Influence of racism and context on ethnic differences in adolescent mental health trajectories : the Determinants of Adolescent Social well-being and Health (DASH) study

Objective: There is some evidence that ethnic density in neighbourhoods may buffer the adverse effect of racism on mental health in adulthood. We previously reported generally better mental health for ethnic minority groups in early adolescence, despite socioeconomic disadvantage. Here we investigate whether this resilience persists in late adolescence, and the interplay between racism and the context of schools and neighbourhoods. Methods: Repeated measures of psychological well-being (Total Difficulties Score (TDS) from Goodman's Strength and Difficulties Questionnaire), racism and characteristics of individuals (socio-economic position and family type), schools (n=49) and neighbourhoods (n=3495) were obtained for 4744 adolescents aged 11-17, surveyed at two waves (2003, 2005/06). Ethnic groups included: White-UK (n=870); Mixed-Black Caribbean/White (n=260); Indian (n=417); Pakistani (n=295); Bangladeshi (n=149); Black Caribbean (n=770); Nigerian/Ghanaian (n=501); Other-African (n=380); Other ethnicities (n=1102). Context measures included neighbourhood White-ethnic density and deprivation, school White-ethnic density, academic performance, and free school-meals. Results: The prevalence of racism increased between wave 1 (girls: 18%; boys: 19%) and wave 2 (girls: 29%; boys: 27%). Ethnic minority adolescents reported better mental health (lower TDS) throughout adolescence relative to White-UK peers, particularly among Nigerian/Ghanaian boys (coefficient (95% CI)): -2.27 (-2.97, -1.58) and Indian girls: -1.74 (- 2.54, -0.94), adjusted for age, socio-economic position, racism and context. TDS improved throughout adolescence for boys. Effects of racism were not ethnic specific, but associated with higher TDS for boys: 1.50 (1.21, 1.79) and girls: 1.88 (1.54, 2.22) through adolescence. Compared to White-UK peers, ethnic minority adolescents were disproportionately situated in less well-performing schools with higher rates of free-meals and lower White-ethnic densities, and more deprived neighbourhoods with lower White-ethnic densities. No variation in TDS was observed between schools or neighbourhoods. Context had little effect on TDS directly, or through modification of the association between racism and TDS. Conclusions: Poorer ethnic-specific mental health patterns among adults do not appear to emerge during adolescence. Boys from all ethnic minority groups and Indian girls reported better mental health. Racism was associated with poorer mental health for all ethnic groups. Targeting intervention in adolescence may be a critical opportunity for preventing ethnic differences in mental health in later life.