Economic inequalities in adolescents’ internalising symptoms: longitudinal evidence from eight countries

Abstract

Background: Research, mainly conducted in Europe and North America, has shown an inequitable burden of internalising mental health problems among adolescents from poorer households. We investigated whether these mental health inequalities differ across a diverse range of countries and multiple measures of economic circumstances. Methods: In this longitudinal observational cohort study, we analysed data from studies conducted in eight countries (Australia, Ethiopia, India, Mexico, Peru, South Africa, the UK, and Viet Nam) across five global regions. All studies had self-reported measures of internalising symptoms using a validated scale at two timepoints in adolescence; a measure of household income, household consumption expenditure, or subjective wealth; and data collected between 2000 and 2019. Household income (measured in four countries), consumption expenditure (six countries), and adolescents’ subjective assessment of household wealth (five countries) were measured in mid-adolescence (14–17 years). The primary outcome (internalising symptoms, characterised by negative mood, affect, and anxiety) was measured later in adolescence between age 17 and 19 years. Analyses were linear regression models with adjustment. Effect estimates were added to random-effects meta-analyses to aid understanding of cross-country differences. Findings: The overall pooled sample of eight studies featured 18 910 adolescents (9568 [50·6%] female and 9342 [49·4%] male). Household income had a small or null association with adolescents’ internalising symptoms. Heterogeneity (I2 statistic) was 71·04%, falling to 39·71% after adjusting for baseline symptoms. Household consumption expenditure had a stronger association with internalising symptoms (decreases of 0·075 SD in Peru [95% CI –0·136 to –0·013], 0·034 SD in South Africa [–0·061 to –0·006], and 0·141 SD in Viet Nam [–0·202 to –0·081] as household consumption expenditure doubled). The I2 statistic was 74·24%, remaining similar at 74·83% after adjusting for baseline symptoms. Adolescents’ subjective wealth was associated with internalising symptoms in four of the five countries where it was measured. The I2 statistic was 57·09% and remained similar after adjusting for baseline symptoms (53·25%). We found evidence for cross-country differences in economic inequalities in adolescents’ internalising symptoms, most prominently for inequalities according to household consumption expenditure. Subjective wealth explained greater variance in symptoms compared with the objective measures. Interpretation: Our study suggests that economic inequalities in adolescents’ mental health are prevalent in many but not all countries and vary by the economic measure considered. Variation in the magnitude of inequalities suggests that the wider context within countries plays an important role in the development of these inequalities.