ImportanceNonsuicidal self-injury is prevalent in adolescence and associated with adverse clinical outcomes. Effective interventions that are brief, transportable, and scalable are lacking.ObjectiveTo test the hypotheses that an internet-delivered emotion regulation individual therapy for adolescents delivered adjunctive to treatment as usual is superior to treatment as usual only in reducing nonsuicidal self-injury and that improvements in emotion regulation mediate these treatment effects.Design, Setting, and ParticipantsThis 3-site, single-masked, randomized superiority trial enrolled participants from November 20, 2017, to April 9, 2020. Eligible participants were aged between 13 and 17 years and met diagnostic criteria for nonsuicidal self-injury disorder; they were enrolled as a mixed cohort of consecutive patients and volunteers. Parents participated in parallel to their children. The primary end point was at 1 month after treatment. Participants were followed up at 3 months posttreatment. Data collection ended in January 2021.InterventionsTwelve weeks of therapist-guided, internet-delivered emotion regulation individual therapy delivered adjunctive to treatment as usual vs treatment as usual only.Main Outcomes and MeasuresPrimary outcome was the youth version of the Deliberate Self-harm Inventory, both self-reported by participants prior to treatment, once every week during treatment, and for 4 weeks posttreatment, and clinician-rated by masked assessors prior to treatment and at 1 and 3 months posttreatment.ResultsA total of 166 adolescents (mean [SD] age, 15.0 [1.2] years; 154 [92.8%] female) were randomized to internet-delivered emotion regulation therapy plus treatment as usual (84 participants) or treatment as usual only (82 participants). The experimental intervention was superior to the control condition in reducing clinician-rated nonsuicidal self-injury (82% vs 47% reduction; incidence rate ratio, 0.34; 95% CI, 0.20-0.57) from pretreatment to 1-month posttreatment. These results were maintained at 3-month posttreatment. Improvements in emotion dysregulation mediated improvements in self-injury during treatment.Conclusions and RelevanceIn this randomized clinical trial, a 12-week, therapist-guided, internet-delivered emotion regulation therapy delivered adjunctive to treatment as usual was efficacious in reducing self-injury, and mediation analysis supported the theorized role of emotion regulation as the mechanism of change in this treatment. This treatment may increase availability of evidence-based psychological treatments for adolescents with nonsuicidal self-injury.Trial RegistrationClinicalTrials.gov Identifier: NCT03353961