org/sowc08/docs/sowc08.pdf [8 December].
Van Os, J., Driessen, G. E. R., Gunther, N. & Delespaul, P. 2000, 'Neighbourhood
variation in incidence of schizophrenia: Evidence for person-environment interaction',
The British Journal of Psychiatry, vol. 176, no. 3, pp. 243-248.
Various, (May 12), Readers' Comments: Strong support to deny rich the baby bonus
[Online], Available: http://www.news.com.au/comments/0,23600,23682811-2,00.html
[9 December 2008].
Various, (September 29), Readers' Comments: Employers reject maternity leave plan
[Online], Available: http://www.news.com.au/comments/0,23600,24418286-2,00.html
[9 December 2008].
Vere, J., P. 2007, ''Having It All' No Longer: Fertility, Female Labor Supply, and the
New Life Choices of Generation X*', Demography, vol. 44, no. 4, p. 821.
175
Wadsworth, M. 1999, 'Early life', in Social Determinants of Health, eds M. Marmot & R.
G. Wilkinson, Oxford University Press, Oxford.
Wadsworth, M. & Butterworth, S. 2006, 'Early life', in Social Determinants of Health,
eds M. Marmot & R. G. Wilkinson, Oxford University Press, Oxford.
Wagner, A. K., Soumerai, S. B., Zhang, F. & Ross-Degnan, D. 2002, 'Segmented
regression analysis of interrupted time series studies in medication use research',
Journal of Clinical Pharmacy & Therapeutics, vol. 27, no. 4, pp. 299-309.
Wagstaff, A. & van Doorslaer, E. 2000, 'Income inequality and health: what does the
literature tell us?' Annual Review of Public Health [NLM - MEDLINE], vol. 21, p. 543.
Waldman, R. J. 1992, 'Income distribution and infant mortality', The Quarterly Journal
of Economics, vol. 107, p. 1283.
Walker, A. E. & Becker, N. G. 2005, 'Health inequalities across socio-economic groups:
comparing geographic-area-based and individual-based indicators', Public Health, vol.
119, no. 12, pp. 1097-1104.
Wang, M. L., Dorer, D. J., Fleming, M. P. & Catlin, E. A. 2004, 'Clinical Outcomes of
Near-Term Infants', Pediatrics, vol. 114, no. 2, pp. 372-376.
WHO 2004, Priorities to take forward the health equity policy agenda, WHO,
Copenhagan.
Wild, S. & McKeigue, P. 1997, 'Cross sectional analysis of mortality by country of birth
in England and Wales, 1970-92', British Medical Journal, vol. 314, no. 7082, p. 705.
176
Wilkinson, R. & Pickett, K. E. 2009, The spirit level : why more equal societies almost
always do better, Allen Lane, London.
Wilkinson, R. G. 1992, 'National Mortality Rates: The Impact of Inequality?' American
Journal of Public Health, vol. 82, no. 8, p. 1082.
Wilkinson, R. G. 1996, Unhealthy societies: the afflications of inequality, London.
Williams, D. R. & Collins, C. 2001, 'Racial Residential Segregation: A Fundamental
Cause of Racial Disparities in Health', Public Health Reports (1974-), vol. 116, no. 5,
pp. 404-416.
Wood, L., France, K., Hunt, K., Eades, S. & Slack-Smith, L. 2008, 'Indigenous women
and smoking during pregnancy: Knowledge, cultural contexts and barriers to
cessation', Social Science & Medicine, vol. 66, no. 11, pp. 2378-2389.
Yogev, Y. & Catalano, P. M. 2009, 'Pregnancy and obesity', Obstet Gynecol Clin North
Am, vol. 36, no. 2, pp. 285-300, viii.
Young, A. F., Dobson, A. J. & Byles, J. E. 2001, 'Health services research using linked
records: who consents and what is the gain?' Australian and New Zealand Journal of
Public Health, vol. 25, no. 5, pp. 417-420.
Zapata, B. C., Rebolledo, A. & et al. 1992, 'The Influence of Social and Political
Violence on the Risk of Pregnancy Complications', American Journal of Public Health,
vol. 82, no. 5, p. 685.
177
Zinn, C. 1997, 'Health gap between Aboriginal and non-Aboriginal people is widening',
British Medical Journal, vol. 314, no. 7088, p. 1148.
