Article Text

Download PDFPDF
Mental health service use by adolescents of Indian and White origin
  1. Panos Vostanis1,
  2. Nadzeya Svirydzenka1,
  3. Pat Dugard2,
  4. Swaran Singh3,
  5. Nisha Dogra1
  1. 1Department of Psychology, University of Leicester, Leicester, UK
  2. 2Independent Senior Statistician, King's Lynne, UK
  3. 3Department of Health Sciences, University of Warwick, Coventry, UK
  1. Correspondence to Professor Panos Vostanis, University of Leicester, School of Psychology, Greenwood Institute of Child Health, Westcotes House, Westcotes Drive, Leicester LE3 0QU, UK; pv11{at}


Background Despite the available epidemiological evidence on the prevalence of mental health problems in childhood and adolescence, there is limited knowledge on whether there are differences in the level of need and service utilisation by young ethnic minority groups.

Methods Adolescents of 13–15 years from nine schools in two English cities in which children of Indian ethnicity were over-represented (n=2900), completed rating scales on different types of mental health problems, contacts with services and informal supports.

Results Indian adolescents scored significantly lower on general mental health and depression symptoms. They were also less likely than White adolescents to self-report having mental health problems, even for a similar level of need. Among those with mental health scores within the clinical range, Indian adolescents were less likely to have visited specialist services. Instead, they were more likely to first approach family members, teachers or general practitioners.

Conclusions Rather than a blanket approach being applied to policy and service planning to meet the needs of diverse communities of young people, more specific evidence needs to be gained about patterns of referrals of minority groups and their strategy of accessing supportive adults.

  • Child Psychiatry
  • Health services research
  • School Health

Statistics from

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.

What is already known on this topic

  • There is a high level of unmet mental health needs among adolescents.

  • Adolescents are more likely to seek help from family and schools before contacting mental health services.

What this study adds

  • Adolescents of Indian ethnic origin perceive and report lower rates of mental health needs than White adolescents attending the same schools.

  • Indian adolescents are more likely to first seek help from their siblings, extended family or teachers.

  • Accounting for differences in mental health needs, Indian adolescents have lower service contacts with mental health services.


One in 10 children and adolescents has mental health problems that would benefit from assessment and potential treatment. Surveys in different countries have highlighted the extent of unmet needs in accessing specialist mental health services and a lack of accessible, appropriate, effective co-ordinated pathways involving health, education, social care and non-statutory sectors.1 ,2 Adolescents often prefer to seek informal help from support networks such as friends or teachers before approaching specialist child and adolescent mental health services (CAMHS).3 ,4

Despite the increasing cultural diversity in high income countries and the recognised importance of evidence-based policy, it is not as yet clear whether adolescents from ethnic minority backgrounds have different types or severity of mental health problems, and if they are less likely to access mental health and related services than the general population. This research gap can be explained by the lack of sufficient sample size of ethnic minority adolescents in general population surveys.5 The rationale for this study therefore was to establish whether adolescents of Indian and White British ethnicity reported different patterns of contacts with specialist and mental health services and informal supports, accounting for their levels of mental health needs.


The sample was selected by inviting 16 schools to take part, in which children of Indian ethnicity were over-represented compared to the general population based on Census data.6–8 Nine schools in an English city and a London borough participated in the study. Prior to recruitment, the study received University of Leicester research ethics approval. Schools acted as gatekeepers for recruitment and parents consented to their child's participation on the opt-out basis. Pupils were also asked for their consent prior to completing the measures. Overall, there was a response rate of 70% by the young people.

All 13–15-year olds in the nine schools were invited to participate and those who did completed UK-validated measures of general mental health (Strengths and Difficulties Questionnaire (SDQ)),9 depressive (Mood and Feelings Questionnaire short version, (MFQ-Short)),10 eating (SCOFF – acronym for Sick Control One Fat Food),11 and drug and alcohol problems, and of informal and formal services they had either used or would use for help with mental health issues.

