The European Journal of Public Health Advance Access originally published online on October 5, 2005
The European Journal of Public Health 2006 16(4):354-360; doi:10.1093/eurpub/cki200
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Health Inequalities |
Household income and health status in children and adolescents in Britain
Eric Emerson1, Hilary Graham2 and Chris Hatton3
1 Professor of Disability and Health Research, Institute for Health Research, Lancaster University, Lancaster, UK
2 Professor of Social Policy, Institute for Health Research, Lancaster University, Lancaster, UK
3 Professor of Psychology of Health and Social Care, Institute for Health Research, Lancaster University, Lancaster, UK
Correspondence: Professor Eric Emerson, Institute for Health Research, Lancaster University, Lancaster LA1 4YT, UK, tel: +44 01524 592260, fax: +44 01524 592406, e-mail: eric.emerson{at}lancaster.ac.uk
Received January 28, 2005, accepted August 30, 2005
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Background: Mortality, health, and well-being across the lifespan are associated with socioeconomic position (typically operationalised as occupational status). There is some evidence that adolescence represents a period of ‘relative equalisation’ of health inequalities. Our aim was to examine the association between inequalities in household income and health in childhood and adolescence. Methods: Cross-sectional survey using multistage stratified random sample of households in Britain. Information was collected on 10438 children aged 5–15 years. Results: Low levels of equivalised household income was associated with poorer health for 13 out of the 22 indicators examined (odds ratio P < 0.05 for
1 income quintile). Multivariate analyses controlling for child characteristics, parental socioeconomic status and household composition indicated that low levels of equivalised household income increased the odds of poor health for 9 out of the 22 indicators examined. There was little evidence of any systematic differences in the extent of health inequalities across age groups (5–10 and 11–15 years). Conclusion: Household income is related to a range of health outcomes for children and adolescents, even when other indicators of socioeconomic status are taken into account. We found little evidence that adolescence represents a period of relative equalisation of health inequalities.
Keywords: adolescents, children, disability, health inequalities
Studies have documented the adverse effects that poor socioeconomic circumstances (SECs) and low income have on health and mortality across the lifecourse.1–5 In particular, evidence suggests that the experience of socioeconomic disadvantage and poverty in childhood is associated with impaired development and poorer health status in childhood,6–11 and increased mortality and morbidity in adulthood.12,13
Some evidence, however, indicates that the strength of the association between SEC and child health may be moderated by age,14 with some studies finding no consistent association in adolescence.15,16 These observations have led to the hypothesis that adolescence is the exception to the lifetime association between SEC and health, with the negative gradients in health evident in childhood and adulthood flattening, or even reversing, between the ages of 11 and 16 years.15–17 Adolescence is therefore seen to represent a period of ‘relative equalisation’ in health inequalities.17 Evidence in support of the relative equalisation hypothesis appears strongest for measures based on parent's occupation, and for some rather than all dimensions of young people's health, including less severe or restricting disorders, accidents, and self-reported symptoms of psychological distress.15,17
However, there are acknowledged difficulties in measuring the SEC of children and young people, with studies highlighting the limitations of measures based on parental occupation in particular.18 Moreover, there are few studies that have incorporated detailed information on health across a number of domains in both childhood and adolescence. Important tasks, therefore, are to examine the negative patterning of health in datasets with comparable health data for both age groups and using measures of SEC which are inclusive and equivalent, across different household structures (two-parent and one-parent) and across children of different ages. Household income provides such a measure and one which is yet to be systematically examined in the studies, from which the equalisation hypothesis was derived.15,17
In this study, we re-examine the equalisation hypothesis in a national survey of British children aged 5–15 years across a range of measures of physical and mental health, using household income as the primary measure of their socioeconomic position.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Secondary analysis was undertaken on the 1999 Office for National Statistics survey of the mental health of children and adolescents in Great Britain.19 Relevant data files20 were obtained from the UK Data Archive. Full details of the design, instrumentation, and sampling procedure have been presented previously.19
Sampling and data collection
The survey collected information on a multistage stratified random sample of 10438 children between 5 and 15 years of age (83% of the target sample of 12529 eligible children). The sample was stratified by age and sex within 475 postal code sectors in England, Scotland, and Wales. Information was collected by interview with the child's primary carer (in 94% of cases the child's mother) and, wherever possible for children aged 11 years or over, interview with the child themselves; psychometric testing of all children; and by postal questionnaire to the child's teacher.
