You are here

Article Text

Article menu

Download PDFPDF

Research report
Socioeconomic inequality profiles in physical and developmental health from 0–7 years: Australian National Study
  1. Jan M Nicholson1,2,
  2. Nina Lucas1,
  3. Donna Berthelsen2,
  4. Melissa Wake1,3,4
  1. 1Murdoch Childrens Research Institute, Melbourne, Australia
  2. 2Centre for Learning Innovation, Queensland University of Technology, Brisbane Australia
  3. 3Royal Children's Hospital and Centre for Community Child Health, Melbourne, Australia
  4. 4Department of Paediatrics, The University of Melbourne, Melbourne, Australia
  1. Correspondence to Jan M Nicholson, Murdoch Childrens Research Institute, Royal Childrens Hospital, Flemington Road, Parkville, Melbourne, VIC 3052, Australia; jan.nicholson{at}mcri.edu.au

Abstract

Background Early and persistent exposure to socioeconomic disadvantage impairs children's health and wellbeing. However, it is unclear at what age health inequalities emerge or whether these relationships vary across ages and outcomes. We address these issues using cross-sectional Australian population data on the physical and developmental health of children at ages 0–1, 2–3, 4–5 and 6–7 years.

Methods 10 physical and developmental health outcomes were assessed in 2004 and 2006 for two cohorts each comprising around 5000 children. Socioeconomic position was measured as a composite of parental education, occupation and household income.

Results Lower socioeconomic position was associated with increased odds for poor outcomes. For physical health outcomes and socio-emotional competence, associations were similar across age groups and were consistent with either threshold effects (for poor general health, special healthcare needs and socio-emotional competence) or gradient effects (for illness with wheeze, sleep problems and injury). For socio-emotional difficulties, communication, vocabulary and emergent literacy, stronger socioeconomic associations were observed. The patterns were linear or accelerated and varied across ages.

Conclusions From very early childhood, social disadvantage was associated with poorer outcomes across most measures of physical and developmental health and showed no evidence of either strengthening or attenuating at older compared to younger ages. Findings confirm the importance of early childhood as a key focus for health promotion and prevention efforts.

  • Health inequalities
  • early childhood
  • socioeconomic status
  • physical health
  • emotional health
  • child health
  • inequalities SI
  • mental health DI
  • physical function
  • social class

https://doi.org/10.1136/jech.2009.103291

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Exposure to socioeconomic disadvantage in childhood impairs children's health and wellbeing1–7 with adverse effects persisting into adulthood8–10 independently of socioeconomic conditions in later life.8 11 Early childhood is a key period in which targeted interventions can provide strong foundations for physical and developmental health across the life course. However, attempts to promote an optimal start in life for all children are hampered without evidence as to when health and developmental inequalities emerge in response to social disadvantage and how the patterns and magnitudes of inequalities vary across different outcomes. We examine these issues using new population-based data for Australian children in four age groups up to the age of 7 years.

    A stepwise pattern of increasing vulnerability and risk with increasing levels of disadvantage (socioeconomic gradient) has been observed for many health outcomes and populations.12 13 How these social gradients first develop in early life is less clear. Proposed models include the persistence model, where inequalities are established early in life and remain constant; the cumulative effects model, where inequalities are evident from early in life but increase over time as a result of accumulated exposure to adversity; the emergent model, where inequalities first become apparent at older ages; and the childhood-limited model, in which early life inequalities diminish over time.4 14

    Different models may operate for different outcomes. For example, Chen and colleagues4 reported evidence consistent with a persistence model for asthma severity, a cumulative effects model for smoking commencement and a childhood-limited model for injury, high blood pressure and asthma prevalence. Thus, strategies to address childhood health inequalities may differ in their optimal timing depending on the outcomes of interest.

