Context: Migraine is a common neurologic disorder in children and adolescents. However, a comparison of multiple nonpharmacological treatments is lacking.

Objective: To examine whether nonpharmacological treatments are more effective than waiting list and whether there are differences between interventions regarding efficacy.

Data sources: Systematic review and network meta-analysis of studies in Medline, Cochrane, Embase, and PsycINFO published through August 5, 2019.

Study selection: Randomized controlled trials of nonpharmacological treatments in children and adolescents diagnosed with episodic migraine.

Data extraction: Effect sizes, calculated as standardized mean differences (SMDs) for the primary outcome efficacy, were assessed in a random-effects model.

Results: Twelve studies (N = 576) were included. When interventions were classified into groups on the basis of similarity of treatment components, self-administered treatments, biofeedback, relaxation, psychological treatments, and psychological placebos were significantly more effective than waiting list with effect sizes ranging between SMD = 1.14 (95% confidence interval, 0.09 to 2.19) for long-term psychological placebos to SMD = 1.44 (95% confidence interval, 0.26 to 2.62) for short-term self-administered treatments. However, when all interventions were examined individually (ie, 1 node per intervention), none were significantly more effective compared with waiting list, mainly because of lack of statistical power.

Limitations: Because of our focus on pediatric migraine, only a small number of studies could be included.

Conclusions: Our findings reveal that components of nonpharmacological interventions are effective in treating pediatric migraine. Some effects have to be interpreted carefully because they are based on small studies. Future researchers should identify factors associated with individual responses in large, multicentered studies.