Measures
We constructed a series of binary measures of child and adolescent health based on informant responses to survey questions. We grouped these measures under nine categories of health status. Within the limits of the study, these categories captured children's developmental health2 as well as current health conditions and their experience of accidents. The categories were current health status; current physical illness; sensory impairment; physical disabilities; speech/language problems; accidental injury; psychiatric disorder; specific learning disabilities; and intellectual disabilities. Table 1 details the items included in the nine health categories and the sources of information used for each item and category.
|
Overall morbidity was defined as having a disorder in any of the above nine categories. Co-morbidity (for two or more and three or more disorders) was determined by summing the number of the above nine categories for which the child scored positive.
Equivalised household income was the primary measure of socioeconomic circumstances, based on reported gross weekly income from all sources. Data were adjusted (taking category mid-points) for household composition using standard housing costs equivalisation scales.24 We also extracted information on family social class based on the standard Registrar General's classification self-identified head of household, parent's (usually mother's) educational attainment (equivalent to less than or equal to GCSE grade C or above), household composition, child age, sex, and ethnicity.
Analyses
We split the sample into quintiles on the basis of equivalised household income. Dependent variables were the nine categories of developmental or current health described above, overall morbidity, two measures of multiple morbidity, and the nine individual illness/disability categories that occurred with an overall unweighted prevalence of >1% in the data set, giving a total of 22 dependent variables. To test for evidence of equalisation between childhood and adolescence we followed West15,16 in defining adolescence as those aged 11 years and older. We used logistic regression modelling to calculate odds ratios for univariate and multivariate analyses. In the univariate analyses we examined the simple association between equivalised household income and the dependent variables separately for boys and girls for the two age groups of 5–10 and 11–15 years. In the multivariate analyses we examined the association between equivalised household income and the dependent variables for the two age groups, while controlling for child age, sex, ethnicity, household composition (lone parent status and number of children living in the household), family social class, and level of parent education.
Ethics
Ethics committee approval was not sought as the study involved the secondary analysis of anonymous publicly available data.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
For the majority of health indicators (13 of 22), low levels of equivalised household income were associated with poorer health in either the 5- to 10-year age group or 11- to 15-year age group (odds ratio P < 0.01 for one or more income quintile). There was no statistically significant association in either age group between equivalised income and (i) the three general categories of current physical illness, physical disabilities, and sensory impairment and (ii) the specific physical illness categories of hayfever, glue ear, digestive problems, and food allergies. Inverse health gradients were observed for eczema (for boys in the age range of 11–15 years and girls in the age range of 5–10 years) and non-food allergies (for boys in the age range of 11–15 years).
Table 2 presents the association (odds ratios) between income quintile and health indicators for boys (upper panel) and girls (lower panel). Also included is information on the unweighted prevalence of the indicator, an indicator of the predictive power of income in accounting for variation in the prevalence (Nagelkerke pseudo R2) and an indication (P-value) of the statistical significance of the overall regression model. Results are only presented for those health indicators significantly associated with income at an alpha level of 0.01. Full tables are available on request.
|
For boys, low income was associated with poorer health status on 12 indicators in the age range of 5–10 years, and 10 indicators in the age range of 11–15 years. For girls, low income was associated with poorer health status on six indicators in each age range. We tested for evidence of the equalisation of health inequalities by comparing across the two age ranges (i) the predictive power from statistical analyses of the predictive power of income (Nagelkerke pseudo R2) and (ii) the odds ratio for the lowest income quintile. For girls, we found no evidence of equalisation of health inequalities (R2 Wilcoxon Z = –0.29; P = 0.773; OR Wilcoxon Z = –0.59; P = 0.556). For boys, we found that, while there was no change in the predictive power of income (R2 Wilcoxon Z = –0.09; P = 0.926), there was some evidence for health equalisation in that the odds ratio for the occurrence of health indicators in the lowest income quintile was marginally lower in the 11- to 15-year age range when compared with the 5- to 10-year age range (OR Wilcoxon Z = 1.74; P = 0.083).
