Temporal dynamics of the relationship between change in depressive symptoms and cannabis use in adolescents receiving psychosocial treatment for cannabis use disorder.

Aims: Cannabis use disorder (CUD) and depression frequently co-occur in youth. How depressive symptoms change over the course of CUD treatment and how they impact substance use treatment outcomes is unknown. In the current study, we examine the temporal relationships between cannabis use and depression in adolescents receiving evidence-based treatments for CUD as part of a multisite clinical trial.

Design: Six hundred adolescents (age 12-18) with a CUD were randomly assigned to substance use treatment from one of five evidence-based psychosocial interventions. We assessed self-reported cannabis use frequency and depressive symptoms at baseline (BL) and again at 3-, 6-, 9, and 12-months. A bivariate latent change model assessed bidirectional effects of baseline levels and time-lagged changes in depressive symptoms and cannabis use on depression and cannabis use outcomes.

Findings: Depressive symptoms (72%) and major depressive disorder (MDD) (18%) were common at BL. Both depression and cannabis use decreased over time and change in cannabis use was significantly associated with change in depressive symptoms (b = 1.22, p = .003). Time-lag analyses showed that within-subject change in depression (from one time point to the next) was predicted by previous depression (b = -0.71, p < .001) but not cannabis use (p = .068), and change (decrease) in cannabis use was predicted by previous (greater) depressive symptoms (b = -1.47, p < .001) but not cannabis use (p = .158), respectively.

Conclusion: These findings indicate an enduring relationship between decreasing cannabis use and decreasing depression among adolescents lasting for 9-months after receiving psychosocial interventions for CUD. The presence of depressive symptoms did not appear to interfere with substance use treatment or attenuate improvements in cannabis use frequency. A decrease in cannabis use was not contingent upon a reduction in depressive symptoms. These findings are limited by the possibility of regression to the mean for both cannabis use and depressive symptoms, and the lack of a nonintervention control group.