Psychological distress among adolescents, and its relationship to individual, family and area characteristics in East London

https://doi.org/10.1016/j.socscimed.2006.02.012Get rights and content

Abstract

This paper identifies factors associated with variation in psychosocial distress among adolescents in a relatively deprived and ethnically diverse inner city setting in London, UK. The research draws on literature which discusses whether neighbourhood socio-economic conditions are associated with mental health, as well as attributes of individual adolescents and their families. We report an analysis of data from the Research with East London Adolescents: Community Health Survey (RELACHS). The survey collected data on mental health measured by the Strengths and Difficulties Questionnaire (SDQ), and on various aspects of individual and family circumstances. These data were linked with information about social and economic conditions in ‘middle layer standard output areas’ (MSOA) used for the population Census 2001, having a mean total population of 6767 in the study area. Census statistics including the Indices of Deprivation for 2004 proposed by the Office of the Deputy Prime Minister, were used to describe the socio-economic conditions within these areas. Although the socio-economic disparities among small areas were not typical of those across the whole of the country, there were differences in levels of deprivation and crime, social fragmentation, and ethnic composition. A Bayesian regression analysis using Gibbs sampling in the programme WinBugs investigated whether there was variability in SDQ at both individual and area (MSOA) level, and whether the predictor variables at both levels were significantly associated with SDQ. Individuals from Asian or Black ethnic groups, and those in families with harmonious relationships and no financial stress had significantly lower SDQ scores, i.e. better health. Those who had special educational needs or long standing illness, or were from reconstituted families had significantly worse SDQ scores. About 6% of the variation in SDQ was associated with area differences. However, this area variation was not related to differences in area indicators of socio-economic deprivation, crime or social fragmentation. There was a complex association between SDQ and ethnic composition of neighbourhoods.

Introduction

The research reported here analysed data from the Research with East London Adolescents: Community Health Survey (RELACHS), an ethnically diverse sample of adolescents, living in an area of London, UK, which is mixed in terms of economic and social conditions and ethnic composition of the population. The aim was to explore whether mental health of adolescents in the sample varied in relation to neighbourhood of residence, after allowing for relevant individual characteristics. If so, we were also interested to determine whether the area differences were associated with material poverty, social disorder and fragmentation or with ethnic composition in the neighbourhood. The characteristics of the sample provided an unusual opportunity to examine ethnic group differences while controlling for socio-economic attributes of the sample.

Our research was informed by a number of publications, discussed below, that have reported or reviewed research on the relationship between the psycho-social health and development of adolescents and the socio-economic conditions in their neighbourhoods of residence.

We were particularly concerned with studies that have combined data on individual people and their families with information on their residential area, and have examined neighbourhood variation in adolescent health and development after controlling for relevant individual and family factors. Before discussing the literature relating to area effects, it is therefore important to note the individual and family characteristics that are known to be associated with adolescent mental health and development and are often controlled for in studies of neighbourhood differences described below. It is beyond the scope of this paper to review the very large literature on these predictors of psycho-social health. The two most pertinent sources for this paper are the study of Mental Health of Children and Adolescents in Great Britain, a survey by the Office of National Statistics (ONS) based on a large national sample of people aged 5–15 years in Britain in 1999 (Meltzer, Gatward, Goodman, & Ford, 2000), and the Health Survey for England findings on Health of Young People 1995–7 also based on a national survey of the English population (McMunn, Bost, Nazroo, & Primatesta, 1998). These are especially relevant here since both surveys used the Strengths and Difficulties Questionnaire (SDQ) as to identify likely cases of mental disorder (which in the ONS survey were verified using other clinical assessments). The SDQ was also employed in the RELACHS study analysed here. In summary, Meltzer et al. (2000) and McMunn et al. (1998) reported that the risk of psychiatric disorder increased with age. For 11–15 year olds the general prevalence of psychiatric disorders was higher among boys, but emotional distress, anxiety and depression was more common in girls. Children with mental illness were more likely to have physical illness or poor general health, and to have special educational needs (SEN). These national surveys did not include sufficient numbers from minority ethnic groups to make detailed comparisons, but Meltzer et al. (2000) found that ‘Asian’ (especially Indian) children had lower rates of psychiatric disorder than ‘White’ or ‘African Carribean’ children. (Results from the RELACHS survey reported elsewhere (Stansfeld et al., 2004) have also shown that controlling for age, sex and socio-economic variables, Bangladeshi respondents in East London were less likely than other groups to report distress on the SDQ.) Meltzer et al. (2000) and McMunn et al. (1998) also reported family factors associated with mental disorder. Risk of disorder is associated with socio-economic position, being lowest where the head of household is in a professional social class and higher in families of unskilled or unemployed workers. Children of lone parents, or living in reconstituted families have higher rates of disorder compared with families where both natural parents are married and living with the family. Children in large families of 5 children or more had a higher risk of disorder than those in small families. Discordant parental relationships and stressful events were also more common in families of children with mental disorders. These results suggest that a range of individual and family factors might need to be controlled for in studies aiming to identify neighbourhood variation. However, further analyses of the ONS data (Ford, Goodman, & Meltzer, 2004) has demonstrated that several of these variables are closely intercorrelated, so that it is important to consider carefully whether they all have independent effects on adolescent mental health. Furthermore, the correlations with individual attributes varied depending which type of psychiatric morbidity was considered. McMunn, Nazroo, Marmot, Boreham, and Goodman (2001) also report analyses of the Health Survey for England showing that the association between lone-parenthood and mental distress measured by the SDQ is explained by economic factors.

