Examining dynamic patterns of problematic cannabis use: Results from a multilevel network analysis

Young adults who engage in problematic cannabis use report lower work and interpersonal functioning yet are less likely to seek treatment, necessitating alternative methods for assessing and intervening on problematic cannabis use (e.g., mobile health applications to self-monitor drivers of cannabis use in daily life). However, previous work examining maintenance models of problematic cannabis use has primarily focused on modeling predictors of cannabis use as measured using static retrospective report rather than examining cannabis use as a series of interactions that unfold in everyday life. In this study, we analyzed ecological momentary assessment data (T = 3,230 observations) from 65 young adults who reported problematic cannabis use (Cannabis Use Disorders Identification Test-Revised: M = 10.38, SD = 4.35) and an interest in reducing their use. We used multilevel network analyses to model associations among biopsychosocial factors that aligned with social learning, self-medication, and experiential avoidance theories of substance use. Network models demonstrated consistent associations between socioenvironmental triggers and cannabis cravings, use, and intoxication that were nearly all clinically meaningful in size (bs > 0.10). Results indicated a statistically significant association between negative and positive affect with cannabis use and intoxication, respectively; however, these associations were not clinically meaningful in size. There were no clinically meaningful associations between coping strategies and cannabis use variables. Findings advance our understanding of cannabis use in everyday life, which is critical for refining more dynamic conceptualization of substance use and improving the precision of clinical assessments. (PsycInfo Database Record (c) 2025 APA, all rights reserved).