    Published research into the emergence of childhood health gradients has important limitations. Most studies have examined the effects of socioeconomic circumstances across wide, composite age ranges (eg, all of childhood or adolescence)3 7 8 10 11 15 potentially obscuring age-related patterns within developmental periods. Relatively little is known about social gradients in developmental outcomes, with much of the research employing dichotomous socioeconomic indicators such as family poverty.2 5 16 Thus, it is unclear whether poor developmental outcomes exhibit threshold effects (evident only when a certain level of disadvantage is exceeded), gradient effects (linear declines with increasing disadvantage) or accelerating effects (progressively stronger declines with increasing disadvantage) as suggested by some recent studies.17–19 Further, most research has examined socioeconomic patterns for single childhood outcomes1 or for multiple outcomes within the physical3 4 or developmental17 18 20 health domains. Research into outcomes across both domains simultaneously is rare15 and mostly focused on long-term, rather than immediate, health consequences.20–22

    We address these limitations by examining the strength and patterning of health inequalities across 10 physical and developmental health outcomes using multi-wave, cross-sectional data from the Longitudinal Study of Australian Children (LSAC).23 24 Our first aim is to identify which outcomes show evidence of inequalities across the age range of 0–7 years. Our second aim is to determine whether patterns of inequalities are consistent with: (a) persistence, cumulative, emergent or early childhood-limited models of timing; and (b) gradient, threshold or accelerating patterns.

    Method

    Sample

    Data are from waves 1 (2004) and 2 (2006) of LSAC. Two cohorts were recruited at wave 1: 5107 infants (aged 3–19 months, 53.6% response rate) and 4983 children (aged 4 years 3 months to 5 years 7 months, 48.5% response rate).24 Cohorts were reassessed 2 years later when aged 2–3 years (N=4606, 90.2% retention) and 6–7 years (N=4464, 89.6% retention). Our analyses use data from both cohorts at waves 1 and 2 to provide cross-sectional results at four ages: 0–1, 2–3, 4–5 and 6–7 years.

    Study design and sample information for LSAC are detailed elsewhere.23 24 Approval for the study was granted by the Australian Institute of Family Studies Ethics Committee. Briefly, a two-stage clustered sampling design was used with postcodes randomly selected within strata for state and capital city versus other. Children within postcodes were then randomly selected using the Medicare database in which >90% of infants and 98% of 4-year-old Australian children are enrolled. The LSAC cohorts are broadly representative of the Australian population.25 26 At wave 1, children with more highly educated parents were over-represented (by 10%), while single-parent, non-English speaking families and those living in rental properties were slightly under-represented. Wave 2 retention for both cohorts was slightly lower for children with less highly educated parents, from non-English speaking backgrounds and living in rental properties.

    Measures

    Data were obtained from face-to-face interviews and questionnaires with the child's primary carer (97% were child's mother), direct child assessment and teacher questionnaires.

    Socioeconomic position (SEP) was assessed using the composite variable provided in the LSAC datasets27 that ranks each family's relative socioeconomic position based on parental income, education and occupational prestige. Composite SEP measures have been used in other cohort studies28 and provide a representation of overall socioeconomic circumstances not achieved by single indicators.1 The LSAC SEP measure was developed as follows. Mother's and father's annual income from all sources, including pensions and allowances, was summed and transformed using natural logarithms to smooth the distribution. Educational attainment was estimated from the number of years of education each parent had completed, each ranging from 0 years for those who had never attended school to an assigned maximum of 20 years for those who had completed a post-graduate degree. Parents' occupational status was based on current or most recent occupation and rated using the ANU4 score29 developed from Australian census data to group occupations by skill level and occupation type. Each component measure was then standardised (mean of 0, SD of 1), summed and divided by the number of parents in the home. This averaged score was then re-standardised producing a final continuous measure of SEP.27

    For the current study, quintiles were computed based on the distribution of SEP scores for all ages combined. These cut-points were then applied to each age sample with group 1 representing the least and group 5 the most disadvantaged group.

    Physical health indicators were general health, special healthcare needs, illness with wheeze, injury and sleep problems (table 1). Although not previously linked to health inequalities, sleep problems are associated with poorer health-related quality of life, psychosocial and behavioural problems, and risk for obesity.41 Physical health indicators were based on parent report and dichotomised according to recommended cut-points (table 1).

