Objective: To investigate whether dance intervention influenced self-rated health for adolescent girls with internalizing problems. Design: Randomized controlled intervention trial with follow-up measures at 8, 12, and 20 months after baseline. Setting: A Swedish city with a population of 130 000. Participants: Girls aged 13 to 18 years with internalizing problems, ie, stress and psychosomatic symptoms. A total of 59 girls were randomized to the intervention group and 53 were randomized to the control group. Intervention: The intervention comprised dance classes twice weekly during 8 months. Each dance class lasted 75 minutes and the focus was on the joy of movement, not on performance. Main Outcome Measures: Self-rated health was the primary outcome; secondary outcomes were adherence to and experience of the intervention. Results: The dance intervention group improved their self-rated health more than the control group at all follow-ups. At baseline, the mean score on a 5-point scale was 3.32 for the dance intervention group and 3.75 for the control group. The difference in mean change was 0.30 (95% CI, -0.01 to 0.61) at 8 months, 0.62 (95% CI, 0.25 to 0.99) at 12 months, and 0.40 (95% CI, 0.04 to 0.77) at 20 months. Among the girls in the intervention group, 67% had an attendance rate of 50% to 100%. A total of 91% of the girls rated the dance intervention as a positive experience. Conclusions: An 8-month dance intervention can improve self-rated health for adolescent girls with internalizing problems. The improvement remained a year after the intervention. JAMA Pediatr. 2013;167(1):27-31. Published online November 12, 2012. doi:10.1001/jamapediatrics.2013.421