Dance-Based Interventions for Mental Health: A Comprehensive and Quality-Sensitive Umbrella Review of Definitions,...

Dance-based interventions are increasingly used to support mental health and well-being across clinical, community, and educational contexts. However, the review-level evidence remains conceptually and methodologically fragmented. This quality-sensitive umbrella review synthesized 90 review articles identified across five databases. Systematic reviews and meta-analyses were appraised using AMSTAR 2, while scoping, narrative, literature, and clinical reviews were assessed using adapted review-appraisal criteria. Findings from systematic reviews and meta-analyses were used primarily to interpret effectiveness evidence, whereas nonsystematic reviews were used to map conceptual definitions, practice traditions, and measurement issues. The synthesis showed that the field is fragmented at three interrelated levels. First, intervention labels and delivery models are inconsistent, ranging from clearly defined dance/movement therapy to broad and underspecified uses of dance. Second, effectiveness is conceptualized differently across intervention traditions, with therapy-oriented reviews emphasizing symptom reduction and non-therapy-oriented reviews foregrounding well-being, vitality, social connection, and embodied experience. Third, outcome assessment relies predominantly on generic psychological scales, with limited use of movement-sensitive or dance-specific measures. Quality-sensitive evidence suggests that dance-based interventions show promise for well-being, social connection, emotional regulation, and selected symptom outcomes, but confidence remains limited by low review quality, heterogeneous populations, inconsistent intervention reporting, and insufficient attention to primary-study overlap. Future research should move beyond asking whether dance works in general and instead specify what kind of dance, for whom, through which mechanisms, and according to which model of mental health and well-being.