Emotional distress in adolescents in 2018 and 2022: a comparison of cross-sectional national probabilistic samples from six countries.

Abstract

Most evidence on how the COVID-19 pandemic affected adolescent mental health relies on non-probability samples, which can yield unreliable estimates. This study aimed to compare emotional distress among individuals aged 15 years in 2018 and 2022 using nationally representative data, and to examine how sociodemographic correlates of distress differed between these periods. We analysed data from the 2018 and 2022 waves of the Programme for International Student Assessment (PISA) in five countries and one territory (Hong Kong, Ireland, Mexico, Panama, Spain, and United Arab Emirates). PISA is a periodic multi-country cross-sectional assessment of students aged 15 years. Emotional distress was measured using items capturing symptoms of anxiety, depression, irritability, and somatisation. Quantile regression models assessed distributional changes between 2018 and 2022; prevalence ratios were calculated to evaluate changes in the proportion of adolescents in 2022 scoring above 2018-based percentiles. We included 80 403 adolescents in 2018 and 77 708 in 2022. Emotional distress increased modestly at the median level (0·11 standardised units [95% CI 0·10-0·12]; p<0·0001), with sharper rises among females (0·20 [0·19-0·22]; p<0·0001) than males (0·02 [0·00-0·03]; p=0·015). However, increases were progressively higher at higher percentiles: 0·22 ([0·21-0·23]; p<0·0001) for the 85th percentile, 0·24 ([0·22-0·25]; p<0·0001) for the 90th percentile, and 0·32 ([0·30-0·34]; p<0·0001) for the 95th percentile. Accordingly, prevalence ratios of 2022 proportions above 2018 cutoffs increased progressively at the median (1·15 [1·13-1·16]), 85th (1·66 [1·62-1·70]), 90th (1·88 [1·83-1·94]), and 95th percentiles (2·37 [2·28-2·47]). Emotional distress among adolescents in most participating countries was higher in 2022 than in 2018, with disproportionate increases in the number of those reporting high degrees of distress. This pattern suggests a widening of vulnerabilities and highlights an urgent need for scalable and context-sensitive responses, such as strengthening school-based supports, integrating mental health into primary care, and expanding digital and community-based interventions. None.