Turning to the more specific focus of this paper, we refer here to some fairly comprehensive overviews of research on neighbourhood factors in adolescent health produced up to 2002, and update these with reference to some more recently published research. Previous reviews have been made, for example, by Duncan, Brooks-Gunn, and Klebanov, 1994, Kupersmidt, Griesler, DeRosier, Patterson, and Davis (1995), Brooks-Gunn, Duncan, and Aber (1997), Lai (1999), Leventhal and Brooks-Gunn (2000) and Sampson, Morenoff, and Gannon-Rowley (2002). Many of these suggested that neighbourhood conditions and processes were associated with adolescent's mental health and well-being, and several neighbourhood factors similar to those tested in our research are identified as significant.

For example, Sampson et al. (2002) reviewed about 40 research papers published up to 2001, which had examined the relationships between neighbourhood factors and adolescent behaviour disorders or health related outcomes. They comment on the difficulties in comparing results of studies carried out in different settings, using different measures and for different groups of adolescents. They nevertheless concluded (Sampson et al., 2002, p. 459) that “overall it appears that concentrated poverty, disorder and low neighbourhood cohesion are linked to mental distress”.

In addition to this review by Sampson and colleagues, several of the other authors listed above identify broadly similar groups of neighbourhood factors relevant to adolescent health and propose hypothetical pathways by which neighbourhood conditions may influence child and adolescent health and development. For example Brooks-Gunn et al. (1997), Leventhal and Brooks-Gunn (2000) invoke: economic differences that result in varying access to institutional and collective resources and facilities for collective socialisation of young people; the impact of social cohesion and stability on relationships with family and peers in the community; social factors influencing ‘contagion’, whereby young people adopt the norms for behaviour and attitudes prevailing in the community, and the efficacy with which members of the community reinforce these norms. Kupersmidt et al. (1995) also put forward several theories to explain how socio-economic conditions in neighbourhoods might be associated with young people's well-being. One possibility is that exposure to health damaging risks may be multiplied in poor neighbourhoods resulting in poorer health in deprived areas. Alternatively, more beneficial circumstances in wealthier neighbourhoods may ‘buffer’ young people against family disadvantage or may enhance the advantages for those who have fewer individual or family risks, resulting in better health in such settings. Also, the person–environment fit may be important for adolescents and disparities between the individual's own specific circumstances and conditions typical of their neighbourhood social and economic surroundings may be damaging to mental health (Kupersmidt et al., 1995).

Thus, a considerable body of theoretical and empirical research had built up in the 1990s relating to adolescent mental health and neighbourhood factors. Since 2001, further studies specifically focused on individual and neighbourhood factors associated with adolescent mental health or distress have added to the evidence concerning the links with poverty, social disorder and lack of social cohesion (or social fragmentation) in neighbourhoods. These suggest that associations may be quite complex and may vary between settings or across demographic groups.