The study sample comprised 2900 adolescents, with a mean age of 13.97 years (SD=0.7). There were 1284 (44%) boys and 1610 (56%) girls, of whom 1087 (37%) were British Indian and 414 (14%) British adolescents (including White British/English/Welsh/Scottish/Irish classifications—hereafter referred to as White); it is these two groups we will be focusing on in the analysis. Other White adolescent subgroups such as those from Eastern Europe were not included in this analysis. Both ethnic groups showed comparable socio-demographic composition, with N=159 (14.6%) and N=70 (16.9%) Indian and White British adolescents eligible for free school meals, respectively. We also recorded their family circumstances: living with parents or extended family, or living with a single parent or step family. The percentages of those self-reporting a mental health problem were 16% for those children living with single parents or step families, and 8% for those with parents or extended family.

Mental health questionnaire scores were dichotomised according to previously established cut-offs that distinguished clinical from non-clinical subjects, that is, adolescents who would benefit from assessment and possible intervention. SDQ scores vary from 0 to 40 with 0–15, 16–19 and 20–40 falling within the normal, borderline and abnormal (clinical) range (the latter group being likely to present with mental health problems). An MFQ-Short score of 8 and above indicated a young person at risk of depression. Those young people with scores of 3 and higher on a 0–5 SCOFF score range were at risk of having or developing an eating disorder. χ2 tests of association and log-linear model analysis were completed for between-group comparisons (degrees of freedom in brackets). We used logistic regression to predict the probability of self-reporting mental health problems while considering our measured or categorical variables together (like ethnicity, who they live with, etc). In the regression, we used a single dichotomous variable as a mental health indicator (MHindicator=1 if there is an abnormal score on SDQ or MFQ-Short, 0 otherwise).


Indian adolescents were significantly less likely to score within the clinical range of general mental health and depression symptoms than White adolescents (SDQ borderline (13% vs 17%) and abnormal (5% vs 13%) scores, p=0.000; MFQ-Short abnormal: 21% vs 30%, p=0.001, respectively—figures 1 and 2). These differences were not associated with different economic status as reflected by being on free school meals.12 There were no differences between the two school areas. In both ethnic groups, more females scored within the abnormal range on the MFQ-Short scale (Indian adolescents:15% male vs 26% female, p=0.000; White British adolescents: 21% male vs 35% female, p=0.004). SCOFF scores revealed gender differences only in the White sample (4% male vs 10% female, p=0.019). There were no gender differences on SDQ-Total scores in either gender group. For both ethnic groups, those living with a single parent or a step family were more likely to be in the borderline or abnormal ranges of SDQ and MFQ-Short scales and to self-report mental health problems. Of those living with single parents or step families, 34% had a score in the abnormal range on SDQ or MFQ-Short, but only 26% of those living with parents or extended family.

Figure 1

Rates of Strengths and Difficulties Questionnaire (SDQ)-total scores in Indian and White adolescents.

Figure 2

Rates of Mood and Feelings Questionnaire (MFQ)-Short scores in Indian and White adolescents.

Within the total sample, Indian adolescents were significantly less likely to self-report mental health problems than their White British counterparts (7% vs 16%, p=0.000). When only those adolescents with borderline and abnormal overall mental health scores (SDQ) were considered, Indian adolescents were still significantly less likely to self-report mental health problems (10% vs 32%, p=0.000 for borderline; and 22% vs 44%, p=0.000 for abnormal). A similar pattern was found among those with abnormal depression (MFQ-Short) and eating (SCOFF) scores (MFQ-Short: 21% vs 37%, p=0.003; SCOFF: 12% vs 38% p=0.003). So, for similar levels of abnormality as measured by these three scales, Indian adolescents were less likely to self-report mental health problems.

