Session: Diverse Applications of Time-Varying Effect Modeling to Answer Important Questions in Prevention Science (Society for Prevention Research 24th Annual Meeting)

SESSION INTRODUCTION. Time-varying effect modeling (TVEM) is a novel statistical technique that offers exceptional potential toward improving our understanding of time- and age-varying effects in prevention science. This session consists of six papers that apply TVEM to elucidate the emergence and duration of risk factor associations with substance use and related problem behavior across adolescence and young adulthood, in order to provide new information that may assist in the timing and targeting of preventive interventions.

Paper 1 presents a conceptual introduction to TVEM and provides findings relevant to marijuana-use prevention from TVEM analyses across four distinct time metrics; including historical time (across cohorts), developmental time (over age), age-of-onset effects, and time relevant to an event (since birth of child). 

Paper 2 examines nationally representative survey data and finds that males and Hispanic youth are at increased risk for e-cigarette use (e-cig use) beginning during early-to-mid adolescence, and that e-cig use is more strongly related to traditional cigarette smoking before versus after age 15.

Paper 3 further unpacks the association between e-cig use and smoking in a longitudinal study of youth, showing that the directionality of this association depends on adolescents’ baseline level of nicotine dependence (ND). E-cig use was positively associated with cigarette smoking for youth with moderate-to-high levels of baseline ND, but was negatively associated with smoking at low levels of ND.

Paper 4 examines longitudinal data from Add Health, and finds that sexual behavior in adolescence is associated with depression (particularly for females), whereas sexual behavior in young adulthood is associated with depression for individuals who attend religious services frequently and who have parents strongly disapproving of adolescent sexual activity.

Paper 5 reports on the age-varying protective association between family management and adolescent marijuana use, and finds the strongest association between ages 14-16. This age-varying association was stronger for girls, youth with lower behavioral disinhibition, and in families where parents did not report marijuana use, with each of these differences emerging at different ages.

Paper 6 examines data from the Monitoring the Future panel study and shows that the association between perceived friends’ marijuana use and adolescents’ own marijuana use increased from ages 18-30, with the association being strongest for males, white youth, and youth from parents without a college education. 

Brief discussion focused on the implications of findings for research and prevention will follow each of the paper presentations.