Effectiveness research on depression prevention usually compares pre- to post-intervention outcomes across groups, but this aggregation across individuals may mask heterogeneity in symptom change trajectories. Hence, this study aimed to identify subgroups of adolescents with unique trajectories of change in a school-based depression prevention trial. It was also examined how trajectory membership was associated with the intervention conditions, depressive symptoms at 12-month follow-up, and baseline predictors. Hundred-ninety adolescent girls (M_age = 13.34; range = 11–16 years) with subclinical depression at screening (M = 57 days before pre-test) were allocated to four conditions: a face-to-face, group-based program (OVK), a computerized, individual program (SPARX), OVK and SPARX combined, and a monitoring control condition. Growth Mixture Modeling was used to identify the distinct trajectories during the intervention period using weekly depressive symptom assessments from pre-test to post-test. Analyses revealed three trajectories of change in the full sample: Moderate-Declining (62.1% of the sample), High-Persistent (31.1%), and Deteriorating-Declining (6.8%) trajectories. Trajectories were unrelated to the intervention conditions and the High-Persistent trajectory had worse outcomes at follow-up. Several baseline factors (depression severity, age, acceptance, rumination, catastrophizing, and self-efficacy) enabled discrimination between trajectories. It is concluded that information about likely trajectory membership may enable (school) clinicians to predict an individual’s intervention response and timely adjust and tailor intervention strategies as needed.

抑郁症预防的有效性研究通常比较各组干预前后的结果，但这种跨个体的聚合可能掩盖症状变化轨迹的异质性。因此，本研究旨在确定在以学校为基础的抑郁症预防试验中具有独特变化轨迹的青少年亚组。还检查了轨迹成员与干预条件、12 个月随访时的抑郁症状和基线预测因素之间的关系。190 名青春期女孩（M_年龄 = 13.34；范围 = 11-16 岁）在筛查时患有亚临床抑郁症（M = 预测试前 57 天）被分配到四个条件：面对面、基于小组的程序 (OVK)、计算机化、个人程序 (SPARX)、OVK 和 SPARX 组合以及监控条件。生长混合模型用于使用从测试前到测试后的每周抑郁症状评估来识别干预期间的不同轨迹。分析揭示了整个样本的三个变化轨迹：中度下降（占样本的 62.1%）、高度持续（31.1%）和恶化 – 下降（6.8%）轨迹。轨迹与干预条件无关，高持续轨迹在随访中的结果更差。几个基线因素（抑郁严重程度、年龄、接受度、反刍、灾难性和自我效能感）能够区分轨迹。