Zubrick, S., Lawrence, D., Silburn, S., Blair, E., Milroy, H., Wilkes, T., Eades, S.,
D’Antoine, H., Read, A., Ishiguchi, P. & Doyle, S. 2004, The Western Australian
Aboriginal Child Health Survey: The Health of Aboriginal Children and Young People.,
Telethon Institute for Child Health Research, Perth.
178
APPENDIX A: SOCIOECONOMIC INDEXES FOR AREAS (SEIFA)
179
CONSTRUCTING THE SOCIOECONOMIC INDEXES FOR AREAS
The Socioeconomic Indexes for Areas (SEIFA) are a set of area-level composite
indices that reflect the characteristics of individuals, families and households. The
indices are produced by the Australian Bureau of Statistics (ABS) using the National
Census20
data and have been created since 1990 (using the 1986 census data). While
several indices have been created following each census, three of those have been
consistent across all census years. They are the Index of Relative Socioeconomic
Disadvantage (IRSD), Index of Economic Resources (IEO) and Index of Education and
Occupation (IEO). The IRSD only distinguishes between relative disadvantage and a
lack thereof, the IER provides an indication of access to economic resources, and the
IEO provides an indication of qualifications and occupations (Pink 2008c).
There are several steps involved in creating the SEIFA indices, with the initial step
involving the selection of relevant census variables for each index (McLennan 1990;
Castles 1994; McLennan 1998; Trewin 2004; Pink 2008c). For example, for the Index
of Relative Socioeconomic Disadvantage (IRSD) variables associated with relative
disadvantage, such as people’s access to material and social resources are included
(e.g. employment, income, education, household occupancy and fluency in English).
Following the selection of these variables, principal components analysis (PCA) is used
to create each composite index. PCA essentially summarises a large set of correlated
variables and produces the principal components, which are a reduced set of variables.
For all SEIFA indices, the first principal component is used, ‘as it is the single
transformed variable that best summarises the common trend underlying the original
set of variables’ (Pink 2008c)21
.
20 In Australia, the National Census is compulsory and is completed by the total population. 21 For more information regarding principal components analysis see Jolliffe, I. T. 1986, Principal component analysis, Springer-Verlag, New York ; Berlin :.
180
SEIFA VARIABLES
The following tables contain a list of the variables and variable loadings and weights
from the 2006 census for the Index of Relative Socioeconomic Disadvantage, Index of
Economic Resources and Index of Education and Occupation.
181
Table 20: Index of Relative Socioeconomic Disadvantage Variable loading
Variable weight
Variable description
–0.85 –0.33 % Occupied private dwellings with no Internet connection
–0.76 –0.30 % Employed people classified as Labourers
–0.76 –0.30 % People aged 15 years and over with no post-school
qualifications
–0.76 –0.30 % People with stated annual household equivalised
income between $13,000 and $20,799 (approx. 2nd
–0.70
and 3rd
deciles)
–0.27 % Households renting from a Government or Community
organisation
–0.70 –0.27 % People (in the labour force) unemployed
–0.67 –0.26 % Families that are one parent families with dependent
offspring only
–0.67 –0.26 % Households paying rent who pay less than $120 per
week (excluding $0 per week)
–0.61 –0.24 % People aged under 70 who have a long-term health
condition or disability and need assistance with core
activities
–0.57 –0.22 % Occupied private dwellings with no car
–0.52 –0.20 % People who identified themselves as being of Aboriginal
and/or Torres Strait Islander origin
–0.52 –0.20 % Occupied private dwellings requiring one or more extra
bedrooms (based on Canadian National Occupancy
Standard)
–0.51 –0.20 % People aged 15 years and over who are separated or
divorced
–0.51 –0.20 % Employed people classified as Machinery Operators and
Drivers
