As adolescent substance use declines, internalizing symptoms increase: identifying high-risk substance using groups and the role of social media, parental supervision, and unsupervised time

ABSTRACT Multiple data sources indicate that adolescent psychopathology, particularly internalizing symptoms, is at historically unprecedented highs in the United States. These include cognitive (low self-esteem, self- derogation), affective (depressive affect), and social (loneliness) dimensions of adolescent internalizing symptoms, which have been rapidly increasing since ~ 2009. Coinciding with these trends has been declines in adolescent alcohol and other drug (AOD) use (except marijuana), contrary to what would be expected given the historically strong relationship between AOD use and internalizing symptoms. Declines are also apparent in high intensity alcohol use (e.g. 10+ drinks per drinking occasion), high-frequency AOD use (although not marijuana), and simultaneous use of AOD (including marijuana). The strength of the relationship between internalizing symptoms and AOD use among adolescents is also decreasing; for the first time approaching null in 2016, which has serious implications for risk factor assessment, prevention and intervention. Divergences may vary across demographic subgroups, however. Little work has estimated why these diverging trends are occurring; the most prominent hypothesis is smartphones and social media. These technologies are hypothesized to underlie less face-to-face time among adolescents, increase real-time parental monitoring, and negative feelings such as envy and low self-worth. Such a shift may underlie decreases in AOD use (less unsupervised time with other adolescents, more parental monitoring) and increases in internalizing symptoms (negative feelings and self-worth). However, existing literature is not based on nationally-representative data, and the relationship with AOD use has not been investigated. The present study will utilize the Monitoring the Future (MTF) cross-sectional surveys of ~45,000 adolescents per year. MTF includes a breadth measures related to internalizing symptoms, substance use, adolescent interaction, as well as a diverse array of potential confounders. We will address three aims: (1) Examine time trends in the relationship between internalizing symptoms (low self-esteem, self-derogation, depressive affect, loneliness) with any, high-intensity, high frequency AOD use, testing the magnitude of the relationship across time and by race, sex, and SES; (2) Examine time trends in the relationship between internalizing symptoms and simultaneous use of AOD (e.g., alcohol and marijuana use, alcohol and opioid use, others), testing the magnitude of the relationship across time and by subgroups; (3) Test the extent to which social media use, parental supervision, and unsupervised time with friends are associated with internalizing symptoms, AOD use, simultaneous use of AOD, and whether the magnitude of these associations explains changes in trends over time. Methods to address these aims will include time-varying effect modeling as well as parallel process growth models. This study will provide evidence for public health action by delineating the relationship between AOD and internalizing symptoms over time, and by examining the role of new technology and the changing landscape of adolescent interaction.