Registries and biobanks for juvenile dermatomyositis (JDM) have generated statistical power to help understand pathogenesis and determine treatment and long-term outcomes in this rare and heterogeneous disease. Genotype, autoantibodies, muscle histology and early clinical features may predict prognosis and guide personalised treatment. While corticosteroids and disease-modifying anti-rheumatic drugs improve outcomes, there remain children who experience refractory disease. Ongoing research into the aberrant immune response and novel biological targets is necessary. Best practice guidelines promote prompt stepwise treatment, and there is growing appreciation of the role of exercise in improving prognosis. Validated tools standardise assessment of disease activity and damage in musculoskeletal, mucocutaneous, pulmonary, cardiac, gastrointestinal and endocrine systems. Recently, an internationally agreed dataset for JDM has been defined for clinical practice and incorporation into registries. In the future, with bigger datasets, statistical models may guide stratification for personalised medicine and discern the most relevant outcome markers for research.