In a logistic regression analysis, we first included the variables of gender, free school meals, MHindicator, and who they live with as predictors of self-report of mental illness along with all two-way and three-way interactions. As a result of this analysis, only the MHindicator and the interaction between the MHindicator and who they live with made a significant contribution. We retained these two variables in our logistic regression model. This model was a significant predictor of the probability of a self-report (χ2(3)=73.94, p=0.000). Residuals, leverage and Dfbeta values did not indicate lack of fit. We then included ethnic group variable to the model to see whether, after allowing for other variables that may influence self-reporting, ethnic group still makes an additional contribution. Ethnic group is indeed significant (χ2(1)=11.82, p=0.001). The odds of Whites self-reporting are 2.24 times the odds of Indians self-reporting (95% CI 1.42 to 3.50). Therefore, the lower rate of self-reporting for Indian children is not accounted for by other variables which could have affected self-reporting.

Among those adolescents who self-reported mental health problems, a much higher proportion of White than Indian adolescents had used or would seek help from CAMHS (41% vs 13%; p=0.001). When considering only those who scored in the SDQ abnormal or borderline range, a lower proportion of Indian than White adolescents had previously visited CAMHS (14% vs 35%; p=0.032). The same pattern of CAMHS visits was found for adolescents scoring within the clinical depression range on the MFQ-Short (11% Indian vs 27% White; p=0.032).

Logistic regression was less powerful in determining predictors of CAMHS, as only about 20% of our sample responded to the question (they were not required to answer the question about CAMHS use, unless they self-reported a mental health problem). However, after using gender, MHindicator, who they live with, whether they get free school meals, whether they self-report a mental health problem and the two-way and three-way interactions among them, only self-reporting was a significant predictor (χ2(1)=8.99, p=0.002). When we added ethnic group to the model, it just failed to reach significance (χ2(1)=3.71, p=0.054). The model estimates that the odds of Whites using CAMHS are 2.21 times the odds of Indians using it, but the 95% CI overlaps 1 (0.99 to 4.50), reflecting the failure to reach significance at 5%.

χ2 tests and logistic regressions indicate that Indian adolescents self-reported fewer problems, even when taking into account their lower levels of need, and they were also less likely to use services, even when they needed to. There was no difference between the two groups on perceived helpful factors or barriers in accessing services.

In relation to informal supports, and who they would be approaching first, there was no significant difference in the proportions of young people who had talked or would talk to parents, but Indian adolescents were more likely to approach siblings (p=0.002) or other family members (p=0.007). In contrast, White adolescents were more likely to talk to their friends first (p=0.046). At school, significantly more Indian adolescents were likely to talk to a teacher (p=0.000), as well as visit their GP (p=0.000), while there were no differences in the proportions approaching social workers or paediatric hospital services. There were no between-group differences in other help-seeking categories, with only 5% being comfortable in using a self-help group and 9% a helpline, which rose to 18% for accessing educational literature, 28% for relevant websites and 8% for chat rooms. Finally, a similar proportion of 17% would approach child mental health services in the future.


Many health policies, including those on the provision of child mental health services, explicitly refer to the diverse and often unmet needs of ethnic minority groups. These policies are, however, based on a variable degree and quality of evidence on whether such assumptions are correct.13 The findings of this study on a particular and substantive ethnic population in the UK, that is, adolescents of Indian ethnic origin, indicate interesting patterns in their perceived and actual mental health problems, and their contacts with services and informal supports.

Indian adolescents were found to have lower mental health needs than White adolescents attending the same schools12 and, when considering only those in the abnormal or borderline range of mental health rating scales, they perceived themselves to have fewer problems and had lower service utilisation rates. In contrast, they did not perceive different factors that hindered them from using services. These findings indicate the existing clinical trends in CAMHS use among ethnic minority groups. Whereas they might generalise to the ethnic groups described in this study living in similar conditions for the reasons we describe below, one must exercise caution at generalising them to other groups and environments.

