Previous studies on the impact of cannabis use disorders (CU) on outcome in psychosis were mostly based on non-representative samples, have often not controlled for confounders and rarely focused on adolescents. Thus, the aims of the present study were to assess;

• (i) prevalence of CU;

• (ii) pre-treatment and baseline differences between CU and those without CU (NCU); and

• (iii) the impact of baseline and course of CU on 18-month outcomes in a representative cohort of adolescents with early onset first episode psychosis (EOP).

The sample comprised 99 adolescents (age 14 to 18) with EOP (onset age 14 to 17), admitted to the Early Psychosis Prevention and Intervention Centre in Australia. Data were collected from medical files using a standardized questionnaire.

Prevalence of lifetime CU was 65.7%, baseline CU 53.5%, and persistent CU 26.3%. Baseline CU compared to NCU was associated with higher illness-severity, lower functioning, less insight, lower premorbid functioning and longer duration of untreated psychosis. Compared to both NCU and those who decreased or stopped CU during treatment, only persistent CU was linked to worse outcomes and more service disengagement. Effect sizes were medium controlling for relevant confounders. Medication non-adherence did not explain the link between persistent CU and worse outcome.

The prevalence of CU in adolescents with EOP is high, while only persistent CU use was associated with worse outcome with medium effect sizes. Specific needs of adolescent patients with respect to cannabis interventions within integrated care settings should be addressed in future studies.