We then conducted multivariate analyses on the 22 indicators of health status. The aims of these analyses were to attempt to control for cohort effects associated with variations between the two age groups with respect to child gender and ethnicity and to examine the separate contribution of household income after taking account of other dimensions of negative socioeconomic circumstances in childhood and adolescence likely to moderate the strength of the association between household income and health. In these analyses, we included parental occupational status, informant (primarily maternal) educational attainment, and household composition (lone parent or couple and number of children in the household).6 The characteristics of the two age cohorts with respect to these variables is presented in table 3.
|
Across the 22 indicators and two age groups equivalised household income was associated with variations in health status in 15 analyses. Of the control variables informant educational attainment was associated with variations in health status in 11 analyses; lone parent status in 8 analyses; parental occupational status in 3 analyses; and the number of children in the household in 2 analyses.
Results are presented in table 4 for those dependent variables significantly associated with income at an alpha level of 0.01. In this table we also, as previously, report an indicator of the predictive power and statistical significance of the logistic regression model. In addition, we list the other variables related to socioeconomic status (social class, educational attainment, and household composition) that were significantly associated with health outcomes. Full tables are available on request.
|
We again tested for evidence of the equalisation of health inequalities by comparing, across the two age ranges, the odds ratio for the lowest income quintile. We found little evidence of the equalisation of health inequalities, with little change in income-based inequalities for overall health, total morbidity, and co-morbidity and an exacerbation of income-based health inequalities for psychiatric disorders and intellectual disability. For only two of the disorders included in the index of ‘current physical illness’ (digestive problems and headaches), was there evidence for the equalisation of income-based health inequalities between childhood and adolescence.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Principal findings
Our results suggest that (i) equivalised household income is associated with adverse health outcomes for children and adolescents across a range of indicators and that (ii) this association holds true after controlling for parental occupational status, maternal education, and household composition. Overall, our data did not support the hypothesis that adolescence represents a period of ‘relative equalisation’ of health inequalities.
Strengths and weaknesses of the study
The main strengths of our study derived from the use of a large nationally representative sample of children and adolescents, with a high response rate and multiple indicators of health. The main weaknesses were the absence of self-assessed measures of health, the unknown validity of the measure of self-reported household income used in the study, and the impossibility of attributing causality to the result of a cross-sectional study.
Measures of health based on self-assessment are increasingly used in studies of adolescents15,16,25,26; at younger ages, parents continue to be used as the primary informants.26,27 As our study required comparable health data for children aged 5–15 years, the measures were inevitably limited to those based on parental report and professional assessment. With regard to the measurement of household income, we consider it probable that the use of broad bands of income in the analyses (equivalised income quintiles) is likely to reduce the degree of error in the data. The use of a cross-sectional design meant that we could only compare the socioeconomic patterning of health among children of different ages and not among the same children at different time points. While we cannot rule out the possibility of cohort effects, we consider that they are likely to be modest, given the relatively narrow age range (age 5–15 years) of the study.
Meaning of the study
Our results add to the growing body of evidence relating to the existence of health inequalities among children and young people.6–11 They also extend this literature, which has primarily focused on the association between parental socioeconomic position (proxied by parental occupation) and child/adolescent health by pointing to the importance of household income.