Before we consider the implications, we acknowledge several limitations of the presented study. Self-report measures rather than direct clinical interviews were used to establish prevalence of mental health problems. Although less accurate than clinical interviews, the measures used are well established and with satisfactory psychometric properties. Second, the sampling procedure varied in some schools, as some preferred to manage the process internally. However, the administration of measures followed a strict procedural script. Finally, some young people in our sample were first-generation immigrants (meaning they had been born elsewhere and had recently moved to the UK); therefore, English might not have been their first language. Since all our instruments were presented in English, we need to consider whether particular mental health terms were interpreted differently between native and non-native English speakers. Although this consideration is mitigated by the fact that all the adolescents showed proficient levels of English to be able to attend a public secondary school in the UK, a study investigating first-generation immigrants’ interpretation of mental health terms would be a useful addition to the literature, but is, however, outside the scope of this paper.

Differences in use of CAMHS between Indian and White British adolescents may reflect referrer bias in who is referred, and that GPs may be less likely to refer Indian children, for which there may be various explanations. When referral patterns were analysed for different agencies and professional groups, Daryanani et al13 found that general practitioners were more likely to over-refer white children; while specialist doctors (paediatricians) were more likely to refer Black and South Asian children, education services to refer Black children, and social workers to refer mixed race children. Another reason for lower rates of CAMHS visits by adolescents of ethnic minority descent is that specialist services may not be easily accessible to the local ethnic community or they might not be aware of existing specialist mental health services, especially if the family are non-British recent migrants. One way to solve this issue would be the facilitation of ethnic community engagement with the specialist service through a number of flexible approaches, like family involvement, as family units are likely to play a role of trusted advisors within that community. Previous research found that young people were remarkably consistent with their views that specialist services needed to be more visible in raising awareness and having a physical presence.14 These opinions clearly correspond to the finding that none of the participants indicated an awareness of specific services for young people with mental health problems, although they did talk about professionals who might be able to help. Young people from the Dogra et al4 study also suggested that services might be improved by better advertisement through TV, magazines, library posters, or leaflets, geographical diversity of clinics and talking in schools with professionals, but also through educating teachers.

Ethnic differences in self-report of mental health problems as well as lack of differences in reported barriers to using the specialist mental health services suggest differences in the conceptualisation of what constitutes a mental health problem or disorder. It can also cue ethnic differences in underpinning fears of stigma, which in turn impact on their help-seeking behaviours.15 It could also be explained by the availability of more support within their extended family or community networks, with help being sought from external services when these links are weak or absent, when their problems become more severe, or when these problems are related to issues within the family, that is when these Indian young people do not have access to the family support available to their peers. Being more likely to approach teachers could reflect higher expectations of adults and/or clearer generational boundaries in Indian adolescents. They were also more likely to visit their general practitioners, but less likely to access specialist CAMHS.

The role of ethnicity in adolescent mental health and their access to services is evident; however, it is not generally the ethnic membership itself that manifests in specific needs or barriers for accessing services, but rather the cultural and behavioural values and norms it carries with it. In this paper, we have considered additional socio-demographic factors as contributors to lower access to CAMHS in Indian adolescents in order to account for the complex nature of the issue; however, these were not significant. Where we are confident in the approach taken to explore the needs of particular ethnic communities, different methodologies need to be employed by future research in contributing to our understanding of why these needs arise and the best ways to alleviate them.

The diverse needs of these groups can be oversimplified by policy guidance which assumes homogeneity. Finally, although there is evidence that stigma is consistent across communities,16 future research should explore whether ethnicity affects attitudes towards mental health, which in turn should inform antistigma mental health campaigns. Improved links between adolescent and adult services would minimise the possibility of early mental health problems resurfacing through more severe presentations during adult life.


We are grateful to all the young people and their schools for their participation in this study; and to Dr David Clarke from the Leicestershire Partnership NHS Trust for his support.



  • Contributors PV: conceptualisation, data analysis and writing up. NS: organisation, data collection and analysis. PD: data analysis. SS: conceptualisation and writing up. ND: conceptualisation and writing up.

  • Funding This study was supported by the LNR Clinical Research Networks.

  • Competing interests None.

  • Ethics approval Leicester University Ethics Committee.

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