Childhood conduct problems, potential snares in adolescence and problematic substance use in Brazil

ObjectiveExamine whether the relationship between childhood conduct problems and substance use in adulthood is explained through potential snares (police arrest, gang membership, school non-completion) in a Brazilian population-based birth cohort. MethodData were analyzed from 4,599 young people from the 1993 Pelotas Birth Cohort in Brazil. The exposure was conduct problems at age 11 years. Outcomes included hazardous alcohol consumption and illicit drug use (age 22 years). Mediators included police arrest (by age 18 years), gang membership (ages 18 and 22 years), and school non-completion (by age 22 years). We performed counterfactual mediation using the parametric g-computation formula to estimate the indirect effect via all three mediators simultaneously. ResultsAfter adjusting for confounders (including comorbidity), conduct problems were weakly associated with police arrest [OR (95% CI)=1.45 (0.97, 2.16)] and school non-completion [OR (95% CI)=1.46 (1.22, 1.74)], but not with gang membership. Police arrest and gang membership were associated with illicit drug use [OR (95% CI)=3.84 (2.46, 5.99); OR (95% CI)=7.78 (4.30, 14.10), respectively] and with hazardous alcohol use [OR (95% CI)=1.60 (1.08, 2.38); OR (95% CI)=1.88 (1.07, 3.30)]. However, there was no association between school non-completion and either outcome after confounder adjustment. There was little evidence for an indirect effect of conduct problems on hazardous alcohol use and illicit drug use via all three mediators after confounder adjustment. ConclusionThe indirect effects of childhood conduct problems on substance use in early adulthood via police arrest, gang membership, and school non-completion were very small, after accounting for confounders and comorbidity.