Back

Early adolescent substance use patterns and associated factors: A longitudinal analysis of the Adolescent Brain Cognitive Development Study

Otero-Leon, D. F.; Harwood, M. I.; Jalali, M. S.; Dong, H.

2025-06-24 addiction medicine
10.1101/2025.06.24.25330203 medRxiv
Show abstract

BackgroundEarly substance use is associated with long-term negative health and behavioral outcomes. While extensive research has examined predictors of substance use in adolescents and adults, relatively little is known about the risk factors that drive substance use in younger children. Furthermore, most studies focus on individual substances in isolation, rather than exploring the use of multiple substances. This study aims to identify demographic, family, and peer-related factors associated with alcohol, nicotine, cannabis, and multiple substance use in children aged from 9 to 13. MethodsWe analyzed longitudinal data from 11,868 children enrolled in the Adolescent Brain and Cognitive Development (ABCD) Study (version 5.1), incorporating observations collected from baseline, annual follow-ups, and mid-year phone interviews. Substance use outcomes were categorized into five mutually exclusive groups: no use, alcohol only, nicotine only, cannabis only, and use of two or more substances. We applied a generalized estimating equation model to assess the associations between substance use outcomes and demographic, family, and peer characteristics. ResultsPeer substance use was consistently associated with all categories of child substance use, particularly for the use of two or more substances. Parental drug use, permissiveness toward alcohol, and lower parental education were also linked to increased risk. Older age was associated with alcohol, nicotine, and multiple substance use. ConclusionsFindings highlight the importance of peer and family environments in shaping early substance use behaviors. Early prevention efforts should consider these factors to effectively target at-risk children.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.