–0.44 –0.17 % People aged 15 years and over who did not go to school
–0.44 –0.17 % Employed people classified as Low Skill Community and
Personal Service Workers
–0.33 –0.13 % People who do not speak English well
Taken from (Pink 2008c)
182
Table 21: Index of Economic Resources Variable loading
Variable weight
Variable description
–0.71 –0.31 % People with stated annual household equivalised
income
between $13,000 and $20,799 (approx. 2nd and 3rd
deciles)
–0.70 –0.30 % Families that are one parent families with dependent
offspring only
–0.69 –0.30 % Occupied private dwellings with no car
–0.68 –0.29 % Households renting from a Government or
Community organisation
–0.64 –0.28 % Households paying rent who pay less than $120 per
week (excluding $0 per week)
–0.62 –0.27 % People aged 15 and over who are unemployed
–0.58 –0.25 % Households that are lone person households
–0.47 –0.20 % Occupied private dwellings requiring one or more
extra bedrooms (based on Canadian National
Occupancy Standard)
0.33 0.14 % Households owning the dwelling they occupy
(without a mortgage)
0.45 0.20 % Occupied private dwellings with at least one person
who is an owner of an unincorporated enterprise
0.52 0.23 % Households paying mortgage who pay more than
$2,120 per month
0.55 0.24 % Households owning the dwelling they occupy (with a
mortgage)
0.55 0.24 % Households paying rent who pay more than $290
per week
0.63 0.27 % People with stated annual household equivalised
income
greater than $52,000 (approx. 9th and 10th deciles)
0.68 0.29 % Occupied private dwellings with four or more
bedrooms
Taken from (Pink 2008c)
183
Table 22: Index of Education and Occupation Variable loading
Variable weight
Variable description
–0.88 –0.41 % People aged 15 years and over whose highest level
of school completed is Year 11 or lower
–0.87 –0.40 % People aged 15 years and over with no post-school
qualifications
–0.79 –0.36 % Employed people who work in a Skill Level 5
occupation
–0.66 –0.31 % Employed people who work in a Skill Level 4
occupation
–0.50 –0.23 % People (in the labour force) unemployed
–0.50 –0.23 % People aged 15 years and over with a certificate
qualification
0.56 0.26 % People aged 15 years and over at university or
other tertiary institution
0.76 0.35 % People aged 15 years and over with an advanced
diploma or diploma qualification
0.85 0.39 % Employed people who work in a Skill Level 1
occupation
Taken from (Pink 2008c)
184
APPENDIX B: OCCUPATION GROUPS
185
Unit Group
Occupation title Skill Level
1000 Managers (not further defined) 1
1110 Chief Executives, Managing Directors and Legislators
(not further defined)
1
1111 Chief Executives and Managing Directors 1
1112 General Managers 1
1113 Legislators 1
1210 Farmers (not further defined) 1
1211 Aquaculture Farmers 1
1212 Crop Farmers 1
1213 Livestock Farmers 1
1214 Mixed Crop and Livestock Farmers 1
1311 Advertising and Sales Managers 1
1321 Corporate Services Managers 1
1322 Finance Managers 1
1323 Human Resource Managers 1
1324 Policy and Planning Managers 1
1325 Research and Development Managers 1
1331 Construction Managers 1
1332 Engineering Managers 1
1333 Importers, Exporters and Wholesalers 1
1334 Manufacturers 1
1335 Production Managers 1
1336 Supply and Distribution Managers 1
1341 Child Care Centre Managers 1
1342 Health and Welfare Services Managers 1
1343 School Principals 1
1344 Other Education Managers 1
1351 ICT Managers 1
1390 managers/proprietors (not further defined) 1
1391 Commissioned Officers (Management) 1
1392 Senior Non-commissioned Defence Force Members 1
1399 Other Specialist Managers 1
1411 Cafe and Restaurant Managers 2
1412 Caravan Park and Camping Ground Managers 2
1413 Hotel and Motel Managers 2
1414 Licensed Club Managers 2
186
1419 Other Accommodation and Hospitality Managers 2
1421 Retail Managers 2
1491 Amusement, Fitness and Sports Centre Managers 2
1492 Call or Contact Centre and Customer Service
Managers
2
1493 Conference and Event Organisers 2
1494 Transport Services Managers 2