Recent evidence concerning the relationship between area poverty and adolescent mental health includes a study from the US by Beyers, Bates, Pettit, and Dodge (2003) showing that boys were particularly likely to display behavioural problems in less affluent areas. Goodman, Huang, Wade, and Kahn (2003) found that poverty at both household and school levels was associated with depressive symptoms in adolescents. Schneiders et al. (2003) found that in Rotterdam, Netherlands, a composite indicator of area disadvantage, including measures of both economic deprivation and social fragmentation, was associated with adolescent psychological morbidity, after controlling for parental socio-economic status. Drukker, Buka, Kaplan, McKenzie, and van Os (2003, 2005) also found that area deprivation was associated with worse self-reported general health for adolescents in Maastricht, Netherlands, but that mental health outcomes specifically did not show this association. Wickrama and Bryant (2003) found in a national sample of adolescents in the USA, that the beneficial influence of parent–child relationships on mental health are weaker in disadvantaged areas. In contrast, Ford et al. (2004) reported from their analysis of the British ONS survey, already discussed, that, after controlling for the sorts of individual factors summarized above, neighbourhood deprivation was not significantly associated with the risk of psychiatric disorder.

In respect of social cohesion/fragmentation and residential stability, some recent evidence suggests that these factors may operate on adolescent mental health through processes related to efficacy in enforcing social norms. Beyers et al. (2003) report research suggesting that the effects of parental monitoring were more pronounced in areas of greater social stability. A study in the Dutch city of Maastricht by Drukker, Kaplan, Feron, and van Os (2003) used indicators of socio-economic conditions in neighbourhoods of Maastricht including survey data on perceived social capital (including levels of informal social control operating in neighbourhoods). This demonstrated that the measure of informal social control in neighbourhoods was most strongly associated with mental health. Xue, Leventhal, Brooks-Gunn, and Earls (2005) found that residential mobility in the neighbourhood was not associated with mental health of adolescents. However, neighbourhood variation in the proportion of adults involved in organizations was independently associated with variation in adolescent mental state.

The relationships between the neighbourhood prevalence of violent crime, individual exposure to crime and the impacts for mental health may be more complex than assumed before. Simons et al. (2002) reported research on African American children in the US states of Iowa and Georgia, which indicated that exposure to criminal victimization was most strongly associated with a higher risk of depression among young people from poor neighbourhoods compared with those from less deprived areas. However, Rosenthal and Wilson (2003) found that although exposure to violence in the community was associated with psychological distress of individual adolescents, the risk of exposure was not very well predicted by neighbourhood characteristics and Deardorff, Gonzales, and Sandler (2003) report analyses suggesting that stressful events occurring in the domain of the family and of the peer group were more important for adolescents’ mental health than stressors in the wider community.

There has also been a growing interest in whether some neighbourhood factors are particularly important for specific groups of adolescents, and in particular circumstances. For example, Gordon et al. (2003) found that attention-deficit and hyperactivity and impairment of daily functioning were more common for adolescents for whom relative income level of the family was ‘mismatched’ with relative poverty/wealth of the surrounding neighbourhood. They invoke explanations put forward by Kupersmidt et al. (1995) concerning ‘person–environment fit’.

In similar vein, and of particular interest for the present study in an ethnically diverse community, are calls for more research on the differences in these relationships by ethnic group. Following research on ‘ethnic density’ as a factor in health of adults from minority groups (Halpern & Nazroo, 2000; Neeleman, Wilson-Jones, & Wessely, 2005; Subramanian, Acevedo-Garcia, & Osypuk, 2005) researchers are beginning to explore whether neighbourhood ethnic/racial composition and racial solidarity may have variable relevance for adolescent mental health, depending on the ethnicity of the individual. For example, Wickrama, Noh, and Bryant (2005) comment that relatively few studies have investigated how community, family and race factors at different levels influence mental health of young people and that research on neighbourhood factors has focused mainly on economic deprivation rather than other aspects of the social environment such as ethnic composition. They found that for black adolescents, risks of mental distress fell as the ethnic diversity of the community increased, which supports an argument that denser concentrations of ethnic minorities is beneficial for their psycho-social health. However, Xue et al. (2005) found that ethnic minority concentration in the area was not associated with adolescent mental health.

