Nationwide Implementation of Unguided Cognitive Behavioral Therapy for Adolescent Depression: Observational Study of SPARX.

Background: Internet-based cognitive behavioral therapy (iCBT) interventions are effective in clinical trials; however, iCBT implementation data are seldom reported. Objective: The objective of this study is to evaluate uptake, adherence, and changes in symptoms of depression for 12- to 19-year-olds using an unguided pure self-help iCBT intervention (SPARX; Smart, Positive, Active, Realistic, X-factor thoughts) during the first 7 years of it being publicly available without referral in Aotearoa New Zealand. Methods: SPARX is a 7-module, self-help intervention designed for adolescents with mild to moderate depression. It is freely accessible to anyone with a New Zealand Internet Protocol address, without the need for a referral, and is delivered in an unguided "serious game" format. The New Zealand implementation of SPARX includes 1 symptom measure-the Patient Health Questionnaire adapted for Adolescents (PHQ-A)-which is embedded at the start of modules 1, 4, and 7. We report on uptake, the number of modules completed, and changes in depressive symptoms as measured by the PHQ-A. Results: In total, 21,320 adolescents aged 12 to 19 years (approximately 2% of New Zealand 12- to 19-year-olds) registered to use SPARX. Of these, 63.6% (n=13,564; comprising n=8499, 62.7% female, n=4265, 31.4% male, and n=800, 5.9% another gender identity or gender not specified; n=8741, 64.4% New Zealand European, n=1941, 14.3% Māori, n=1202, 8.9% Asian, n=538, 4.0% Pacific, and n=1142, 8.4% another ethnic identity; mean age 14.9, SD 1.9 years) started SPARX. The mean PHQ-A at baseline was 13.6 (SD 7.7) with 16.1% (n=1980) reporting no or minimal symptoms, 37.4% (n=4609) reporting mild to moderate symptoms (ie, the target group) and 46.7% (n=5742) reporting moderately severe or severe symptoms. Among those who started, 51.1% (n=6927) completed module 1, 7.4% (n=997) completed at least 4 modules, and 3.1% (n=416) completed all 7 modules. The severity of symptoms reduced from baseline to modules 4 and 7. Mean PHQ-A scores for baseline, module 4, and module 7 for those who completed 2 or more assessments were 14.0 (SD 7.0), 11.8 (SD 7.9), and 10.5 (SD 8.5), respectively; mean difference for modules 1-4 was 2.2 (SD 5.7; P<.001) and for modules 1-7 was 3.6 (SD 7.0; P<.001). Corresponding effect sizes were 0.38 (modules 1-4) and 0.51 (modules 1-7). Conclusions: SPARX reached a meaningful proportion of the adolescent population. The effect size for those who engaged with it was comparable to trial results. However, completion was low. Key challenges included logistical barriers such as slow download speeds and compatibility with some devices. Ongoing attention to rapidly evolving technologies and engagement with them are required. Real-world implementation analyses offer important insights for understanding and improving the impact of evidence-based digital tools and should be routinely reported.