1499 Other Hospitality, Retail and Service Managers 2
1501 Self-employed 1
1502 Self-employed 2
2000 Professional (not further defined) 1
2111 Actors, Dancers and Other Entertainers 1
2112 Music Professionals 1
2113 Photographers 1
2114 Visual Arts and Crafts Professionals 1
2120 Media professionals (not further defined OR not
elsewhere classified)
1
2121 Artistic Directors, and Media Producers and
Presenters
1
2122 Authors, and Book and Script Editors 1
2123 Film, Television, Radio and Stage Directors 1
2124 Journalists and Other Writers 1
2211 Accountants 1
2212 Auditors, Company Secretaries and Corporate
Treasurers
1
2221 Financial Brokers 2
2222 Financial Dealers 1
2223 Financial Investment Advisers and Managers 1
2231 Human Resource Professionals 1
2232 ICT Trainers 1
2233 Training and Development Professionals 1
2241 Actuaries, Mathematicians and Statisticians 1
2242 Archivists, Curators and Records Managers 1
2243 Economists 1
2244 Intelligence and Policy Analysts 1
2245 Land Economists and Valuers 1
2246 Librarians 1
2247 Management and Organisation Analysts 1
187
2249 Other Information and Organisation Professionals 1
2251 Advertising and Marketing Professionals 1
2252 ICT Sales Professionals 1
2253 Public Relations Professionals 1
2254 Technical Sales Representatives 1
2311 Air Transport Professionals 1
2312 Marine Transport Professionals 1
2321 Architects and Landscape Architects 1
2322 Cartographers and Surveyors 1
2323 Fashion, Industrial and Jewellery Designers 1
2324 Graphic and Web Designers, and Illustrators 1
2325 Interior Designers 1
2326 Urban and Regional Planners 1
2331 Chemical and Materials Engineers 1
2332 Civil Engineering Professionals 1
2333 Electrical Engineers 1
2334 Electronics Engineers 1
2335 Industrial, Mechanical and Production Engineers 1
2336 Mining Engineers 1
2339 Other Engineering Professionals 1
2341 Agricultural and Forestry Scientists 1
2342 Chemists, and Food and Wine Scientists 1
2343 Environmental Scientists 1
2344 Geologists and Geophysicists 1
2345 Life Scientists 1
2346 Medical Laboratory Scientists 1
2347 Veterinarians 1
2349 Other Natural and Physical Science Professionals 1
2410 Teachers (not further defined) 1
2411 Early Childhood (Pre-primary School) Teachers 1
2412 Primary School Teachers 1
2413 Middle School Teachers (Aus) / Intermediate School
Teachers (NZ)
1
2414 Secondary School Teachers 1
2415 Special Education Teachers 1
2421 University Lecturers and Tutors 1
2422 Vocational Education Teachers (Aus) / Polytechnic
Teachers (NZ)
1
188
2491 Education Advisers and Reviewers 1
2492 Private Tutors and Teachers 1
2493 Teachers of English to Speakers of Other Languages 1
2494 Education professional (not elsewhere classified) 1
2511 Dieticians 1
2512 Medical Imaging Professionals 1
2513 Occupational and Environmental Health
Professionals
1
2514 Optometrists and Orthoptists 1
2515 Pharmacists 1
2519 Other Health Diagnostic and Promotion Professionals 1
2521 Chiropractors and Osteopaths 1
2522 Complementary Health Therapists 1
2523 Dental Practitioners 1
2524 Occupational Therapists 1
2525 Physiotherapists 1
2526 Podiatrists 1
2527 Speech Professionals and Audiologists 1
2531 Generalist Medical Practitioners 1
2532 Anaesthetists 1
2533 Internal Medicine Specialists 1
2534 Psychiatrists 1
2535 Surgeons 1
2539 Other Medical Practitioners 1
2541 Midwives 1
2542 Nurse Educators and Researchers 1
2543 Nurse Managers 1
2544 Registered Nurses 1
2610 ICT professional (not further defined OR not
elsewhere classified)
1
2611 ICT Business and Systems Analysts 1
2612 Multimedia Specialists and Web Developers 1
2613 Software and Applications Programmers 1
2621 Database and Systems Administrators, and ICT
Security Specialists
1
2631 Computer Network Professionals 1
2632 ICT Support and Test Engineers 1
2633 Telecommunications Engineering Professionals 1
189
2711 Barristers 1
2712 Judicial and Other Legal Professionals 1
2713 Solicitors 1
2721 Counsellors 1
2722 Ministers of Religion 1
2723 Psychologists 1
