Substance Use
Applying intersectional approaches to studying the trajectories of adolescent alcohol and cannabis use: A combination of intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) and group-based trajectory modeling (GBTM) methods with application to youth in first time contact with the legal system Yu (Seashore) Li* Yu(Seashore) Li Johanna B. Folk Chase Anderson Michelle V. Porche Marina Tolou-Shams Brandon D.L. Marshall
Background: Longitudinal analyses applying intersectional theories to the study of developmental alcohol and cannabis use trajectories among youth are lacking.
Methods: We used intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) and group-based trajectory modeling (GBTM) to examine the relationship between eight intersectional strata defined by racial/ethnic & sexual minority identities and joint trajectories of self-reported, past 4-month alcohol use and cannabis use among justice-involved youth. Participants were recruited following their first ever contact with a family court system in a Northeastern state and followed prospectively for 24 months; assessments were completed every 4 months.
Results: Participants were 397 youth in first time contact with the legal system (30.2% sexual minority, 10.8% Black, 42.3% Latinx, 15.1% other/multi-racial, 31.7% White). Trajectory modeling identified four distinct trajectories of alcohol use: consistently low (43%), early initiators (12%), decreasing use (27%), and consistently high use (18%); see Figure 1A. Similarly, we found four trajectories of longitudinal cannabis use: consistently low (45%), early initiators (8%), decreasing use (20%), and consistently high cannabis use (27%); see Figure 1B. In MAIHDA models, we found a significant main effect for sexual minority status: sexual minority youth were more likely to be assigned to the consistently high alcohol and cannabis use trajectories compared to heterosexual youth. However, we did not observe substantial variation attributable to the intersectional strata defined by sexual minority status and race/ethnicity.
Conclusion: This research demonstrates the feasibility of combining MAIHDA and GBTM approaches to study the intersectional predictors of longitudinal substance use.