    View this table:
    Table 1

    Child physical and developmental health indicators

    Developmental health indicators were socio-emotional competence, socio-emotional difficulties, and communication, vocabulary and emergent literacy skills, measured by parent report, teacher report or direct assessment (table 1). As children's skills in these areas are age-related, it was not possible to assess these outcomes at every age (eg, emergent literacy was not assessed at 0–1 or 2–3 years) or to use the same measurement tools at all ages (eg, different measures were used to assess vocabulary skills at different ages). To facilitate cross-age comparisons, the continuous scores were dichotomised with children scoring in the lowest 15% of the distribution for their age group classified as having a developmental health problem. This cut-point has been used in other studies as a marker of clinically significant problems.42 43

    Analyses

    The cross-sectional associations between SEP and each outcome were examined using logistic regression models. Missing data were not imputed as rates were low (<3% on each item) for parent report and direct assessment measures. Data were analysed using SPSS 16.0 and Stata 10.0.44 45 We applied LSAC sample weights that adjust for initial non-response bias and non-random attrition.26

    Preliminary analyses examined the additional impact of the complex survey design (ie, clustering and stratification) and the correlations arising from repeated measurement with the same participants. First, ‘unadjusted’ models (adjusted for sample weights only) were compared to models where the survey design was also taken into account using first-order Taylor linearisation to estimate the SEs on which the 95% CIs are based. Next, the unadjusted models were compared to models accounting for the correlation between outcome measures for the same participants at wave 1 and wave 2 using information sandwich (‘robust’) estimates of SE.46 Neither the survey design nor the correlations resulted in marked changes to the estimated ORs and 95% CIs, although the survey design was slightly more influential. As it is not possible to account for both influences simultaneously, we present the more conservative estimates (ie, those with the widest CIs), which were those accounting for the survey design. Models were also adjusted for the age of the child's primary carer and child gender. Preliminary analyses revealed mostly non-significant gender by SEP interactions (results by gender available from the authors). Child age in months (within each age group) was not associated with SEP and, thus, not included in the models.

    The associations between SEP and each outcome were examined using logistic regression (aim 1). Tests of interaction between SEP and age were used to assess evidence that observed SEP effects varied by age group (aim 2a). Non-statistically significant interactions were accepted as evidence of persistence effects (ie, no age variations) for which outcomes we present combined data across all four age groups. Where significant interactions were found, data were examined separately by age to determine whether the patterns were consistent with age-limited effects (significant associations at younger but not older ages), emergent effects (significant associations at older but not younger ages) or cumulative effects (significant at all ages, stronger at older ages).

    Regarding aim 2b, we examined associations between SEP and outcomes for linearity. Wald tests were computed based on the difference in the χ2goodness of fit statistic between the models with SEP group modelled either as a categorical variable (using dummy variables to contrast each of the four more disadvantaged groups with the least disadvantaged group) or as a continuous variable. Non-significant differences were accepted as evidence of gradient effects. Significant differences were examined for evidence of threshold effects (significantly elevated ORs for higher but not lower disadvantage groups) or accelerating effects (significantly elevated ORs at all SEP groups compared to the referrant with increasing increments in the ORs between SEP groups at higher levels of disadvantage).

    Sensitivity analyses were conducted to assess the impact of using 15% cut-points to dichotomise the developmental outcomes. While the use of a lower cut-point (10%) and a higher cut-point (20%) produced slightly stronger and weaker SEP associations, respectively, the patterns of association were largely unchanged. We also modelled the developmental measures as continuous variables (standardised to facilitate comparisons) and results were again consistent with those obtained using the dichotomised outcomes (tables available on request).

    Results

    Table 2 displays demographic information at each age. As expected, primary carer age and the proportions of children in single-parent families increased with increasing child age group. There were few differences on other demographic variables.

    View this table:
    Table 2

    Demographic characteristics for each age group

    For the five indicators of physical health, there were no statistically significant SEP by age interactions so age-group combined data are presented. Table 3 shows the proportions of children with each physical health problem and the ORs and 95% CIs by SEP group. Small but significant associations with SEP were found for all physical health indicators. The test for non-linearity was significant for general health and special healthcare needs. Compared to the least disadvantaged group, the odds of poorer general health or special healthcare needs were increased only for the most disadvantaged group (OR 1.6; CI 1.3 to 1.8 and 1.3; CI 1.1 to 1.6, respectively) suggesting threshold effects. SEP associations for illness with wheeze, sleep problems and injury were consistent with linear (gradient) effects.

    View this table:
    Table 3

    OR and 95% CI* of having a physical health problem by socioeconomic position (SEP) group

    Table 4 shows the ORs and CIs for each developmental health problem by SEP group. For socio-emotional competence (measured at ages 2–3, 4–5 and 6–7 years), the SEP by age interaction was non-significant and age-group combined data are presented. For this outcome, the test for non-linearity was significant and consistent with a threshold effect. Compared to the least disadvantaged group, groups 4 and 5 had significantly elevated odds of low socio-emotional competence (OR 1.0, CI 1.0 to 1.4 and OR 1.5, CI 1.3 to 1.8, respectively).

