The Modular Approach to Therapy for Children (MATCH) was developed to address the comorbidities common among clinically referred youth, with beneficial outcomes shown in 2 US randomized clinical trials, where it outperformed both usual clinical care and single disorder-specific treatments. To determine whether MATCH training of clinicians would result in more use of empirically supported treatment (EST) and better clinical outcomes than usual care (UC) in the publicly funded, multidisciplinary context of New Zealand. This multisite, single-blind, computer-randomized clinical effectiveness trial compared MATCH with UC in child and adolescent mental health services in 5 regions of New Zealand. Recruitment occurred from March 2014 to July 2015, and a 3-month follow-up assessment was completed by May 2016. Clinicians at participating child and adolescent mental health services were randomized (1:1) to undertake training in MATCH or to deliver UC, and young people with anxiety, depression, trauma-related symptoms, or disruptive behavior seeking treatment at child and adolescent mental health services were randomized (1:1) to receive MATCH or UC. Participants and research assistants were blind to allocation. Data analysis was performed from April 2016 to July 2017. MATCH comprises EST components for flexible management of common mental health problems. UC includes case management and psychological therapies. Both can include pharmacotherapy. There were 3 primary outcomes: trajectory of change of clinical severity, as measured by weekly ratings on the Brief Problem Monitor (BPM); fidelity to EST content, as measured by audio recordings of therapy sessions coded using the Therapy Integrity in Evidence Based Interventions: Observational Coding System; and efficiency of service delivery, as measured by duration of therapy (days) and clinician time (minutes). The study included 65 clinicians (mean age, 38.7 years; range, 23.0-64.0 years; 54 female [83%]; MATCH, 32 clinicians; UC, 33 clinicians) and 206 young people (mean age, 11.2 years; range 7.0-14.0 years; 122 female [61%]; MATCH, 102 patients; UC, 104 patients). For the BPM total ratings for parents, there was a mean (SE) slope of -1.04 (0.14) (1-year change, -6.12) in the MATCH group vs -1.04 (0.10) (1-year change, -6.17) in the UC group (effect size, 0.00; 95% CI, -0.27 to 0.28; P = .96). For the BPM total for youths, the mean (SE) slope was -0.74 (0.15) (1-year change, -4.35) in the MATCH group vs -0.73 (0.10) (1-year change, -4.32) in the UC group (effect size, -0.02; 95% CI, -0.30 to 0.26; P = .97). Primary analyses (intention-to-treat) showed no difference in clinical outcomes or efficiency despite significantly higher fidelity to EST content in the MATCH group (58 coded sessions; mean [SD], 80.0% [20.0%]) than the UC group (51 coded sessions; mean [SD], 57.0% [32.0%]; F(1,108) = 23.0; P < .001). With regard to efficiency of service delivery, there were no differences in total face-to-face clinician time between the MATCH group (mean [SD], 806 [527] minutes) and the UC group (mean [SD], 677 [539] minutes) or the overall duration of therapy between the MATCH group (mean [SD], 167 [107 days]) and the UC group (mean [SD], 159 [107] days). MATCH significantly increased adherence to EST practices but did not improve outcomes or efficiency. The nonsuperiority of MATCH may be attributable to high levels of EST use in UC in New Zealand. Australian New Zealand Clinical Trials Registry Identifier: ACTRN12614000297628.