Mobile intervention for emerging adults with regular cannabis use: a micro-randomized trial

Cannabis use is prevalent among emerging adults, with regular use increasing risk for problems; yet few seek treatment. Using a micro-randomized trial (MRT), the study aims were to test a personalizing just-in-time adaptive interventions (pJITAI), which is powered by a reinforcement learning (RL) algorithm to improve the JITAI in real-time.

Emerging adults (N = 122) with regular cannabis use were enrolled in a 30-day MRT, and completed a 1-month post-test and 2-month follow-up (e.g., after MRT start; NCT05824754). During the MRT, participants were prompted to complete twice-daily surveys, followed by twice-daily micro-randomizations to intervention message or no message (probabilities determined by the RL algorithm). Primary outcomes were 1-month feasibility and acceptability; MRT results were examined for the effect of intervention message delivery (vs. not) on proximal intervention engagement and cannabis use (next decision point; intent-to-treat).

The app was feasible (89.1% (n = 122/137) downloaded; 77.1% (5642/7320) of twice-daily surveys completed, and acceptable (80.4% (n = 90/112) positively rated the messages for care, support, warmth, and respect). In MRT analyses, delivering a message (vs. not) increased proximal engagement (b = 0.044; 95% CI: 0.018, 0.070) but did not affect proximal cannabis use (b = 0.008; 95% CI: -0.003, 0.020). A clinically meaningful negative correlation was found between average engagement and cannabis use days at 2-months (b = -3.38; 95% CI: -8.68, 1.91).

The pJITAI is promising among not-in-treatment emerging adults, with messages increasing engagement, warranting further refinement to optimize this intervention.

Research was supported by the National Institute of Health [(P50DA054039; K23AA028232; T32AA007477; K23DA059484; P41EB028242] and the Brzoznowski family.