    View this table:
    Table 4

    OR and 95% CI* of poor developmental outcomes by socioeconomic position (SEP) group

    For the remaining developmental health indicators, tests for SEP by age interactions were statistically significant and data are presented separately by age in table 4 and figure 1A–D. SEP was not associated with communication skills at 0–1 years. For socio-emotional difficulties, vocabulary, emergent literacy and communication at older ages, greater disadvantage was associated with higher odds of poor developmental health.

    Figure 1
    Figure 1

    ORs (presented on a log scale) by socioeconomic position quintile for socio-emotional difficulties, and poor communication, vocabulary and emergent literacy skills.

    Linearity was examined separately for each age group. Tests for non-linearity were significant for socio-emotional difficulties at all ages and ORs were consistent with an accelerating relationship: elevated odds apparent at SEP groups 2 or 3 and increasing more steeply for groups 4 and 5 (figure 1A). For communication problems, linear relationships were indicated for the two communication indicators at age 2–3 years, whereas the test for non-linearity was significant at age 6–7 years and showed an accelerating relationship (figure 1B). For vocabulary problems, the SEP relationship was linear for ages 2–3 and 6–7 years and accelerating for age 4–5 years (figure 1C). For emergent literacy (assessed at 4–5 years and with two measures at 6–7 years), all SEP relationships were linear. While the slopes of the lines were similar for emergent literacy skills at 4–5 years and the teacher report of mathematical thinking at 6–7 years, teacher report of language and literacy skills at 6–7 years showed a steeper gradient (figure 1D).

    Discussion

    In this national Australian sample, inequalities in physical and developmental health were evident from the earliest years. With the exception of communication skills at 0–1 years, socioeconomic disadvantage was associated with poorer outcomes across all ages and measures. In general, patterns of inequalities were stronger and more variable for developmental than for physical health illustrating the complex nature of inequalities across early childhood.

    For all physical health measures, the absence of age-modifying effects supported a persistence model of inequalities. For general health and special healthcare needs, models were consistent with a threshold effect at all ages; SEP differences were evident only for the most disadvantaged group compared to the least disadvantaged group. The odds for illness with wheeze, sleep problems and injury increased monotonically with each successive level of disadvantage, consistent with gradient effects. This provides the first evidence for the social patterning of sleep problems in childhood and adds to Chen's research3 by demonstrating stability in the patterning of physical health inequalities within early to middle childhood.

    Associations between developmental health and SEP were more complex and did not readily match the theoretical models. Socio-emotional competence was the only outcome showing both age-invariant patterns and a threshold effect (elevated odds at the two highest levels of disadvantage only). Age-modifying effects were observed for the remaining four developmental health outcomes, but were not consistent with cumulative, emergent or childhood-limited models of vulnerability. With the exception of communication at 0–1 years, children were at increased odds of poor outcomes with each successive level of socioeconomic disadvantage with linear relationships at some ages and accelerating patterns at other ages. This provides mixed support for earlier studies that have found accelerating patterns for similar outcomes.17 18 20 While there was some tendency for the ORs to be higher at ages 4–5 or 6–7 years (compared to 0–1 or 2–3 years), differences were not sufficiently large or consistent to provide evidence for a cumulative pattern of health inequalities.

    The developmental health outcomes were measured by parent report, direct child assessment or teacher report. Comparison of ORs by source of measurement suggested higher estimates for direct assessment and teacher-reported outcomes than for parent-reported outcomes. For example, 80% of the ORs for SEP groups 4 and 5 exceeded 3.0 for the direct assessment and teacher measures compared to 17% above 3.0 for parent-report measures. Other studies have similarly reported stronger inequalities for objective versus self-report or parent-report measures15 implying that parent reports may systematically underestimate SEP effects on children.

    Strengths of this study were that SEP effects were examined for discrete age groups within early to middle childhood and the use of nationally representative cohorts, multiple outcome measures and a composite measure of SEP. Limitations were the cross-sectional analyses and the absence of cross-lagged data precluded testing for cohort effects. These shortcomings can be addressed as further waves of LSAC data enable tracking of individual trajectories over time and same-age comparisons of the cohorts.