Household income avoids the limitations of occupation-based measures used in UK studies from which the equalisation hypothesis has been generated, in two important ways. First, it provides a direct measure of the material circumstances in which children and young people live and one equivalent across households. Occupation-based indicators are known to poorly characterise the SEC of children in households outside the labour market. In addition, the conventional social class classification (the Registrar General's classification) used in UK studies is acknowledged to provide less precise differentiation of the SEC of female-headed than male-headed households and of ethnic minority households than white households.28,29 Second, the use of household income goes some way to addressing the problem of the non-equivalence of parent-derived measures of SEC for children and adolescents. It has been argued that, with children increasingly shaping their own socioeconomic trajectories as they move into their teenage years, parental occupation provides a less relevant and less reliable measure of the SEC of adolescents than of younger children.30 While not wholly avoiding the problem, household income reduces potential biases in results due to the non-equivalence of the SEC measure.
Within the multivariate analyses, equivalised household income was more frequently associated with health outcomes than any other indicator of socioeconomic circumstances. As such, our results suggest that the traditional UK practice of defining children and young people's SEC almost exclusively in terms of parental occupational status may underestimate the strength of socioeconomic gradients in child and adolescent health. The use of income-based indicators is consistent with current approaches in the UK to the measurement of child poverty.31,32
We found little evidence in support of either the overall equalisation of health inequalities or the emergence of negative health gradients in adolescence. Indeed, in the multivariate analyses we found as much evidence for the exacerbation of health inequalities during adolescence as we did for equalisation. The findings contribute to the wider body of research pointing to continuities in health inequalities across childhood, adolescence, and adulthood.
It should also be noted that there was evidence of income-based health inequalities for only a subset of indicators of child health. We found no evidence for income-based inequalities (or any other types of inequalities) for the three general categories of current physical illness, physical disabilities, and sensory impairment. Our results are, however, generally consistent with the wider literature on health inequalities which suggests that negative health gradients are more pronounced for more severe or professionally defined disorders, and that reverse health gradients tend to be restricted to self-reported measures of health.33
Unanswered questions and future research
Future research is required to determine the extent to which equalisation is specific to particular countries (given that the strength of evidence for equalisation has been found in UK, and, specifically, Scottish studies), health domains, modes of measurement (e.g. parent reports, child self-report, and formal diagnoses) and indicators of socioeconomic circumstances.6,7,15,17 It is likely to be of particular importance to identify the separate effects on health of the multiple aspects of socioeconomic disadvantage.6 A fuller understanding of how household circumstances impact on health across childhood and adolescence is required if health and welfare services are to be appropriately resourced and targeted to tackle health inequality at this crucial stage of the lifecourse.
Key points
|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
1 Department of Health. Independent inquiry into inequalities in health: report. London: The Stationary Office, 1998.
2 Keating DP, Hertzman C, editors. Developmental health and the wealth of nations: social, biological and educational dynamics. New York: Guilford Press, 1999.
3 Blane D. The lifecourse, the social gradient and health. In: Marmot M, Wilkinson RG, editors. Social determinants of health. Oxford: Oxford University Press, 1999: 67–80.
4 Kuh D, Power C, Blane, D, Bartley M. Socioeconomic pathways between childhood and adult health. In: Kuh DL, Ben-Shlomo Y. editors. A lifecourse approach to chronic disease epidemiology: tracing the origins of ill-health from early to adult life, 2nd edition. Oxford: Oxford University Press, 2003: 371–95.
5 Yen IH, Syme SL. The social environment and health: a discussion of the epidemiologic literature. Annu Rev Public Health 1999;20:287–308.[CrossRef][ISI][Medline]
6 Aber JL, Bennett NG, Conley DC, Li J. The effects of poverty on child health and development. Annu Rev Public Health 1997;18:463–83.[CrossRef][ISI][Medline]
7 Bradshaw J, editor. Poverty: the outcomes for children. London: Family Policy Studies Centre, 2001.
8 Brooks-Gunn J, Duncan G. The effects of poverty on children and youth. Future Child 1997;7:55–71.[CrossRef][ISI][Medline]