2724 Social Professionals 1
2725 Social Workers 1
2726 Welfare, Recreation and Community Arts Workers 1
3111 Agricultural Technicians 2
3112 Medical Technicians 2
3113 Primary Products Inspectors 2
3114 Science Technicians 2
3121 Architectural, Building and Surveying Technicians 2
3122 Civil Engineering Draftspersons and Technicians 2
3123 Electrical Engineering Draftspersons and Technicians 2
3124 Electronic Engineering Draftspersons and
Technicians
2
3125 Mechanical Engineering Draftspersons and
Technicians
2
3126 Safety Inspectors 2
3129 Other Building and Engineering Technicians 2
3131 ICT Support Technicians 2
3132 Telecommunications Technical Specialists 2
3211 Automotive Electricians 3
3212 Motor Mechanics 3
3213 Automotive technicians 3
3213 Automotive technicians 3
3221 Metal Casting, Forging and Finishing Trades Workers 3
3222 Sheetmetal Trades Workers 3
3223 Structural Steel and Welding Trades Workers 3
3231 Aircraft Maintenance Engineers 3
3232 Metal Fitters and Machinists 3
3233 Precision Metal Trades Workers 3
3234 Toolmakers and Engineering Patternmakers 3
3241 Panel beaters 3
3242 Vehicle Body Builders and Trimmers 3
3243 Vehicle Painters 3
190
3311 Bricklayers and Stonemasons 3
3312 Carpenters and Joiners 3
3321 Floor Finishers 3
3322 Painting Trades Workers 3
3331 Glaziers 3
3332 Plasterers 3
3333 Roof Tilers 3
3334 Wall and Floor Tilers 3
3341 Plumbers 3
3411 Electricians 3
3421 Air-conditioning and Refrigeration Mechanics 3
3422 Electrical Distribution Trades Workers 3
3423 Electronics Trades Workers 3
3424 Telecommunications Trades Workers 3
3511 Bakers and Pastry cooks 3
3512 Butchers and Smallgoods Makers 3
3513 Chefs 2
3514 Cooks 3
3611 Animal Attendants and Trainers 3
3612 Shearers 3
3613 Veterinary Nurses 3
3621 Florists 3
3622 Gardeners 3
3623 Greenkeepers 3
3624 Nurserypersons 3
3911 Hairdressers 3
3921 Binders, Finishers and Screen Printers 3
3922 Graphic Pre-press Trades Workers 3
3923 Printers 3
3931 Canvas and Leather Goods Makers 3
3932 Clothing Trades Workers 3
3933 Upholsterers 3
3941 Cabinetmakers 3
3942 Wood Machinists and Other Wood Trades Workers 3
3991 Boat Builders and Shipwrights 3
3992 Chemical, Gas, Petroleum and Power Generation
Plant Operators
3
3993 Gallery, Library and Museum Technicians 2
191
3994 Jewellers 3
3995 Performing Arts Technicians 3
3996 Sign writers 3
3999 Other Miscellaneous Technicians and Trades
Workers
3
4111 Ambulance Officers and Paramedics 2
4112 Dental Hygienists, Technicians and Therapists 2
4113 Diversional Therapists 3
4114 Enrolled and Mothercraft Nurses 2
4115 Indigenous Health Workers 2
4116 Massage Therapists 2
4117 Welfare Support Workers 2
4118 Health support worker 2
4118 Health and support workers (not further defined) 2
4200 Carers and aides (not further defined) 4
4211 Child Carers 4
4221 Education Aides 4
4231 Aged and Disabled Carers 4
4232 Dental Assistants 4
4233 Nursing Support and Personal Care Workers 4
4234 Special Care Workers 4
4311 Bar Attendants and Baristas 4
4312 Cafe Workers 5
4313 Gaming Workers 4
4314 Hotel Service Managers 3
4315 Waiters 4
4319 Other Hospitality Workers 5
4411 Defence Force Members - Other Ranks 3
4412 Fire and Emergency Workers 3
4413 Police 2
4421 Prison Officers 4
4422 Security Officers and Guards 5
4511 Beauty Therapists 4
4512 Driving Instructors 3
4513 Funeral Workers 3
4514 Gallery, Museum and Tour Guides 4
4515 Personal Care Consultants 4
4516 Tourism and Travel Advisers 4
192
4517 Travel Attendants 3
4518 Other Personal Service Workers 5
4521 Fitness Instructors 4
4522 Outdoor Adventure Guides 4
4523 Sports Coaches, Instructors and Officials 3
4524 Sportspersons 3
5111 Contract, Program and Project Administrators 2
5121 Office Managers 2
5122 Practice Managers 2
5211 Personal Assistants 3
5212 Secretaries 3
5311 General Clerks 4
5321 Keyboard Operators 4
5411 Call or Contact Centre Workers 4
5412 Inquiry Clerks 4
5421 Receptionists 4
5511 Accounting Clerks 4
5512 Bookkeepers 4
5513 Payroll Clerks 4