    Our findings suggest that interventions to reduce inequalities in early childhood should evaluate impacts across multiple health and development domains. The observed inequalities and, thus, potential for intervention gain were largest in children's developmental health outcomes. The breadth of early life outcomes associated with SEP suggests that early interventions designed to address one particular outcome may provide unanticipated benefits in other areas, including physical health. For many of the problems examined in this study, the odds were consistent with either linear or accelerated relationships. Threshold effects were found only for poorer general health, special healthcare needs and lower socio-emotional competence. The possibility that these reflect a lack of measurement variability, rather than an absence of linear or accelerated effects, cannot be excluded.

    Life course and developmental theories postulate that early disadvantage has both an immediate adverse impact on health and development, and that the combination of continuing disadvantage and poor early outcomes set accumulating limitations on children's subsequent trajectories.12 13 47 48 This study found no consistent evidence that the effects of social disadvantage accumulated with age across the early years; it is possible that such accumulating effects may emerge in later childhood or when children's odds for multiple (rather than single) poor outcomes are considered.1 This research also found no evidence that SEP effects weakened with age further highlighting that investment and prevention in the early years should remain a high priority for more disadvantaged children in order to build early human capital across a range of physical and developmental health outcomes.

    What is already known on this subject

    While the negative effects of socioeconomic disadvantage on children's concurrent and later life health is well-known, the patterning of health inequalities across the early years of life remains unclear. In particular, there is little evidence about when inequalities first emerge, whether these remain constant across ages and across diverse physical and developmental health outcomes, and whether inequalities exist at all levels of disadvantage or only after a certain threshold of disadvantage has been exceeded.

    What this study adds

    Using national Australian data, this study demonstrates considerable variability in the social patterning of early-life health and provides new evidence of social inequalities in children's sleep problems. From infancy to school age, we found threshold effects for poor general health, special healthcare needs and socio-emotional competence; linear gradients for illness with wheeze, sleep problems and injury; and linear gradients or accelerating effects varying across age groups for socio-emotional difficulties, communication, language and literacy. The consistent presence of health inequalities across a wide range of health outcomes indicates that early interventions may have broad ranging benefits. However, variability in the patterning of these relationships needs to be considered when targeting interventions to specific age groups or outcomes.

    Acknowledgments

    This paper uses unit record data from Growing Up in Australia—the Longitudinal Study of Australian Children. The study is conducted in partnership between the Department of Families, Housing, Community Services and Indigenous Affairs (FaHCSIA), the Australian Institute of Family Studies (AIFS) and the Australian Bureau of Statistics (ABS). The findings and views reported are those of the authors and should not be attributed to FaHCSIA, AIFS or the ABS. The authors thank Obioha Ukoumunne, Polly Hardy and Helena Romaniuk for statistical advice.

