Abstract: Demographic, Environmental and Attitudinal Predictors of Substance Use Heterogeneity Among South African Adolescents: A Latent Class Analysis (Society for Prevention Research 23rd Annual Meeting)

177 Demographic, Environmental and Attitudinal Predictors of Substance Use Heterogeneity Among South African Adolescents: A Latent Class Analysis

Schedule:
Wednesday, May 27, 2015
Columbia A/B (Hyatt Regency Washington)
* noted as presenting author
Megan S. Schuler, PhD, Postdoctoral Fellow, The Pennsylvania State University, State College, PA
Jessica Salas-Brooks, BS, Graduate Student, The Pennsylvania State University, State College, PA
Linda Lee Caldwell, PhD, Professor, The Pennsylvania State University, University Park, PA
Edward Allan Smith, PhD, Director, The Pennsylvania State University, University Park, PA
Introduction: Adolescent substance use, particularly polydrug use, is associated with a range of negative outcomes, including school dropout, poor physical and mental health outcomes, and sexual victimization. Geography and culture influences substance use trends and availability, yet there is limited research focusing on patterns of adolescent use in developing countries. Specifically, this study focuses on South Africa, where use of inhalants and methamphetamine is more common that in the US.

Methods: Data were collected as part of the HealthWise 2 school-based prevention program, which is administered in 56 public schools in Cape Town, South Africa. Our sample included 4,428 adolescents enrolled in 9th grade (mean age=16.8, 56% female). We investigated heterogeneity of adolescent substance use patterns using latent class analysis based on 5 substance use indicators: lifetime use of alcohol, cigarette, marijuana, methamphetamine, and inhalants. Additionally, we used latent class regression to determine demographic, social/environmental, and attitudinal predictors of class membership. All analyses were conducted in Mplus.

Results: Based on LCA fit statistics and class interpretability, we determined that a 3-class model best fit the data. The classes were as follows: Low Use class, characterized by moderate alcohol use and very low use of all other drugs (53% of youth); Alcohol, Cigarette and Marijuana (ACM) class, characterized by high rates of alcohol, cigarettes, and marijuana (39% of youth), and Polydrug Use class, characterized by high rates of alcohol, cigarettes and marijuana and moderate use of inhalants and methamphetamine (8% of youth). Youth who did not live with their father full time, who were in a relationship, who had ever had sex, or who had experienced a forced sex attempt were more likely to be either in the Polydrug Use or the ACM class, relative to the Low Use class. Youth who reported feeling confident they could avoid a risky situation and who had someone to talk to when making important decisions were less likely to be in the Polydrug Use class, relative to the Low Use class. Males, youth who reported being likely to drop out of school within a year, and youth who reported 5+ sex partners were more likely to be in the Polydrug Use class, relative to the Low Use class.

Conclusions: Encouragingly, the majority of 9th grade students in our sample belonged to the Low Use class. Yet, we also observed the high-risk Polydrug Use class, who have moderate to high usage of all 5 substances. Understanding the patterns of substance use among a representative sample of secondary school students is central to prevention programs seeking to reduce adolescent substance use, such as HealthWise 2.