From the research literature summarized above, we concluded that the evidence concerning area effects on adolescent health is complex and the relationships are not very clearly established, so that there is a case for further work in this field. There is still uncertainty about the significance of neighbourhood effects for adolescents from very socially and ethnically diverse populations of inner cities, which are not always well represented in national survey samples. There is particular interest in the relevance for adolescent mental health of area indicators of economic deprivation, social fragmentation, crime and disorder, and ethnic composition. Analyses which aim to establish independent area effects need to control for a range of individual and family characteristics which may confound the relationships of interest. These may include data on children's age, sex, ethnicity, physical health, and SEN, as well as family social and economic circumstances, parental relationships and family composition.

Section snippets

An empirical analysis of cohort data from a poor inner city area

The present study concerns the relationship between adolescent mental health and neighbourhood conditions using data from a cohort of 2790 individual adolescents sampled from schools in an inner city area of east London, UK. As discussed below, the survey had shown some variability in mental health among these young people, measured by the SDQ, which this study seeks to explain. The study area is characterized by generally high levels of socio-economic deprivation. However, there is a degree of

Method of analysis

The dependent variable in our analysis was psychological distress. The individual/family variables were attributed to the same ‘individual’ level, comprising categorical variables for sex, age group, ethnicity, parental relationships, whether the child lives with a lone parent or in a reconstituted family, and variables indicating where there were no reported income problems, where both parents were unemployed, and where children had special needs or long-term illness, as well as continuous

Characteristics of the sample population

The characteristics of the individuals analysed here are summarized in Table 1. The RELACHS survey cohort comprised a roughly equal proportion of boys (51%) and girls (49%). Some characteristics of the sample reflected the high levels of ethnic minority concentration and of socio-economic deprivation in the area studied. In terms of self-ascribed ethnicity, the sample was very diverse, with 17 different ethnic groups represented in the sample; the majority came from an ethnic ‘minority’

Discussion

Our findings support the results of other research on variation in health at the individual and the area level, reviewed above, in that a relatively small, but significant proportion of the variation in SDQ in this sample adolescents is associated with area of residence. However, while there is some variability in area characteristics, such as material deprivation, crime and social fragmentation, these were not found to be strongly statistically associated with our measure of mental illness.

Acknowledgements

This research was supported by the Economic and Social Research Council, though project grant RES 000 22 0221. We would like to thank anonymous reviewers for their helpful comments on an earlier version of this paper.

References (39)

  • P. Congdon

    The epidemiology of suicide in London

    Journal of Royal Statistical Society, Series A

    (1996)
  • P. Congdon

    A multivariate model for spatio-temporal health outcomes with an application to suicide mortality

    Geographical Analysis

    (2004)
  • J. Deardorff et al.

    Control beliefs as a mediator of the relation between stress and depressive symptoms among inner city adolescents

    Journal of Abnormal Child Psychology

    (2003)
  • G. Duncan et al.

    Economic deprivation and early childhood development

    Child Development

    (1994)
  • T. Ford et al.

    The relative importance of child, family, school and neighbourhood correlates of childhood psychiatric disorder

    Social Psychiatry and Psychiatric Epidemiology

    (2004)
  • R. Goodman

    A modified version of the Rutter Parent Questionnaire including extra items on children's strengths: A research note

    Journal of Child Psychology & Psychiatry & Allied Disciplines

    (1994)
  • R. Goodman

    The strengths and difficulties questionnaire: A research note

    Journal of Child Psychology & Psychiatry & Allied Disciplines

    (1997)
  • R. Goodman

    Using the Strengths and Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders

    British Journal of Psychiatry

    (2000)
  • D. Halpern et al.

    The ethnic density effect: Results from a national community survey of England and Wales

    International Journal of Social Psychiatry

    (2000)
  • Cited by (0)

    View full text