    References

    1. ↵
      1. Bradley RH,
      2. Corwyn RF
      . Socioeconomic status and child development. Annu Rev Psychol 2002;53:371–99.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Brooks-Gunn J,
      2. Duncan G
      . The effects of poverty on children. Future Child 1997;7:55–71.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Chen E,
      2. Martin AD,
      3. Matthews KA
      . Socioeconomic status and health: Do gradients differ within childhood and adolescence? Soc Sci Med 2006;62:2161–70.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Chen E,
      2. Matthews KA,
      3. Boyce WT
      . Socioeconomic differences in children's health: How and why do these relationships change with age? Psychol Bull 2002;128:295–329.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Duncan GJ,
      2. Brooks-Gunn J,
      3. Yeung W,
      4. et al
      . How much does childhood poverty affect the life chances of children? Am Sociol Rev 1998;63:406–23.
      OpenUrlCrossRefWeb of Science
    6. ↵
      1. Entwistle DR,
      2. Alexander KL
      . Entry into school: The beginning school transition and educational stratification in the United States. Annu Rev Sociol 1993;19:401–23.
      OpenUrlCrossRefWeb of Science
    7. ↵
      1. Victorino CC,
      2. Gauthier AH
      . The social determinants of child health: Variations across health outcomes—a population-based cross-sectional analysis. BMC Pediatr 2009;9:53.
      OpenUrlCrossRefPubMed
    8. ↵
      1. Kaplan GA,
      2. Turrell G,
      3. Lynch JW,
      4. et al
      . Childhood socioeconomic position and cognitive function in adulthood. Int J Epidemiol 2001;30:256–63.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Poulton R,
      2. Caspi A,
      3. Milne BJ,
      4. et al
      . Association between children's experience of socioeconomic disadvantage and adult health: A life-course study. Lancet 2002;360:1640–5.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Turrell G,
      2. Lynch JW,
      3. Kaplan GA,
      4. et al
      . Socioeconomic position across the lifecourse and cognitive function in late middle age. J Gerontol B Psychol Sci Soc Sci 2002;578:S43–51.
      OpenUrl
    11. ↵
      1. Beebe-Dimmer J,
      2. Lynch JW,
      3. Turrell G,
      4. et al
      . Childhood and adult socioeconomic conditions and 31-year mortality risk in women. Am J Epidemiol 2004;159:481–90.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Keating DP,
      2. Hertzman C
      , eds. Developmental health and wealth of nations: social, biological and educational determinants. New York: Guildford Press, 1999.
    13. ↵
      1. Kuh D,
      2. Ben-Shlomo Y
      1. Kuh D,
      2. Ben-Shlomo Y
      . Introduction. In: Kuh D, Ben-Shlomo Y, eds. A life course approach to chronic disease epidemiology. Oxford: Oxford University Press, 2004:1–14.
    14. ↵
      1. Prus SG
      . Age, SES, and health: a population level analysis of health inequalities over the lifecourse. Sociol Health Illn 2007;29:275–96.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Emerson E,
      2. Graham H,
      3. Hatton C
      . Household income and health status in children and adolescents in Britain. Eur J Public Health 2005;16:354–60.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Duncan GJ,
      2. Ziol-Guest KM,
      3. Kalil A
      . Early-childhood poverty and adult attainment, behaviour and health. Child Dev 2010;81:306–25.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Kiernan KE,
      2. Mensah FK
      . Poverty, maternal depression, family status and children's cognitive and behavioural development in early childhood: A longitudinal study. J Soc Policy 2009;38:569–88.
      OpenUrlCrossRefWeb of Science
    18. ↵
      1. Mensah FK,
      2. Kiernan KE
      . Gender differences in educational attainment: Influences of the family environment. Br Educ Res J 2010;36:239–60.
      OpenUrlCrossRef
    19. ↵
      1. Tong S,
      2. Baghurst P,
      3. Vimpani G,
      4. et al
      . Socioeconomic position, maternal IQ, home environment and cognitive development. J Pediatr 2007;151:284–8.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Gershoff ET,
      2. Aber JL,
      3. Raver CC,
      4. et al
      . Income is not enough: Incorporating material hardship into models of income associations with parenting and child development. Child Dev 2007;78:70–95.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Fergusson DM,
      2. Horwood LJ,
      3. Lynskey M
      . The childhoods of multiple problem adolescents: A 15-year longitudinal study. J Child Psychol Psychiatry 1994;35:1123–40.
      OpenUrlPubMedWeb of Science
    22. ↵
      1. Mensah FK,
      2. Hobcraft J
      . Childhood deprivation, health and development: Associations with adult health in the 1958 and 1970 British propective birth cohort studies. J Epidemiol Community Health 2008;62:599–606.
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Sanson A,
      2. Nicholson J,
      3. Ungerer J,
      4. et al
      . Introducing the Longitudinal Study of Australian Children: LSAC discussion paper no.1. Melbourne, Australia: Australian Institute of Family Studies, 2002.
