Abstract
OBJECTIVES/GOALS: Adolescence represents a critical period for substance use initiation. Various factors may contribute to trying a sip or single puff of a substance, that could lead to more frequent use. However, less is known about how predictors from multiple domains converge to impact risk for general substance use initiation. METHODS/STUDY POPULATION: The Adolescent Brain Cognitive Development (ABCD) study is a multi-site longitudinal study following youth into early adulthood. The present study included 7,644 ABCD children who reported no lifetime substance use (including any experimentation) at baseline (ages 9–10). Our primary aim was to use a random forest classification model to predict binary substance use initiation, defined as trying any non-prescribed substance (e.g., alcohol, tobacco, cannabis, non-prescribed medications), during a 2-year follow-up after baseline. A total of 402 variables from the following categories were examined as predictors: demographics, peer substance use and availability, mental and physical health, culture and environment, biospecimens, neurocognitive functioning, and structural neuroimaging variables. RESULTS/ANTICIPATED RESULTS: Over a two-year follow-up, 751 (9.8%) of substance-naïve children reported trying a substance by age 11. The most common substance was alcohol, followed by cannabis and tobacco. Mean Decrease Accuracy (MDA) values were used to assess the relative importance of each predictor. The overall accuracy of the model in accurately predicting group membership (no substance use initiation vs. substance use initiation) was 57.66%. Of the top 5 predictors, the most important predictor was intent to use alcohol (MDA = .002). The following top predictors were structural neuroimaging variables: volume and surface area of right lateral occipital lobe (MDA = .0009 and .0008, respectively), surface area of right inferior temporal lobe (MDA = .0007), and surface area of left superior frontal lobe (MDA = .0007). DISCUSSION/SIGNIFICANCE: A combination of intent to use alcohol and structural neuroimaging indices were among the top predictors of substance use initiation. Understanding predictors of early substance use experimentation is important for identifying at-risk youth that may require targeted intervention approaches.