Exploring the Components of Impulsivity and Their Relationship to Adolescent Drug Use and Sexual Risk Behaviors

Introduction: Adolescence is a life stage when maladaptive risk behaviors first emerge. A well-recognized and robust predictor of adolescent risk-taking is the multifaceted construct of impulsivity (Whiteside & Lynam, 2001). Despite the fact that impulsivity can include different interrelated yet distinct dimensions like sensation seeking (SS), acting-without-thinking (AWT), and inability to delay gratification (IDG), most studies examining adolescent risk-taking have focused on a single index of impulsivity. Our goal was to focus on the distinct impulsivity dimensions (SS, AWT, IDG) and examine their similarities and differences, both in terms of relative stability/change over the course of adolescence and relationships with different risk behaviors (alcohol and other common drug use, early sexual initiation, unprotected sexual involvement).

Methods: We analyzed five waves of annual assessments from a community cohort of 387 adolescents (52% female; baseline age = 10-12 years) recruited from the Philadelphia area using robust linear regression analyses. Impulsivity dimensions of SS and AWT were assessed using self-report questionnaires, while IDG was assessed using a hypothetical monetary choice procedure (Green et al., 1994). Self-reports of adolescent risk behaviors were collected using computerized self-assisted interviewing techniques.  

Results: Bivariate associations revealed both stability and variability in these personality dimensions, suggesting these are not fixed traits. Some dimensions of impulsivity were more strongly associated (i.e., SS and AWT, r = 0.12-0.44, p < 0.05) during adolescence while others were not significantly linked (e.g., SS and IDG, r = 0.02-0.09, p = n.s.), suggesting that although inter-related these dimensions have unique characteristics. This was further validated by the fact that these dimensions were differentially related to risk behaviors: SS was not as strongly predictive of sexual risk-taking (β = 0.01 – 0.10, p = n.s.), as compared to AWT (β = 0.18-0.24, p <0.05) and IDG (β = 0.12-0.15, p <0.05). Alcohol and other drug use was more strongly related to AWT (β = 0.19-0.28, p <0.05) and SS (β = 0.20-0.28, p <0.05), as compared to IDG (β = 0.07-0.08, p <0.08).

Conclusions: Overall the findings highlight the importance of studying these dimensions individually as predictors of adolescent risk-taking, and understanding their sources of variability (developmental and individual) to better inform preventive interventions. In order to better understand these intra- and inter-individual variations in impulsivity, we are currently modeling heterogeneous latent trajectories of each dimension from early adolescence to early adulthood. These findings will be presented at the conference if selected.