    24. ↵
      1. Soloff C,
      2. Lawrence D,
      3. Johnstone R
      . Sample design (LSAC Technical Paper, No. 1). Melbourne: Australian Institute of Family Studies, 2005.
    25. ↵
      1. Misson S,
      2. Sipthorp M
      . Wave 2 weighting and non-response (LSAC Technical Paper, No. 5). Melbourne: Australian Institute of Family Studies, 2007.
    26. ↵
      1. Soloff C,
      2. Lawrence D,
      3. Misson S,
      4. et al
      . Wave 1 weighting and non-response (LSAC Technical Paper No. 3). Melbourne: Australian Institute of Family Studies, 2006.
    27. ↵
      1. Blakemore T,
      2. Strazdins L,
      3. Gibbings J
      . Measuring family socio-economic position. Australian Social Policy Journal 2009;8:121–68.
      OpenUrl
    28. ↵
      1. Willms JD,
      2. Sheilds M
      . A measure of socioeconomic status for the National Longitudinal Study of Children. New Brunswick: Atlantic Center for Policy Research in Education, Statistics Canada, 1996.
    29. ↵
      1. Jones FL,
      2. McMillan J
      . Scoring occupational categories for social research: a review of current practice, with Australian examples. Work, Employment and Society 2001;15:539–63.
      OpenUrlAbstract/FREE Full Text
      1. Waters E,
      2. Salmon L,
      3. Wake M
      . The parent-form Child Health Questionnaire in Australia: comparison of reliability, validity, structure and norms. J Pediatr Psychol 2000;25:381–91.
      OpenUrlAbstract/FREE Full Text
      1. Bethell C,
      2. Read D,
      3. Stein R,
      4. et al
      . Identifying children with special health care needs: development and evaluation of a short-screening instrument. Ambul Pediatr 2002;2:38–48.
      OpenUrlCrossRefPubMedWeb of Science
      1. Briggs-Gowan MJ,
      2. Carter AS,
      3. Irwin JR,
      4. et al
      . The brief infant-toddler social and emotional assessment: screening for social-emotional problems and delays in competence. J Pediatr Psychol 2004;29:143–55.
      OpenUrlAbstract/FREE Full Text
      1. Goodman R
      . The strengths and difficulties questionnaire: a research note. J Child Psychol Psychiat 1997;38:581–6.
      OpenUrlPubMedWeb of Science
      1. Wetherby AM,
      2. Prizant BM
      . Communication and symbolic behaviour scales—infant/toddler checklist. Baltimore: Paul H. Brookes Publishing Co, 2001.
      1. Fenson L,
      2. Marchman VA,
      3. Thal D,
      4. Bates E,
      5. et al
      . The MacArthur-Bates Communicative Development Inventories user's guide and technical manual (2nd Revised edition), Baltimore, MD: Brookes Publishing Co, 2006.
    36. Statistics Canada. National Longitudinal Survey of Children and Youth (NLSCY) Cycle 3 survey instruments: parent questionnaire. Ottawa, Canada: Statistics Canada, 2000.
      1. Bishop DMV
      . Children's communication checklist. London: Harcourt Assessment Inc, 2003.
      1. Rothman S
      . An Australian version of the Adaptive PPVT-III for use in research. Melbourne, Australia: Australian Council for Educational Research, 2003.
      1. de Lemnos M,
      2. Doig B
      . Who am I: developmental assessment manual. Melbourne, Australia: Australian Council for Educational Research, 1999.
    40. National Center for Education Statistics. Early Childhood Longitudinal Study- Kindergarten (ECLS-K). Washington, DC: Department of Education, 2000.
    41. ↵
      1. Hiscock H,
      2. Canterford L,
      3. Ukoumonne OC,
      4. et al
      . Adverse associations of sleep problems in Australian preschoolers: national population study. Pediatrics 2007;119:86–93.
      OpenUrlAbstract/FREE Full Text
    42. ↵
      1. Sanson AV,
      2. Misson S,
      3. Hawkins MT,
      4. et al
      . The development and validation of Australian indices of child development-Part I: conceptualisation and development. Child Indic Res 2010;3:275–92.
      OpenUrlCrossRefWeb of Science
    43. ↵
      1. Willms JD
      , ed. Vulnerable children. Alberta: University of Alberta Press, 2002.
    44. ↵
      SPSS for Windows. Rel. 16.0.0. Chicago: SPSS Inc, 2007.
    45. ↵
      StataCorp. Stata statistical software: release 10. College Station, TX: StataCorp LP, 2007.
    46. ↵
      1. Hanley JA,
      2. Negassa A,
      3. Edwardes MD,
      4. et al
      . Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364–75.
      OpenUrlAbstract/FREE Full Text
    47. ↵
      1. Fox NA,
      2. Rutter M
      . Introduction to the special section on the effects of early experience on development. Child Dev 2010;81:23–7.
      OpenUrlCrossRefPubMedWeb of Science
    48. ↵
      1. Shonkoff JP
      . Building a new biodevelopmental framework to guide the future of early childhood policy. Child Dev 2010;81:357–67.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Funding This work was supported by the National Health and Medical Research Council (Career Development Awards 390136 to JN, 546405 and 284556 to MW) and the Australian Research Council (DP0774439).

    • Competing interests None.

    • Ethics approval This study was conducted with the approval of the Australian Institute of Family Studies Ethics Committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.