9 Spencer NJ. Poverty and child health. 2nd edition. Oxford: Radcliffe Medical Press, 2000.
10 Bradley RH, Corwyn RF. Socioeconomic status and child development. Annu Rev Psychol 2002;53:371–99.[CrossRef][ISI][Medline]
11 Newacheck PW, Hung YY, Park MJ, et al. Disparities in adolescent health and health care: does socioeconomic status matter? Health Serv Res 2003;38:1235–52.[CrossRef][ISI][Medline]
12 Kuh D, Hardy R, Langenberg C, et al. Mortality in adults aged 26–54 years related to socioeconomic conditions in childhood and adulthood: post war birth cohort study. Br Med J 2002;325:1076–80.
13 Galobardes B, Lynch JW, Davey Smith G. Childhood socioeconomic circumstances and cause-specific mortality in adulthood: systematic review and interpretation. Epidemiol Rev 2004;26:7–21.
14 Chen E, Matthews KA, Boyce WT. Socio-economic differences in children's health: How and why do these relationships change with age? Psychol Bull 2002;128:295–329.[CrossRef][ISI][Medline]
15 West P, Sweeting H. Evidence on equalisation in health in youth from the West of Scotland. Soc Sci Med 2004;59:13–27.[CrossRef][ISI][Medline]
16 West P, Macintyre S, Annadale E, Hunt K. Social class and health in youth: findings from the West of Scotland Twenty-07 study. Soc Sci Med 1990;30:665–73.[CrossRef][ISI][Medline]
17 West P. Health inequalities in the early years: Is there equalisation in youth? Soc Sci Med 1997;44:833–58.[CrossRef][ISI][Medline]
18 Judge K, Benzeval M. Health inequalities: new concerns about the children of single mothers. Br Med J 1993;396:677–80.
19 Meltzer H, Gatward R, Goodman R, Ford T. Mental health of children and adolescents in Great Britain. London: The Stationery Office, 2000.
20 Office for National Statistics Social Survey Division. Mental Health of Children and Adolescents in Great Britain, 1999 [computer file]. Colchester, Essex: The Data Archive [distributor], 13 December 2000. SN: 4227.
21 Goodman R, Ford T, Richards H, et al. The Development and Well-Being Assessment: description and initial validation of an integrated assessment of child and adolescent psychopathology. J Child Psychol Psychiatry 2000;41:645–56.[CrossRef][ISI][Medline]
22 Dunn LM, Dunn LM, Whetton C, Burley J. The British Picture Vocabulary Scale second edition testbook. Windsor: NFER, 1997.
23 Elliott C, Smith P, McCulloch K. British Ability Scales—second edition: administration and scoring manual. Windsor: NFER, 1996.
24 Department of Social Security. Households below average income 1999/00. London: Department of Social Security, 2001.
25 Currie C, Hurrelman K, Settertobulte W, et al. editors. Health and health behaviour among young people: Health behaviour in school-aged children: A WHO Cross-national Study (HBSC) international report. Copenhagen: WHO, 2000.
26 Prescott-Clarke P, Primatesta P. Health Survey for England: the health of young people '95–97. London: The Stationery Office, 1999.
27 Power C, Manor O, Fox J. Health and class: the early years. London: Chapman Hall, 1991.
28 Arber S. Class, paid employment and family roles: making sense of structural disadvantage, gender and health status, Soc Sci Med 1999;32:425–36.
29 Nazroo JY. The health of Britain's ethnic minorities: findings from a national survey. London: Policy Studies Institute, 1997.
30 Glendinning A, Hendry L, Shucksmith J. Lifestyle, health and social class in adolescence. Soc Sci Med 1995;41:235–48.[CrossRef][ISI][Medline]
31 Department of Work and Pensions. Measuring child poverty. London: Department of Work and Pensions, 2003.
32 House of Commons Work and Pensions Committee. Child poverty in the UK: second report of session 2003–04. London: The Stationery Office, 2004.
33 Wagstaff, A. Poverty and health sector inequalities. Bull World Health Organ 2002;80:97–105.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||