5520 Financial and Insurance clerks (not further defined) 4
5521 Bank Workers 4
5522 Credit and Loans Officers 4
5523 Insurance, Money Market and Statistical Clerks 4
5611 Betting Clerks 5
5612 Couriers and Postal Deliverers 5
5613 Filing and Registry Clerks 5
5614 Mail Sorters 5
5615 Survey Interviewers 5
5616 Switchboard Operators 5
5619 Other Clerical and Office Support Workers 5
5911 Purchasing and Supply Logistics Clerks 4
5912 Transport and Despatch Clerks 4
5991 Conveyancers and Legal Executives 2
5992 Court and Legal Clerks 3
5993 Debt Collectors 4
5994 Human Resource Clerks 4
5995 Inspectors and Regulatory Officers 4
5996 Insurance Investigators, Loss Adjusters and Risk 3
193
Surveyors
5997 Library Assistants 4
5999 Other Miscellaneous Clerical and Administrative
Workers
4
6111 Auctioneers, and Stock and Station Agents 3
6112 Insurance Agents 3
6113 Sales Representatives 4
6121 Real Estate Sales Agents 3
6211 Sales Assistants (General) 5
6212 ICT Sales Assistants 5
6213 Motor Vehicle and Vehicle Parts Salespersons 4
6214 Pharmacy Sales Assistants 5
6215 Retail Supervisors 4
6216 Service Station Attendants 5
6217 Street Vendors and Related Salespersons 5
6219 Other Sales Assistants and Salespersons 5
6311 Checkout Operators and Office Cashiers 5
6391 Models and Sales Demonstrators 5
6392 Retail and Wool Buyers 3
6393 Telemarketers 5
6394 Ticket Salespersons 5
6395 Visual Merchandisers 4
6399 Other Sales Support Workers 5
7111 Clay, Concrete, Glass and Stone Processing Machine
Operators
4
7112 Industrial Spray painters 4
7113 Paper and Wood Processing Machine Operators 4
7114 Photographic Developers and Printers 4
7115 Plastics and Rubber Production Machine Operators 4
7116 Sewing Machinists 4
7117 Textile and Footwear Production Machine Operators 4
7119 Other Machine Operators 4
7121 Crane, Hoist and Lift Operators 4
7122 Drillers, Miners and Shot Firers 4
7123 Engineering Production Systems Workers 4
7129 Other Stationary Plant Operators 4
7211 Agricultural, Forestry and Horticultural Plant
Operators
4
194
7212 Earthmoving Plant Operators 4
7213 Forklift Drivers 4
7219 Other Mobile Plant Operators 4
7311 Automobile Drivers 4
7312 Bus and Coach Drivers 4
7313 Train and Tram Drivers 4
7321 Delivery Drivers 4
7331 Truck Drivers 4
7411 Storepersons 4
8110 Cleaner (not further defined) 5
8111 Car Detailers 5
8112 Commercial Cleaners 5
8113 Domestic Cleaners 5
8114 Housekeepers 5
8115 Laundry Workers 5
8116 Other Cleaners 5
8211 Building and Plumbing Labourers 5
8212 Concreters 5
8213 Fencers 4
8214 Insulation and Home Improvement Installers 4
8215 Paving and Surfacing Labourers 5
8216 Railway Track Workers 4
8217 Structural Steel Construction Workers 4
8219 Other Construction and Mining Labourers 5
8311 Food and Drink Factory Workers 5
8312 Meat Boners and Slicers, and Slaughterers 4
8313 Meat, Poultry and Seafood Process Workers 5
8321 Packers 5
8322 Product Assemblers 5
8391 Metal Engineering Process Workers 5
8392 Plastics and Rubber Factory Workers 5
8393 Product Quality Controllers 4
8394 Timber and Wood Process Workers 5
8399 Other Factory Process Workers 5
8410 Farmhands (not further defined) 5
8411 Aquaculture Workers 5
8412 Crop Farm Workers 5
8413 Forestry and Logging Workers 4
195
8414 Garden and Nursery Labourers 5
8415 Livestock Farm Workers 5
8416 Mixed Crop and Livestock Farm Workers 5
8419 Other Farm, Forestry and Garden Workers 5
8511 Fast Food Cooks 5
8512 Food Trades Assistants 5
8513 Kitchenhands 5
8911 Freight and Furniture Handlers 5
8912 Shelf Fillers 5
8991 Caretakers 5
8992 Deck and Fishing Hands 4
8993 Handypersons 5
8994 Motor Vehicle Parts and Accessories Fitters 4
8995 Printing Assistants and Table Workers 4
8996 Recycling and Rubbish Collectors 5
8997 Vending Machine Attendants 5
8999 Other Miscellaneous Labourers 5
9988 inadequate information supplied 0
9989 invalid pensioner 0
9989 other pensioner 0
9991 Student 0
9992 Child/baby 0
9995 housewife/husband 0
9996 retired 0
9997 unemployed 0
9998 volunteers 0
9999 missing 0