Towards a new taxonomy for youth mental health: Clusters of adolescent well-being and socio-emotional difficulties within large community and school-based samples

Background: Current psychiatric taxonomies are limited, and interest in alternative approaches to organise youth mental health has increased. Aims: My research question was - how can we best group adolescents based on mental health difficulties and strengths? One aim was to therefore identify adolescent mental health groups that emerge using person-centred statistical analyses. A second aim was to classify adolescents into groups by applying a promising, but under-used statistical approach. Method: Regarding the first aim, a systematic review of the literature was conducted. Twelve studies were identified (n = 101,792 adolescents). To meet the other aim, a secondary analysis of a large dataset was calculated. An unsupervised machine learning algorithm was used to cluster adolescents (n = 25,568 adolescents) based on socio-emotional and behavioural difficulties as well as psychological well-being. Results: The systematic review revealed around six common cluster-types. They described patterns of flourishing, surviving, struggling, and languishing – the other two were less common and were characterised by a general absence of symptoms. The secondary analysis indicated eight clusters. These ranged from adolescents with high well-being to those with intense, co-occurrent problems. Also identified were groups experiencing discrete problems but with intact well-being. Conclusion: Findings suggest mental health groups that fall along a gradient. It spans from excellent mental health to widespread problems, with discrete patterns of difficulties nested between. The research reported here offers a different perspective on alternative mental health taxonomies, and new evidence that can contribute to this discussion.