The efficacy of three interventions to prevent substance use and sex initiation in subgroups of Hispanic adolescents

The objectives of this dissertation were to identify subgroups of Hispanic adolescents based on their intrapersonal (e.g., positive attitudes towards substance use and risky sexual behavior) and ecodevelopmental (e.g., poor parent-adolescent communication) risks, and to determine whether, and if so to what extent, the efficacy of three interventions designed to prevent substance use and sex initiation varied as a function of the Hispanic adolescent subgroups identified. A total of 266 8th grade, low-income, inner city Hispanic adolescents living in Miami-Dade County, Florida, were randomized to one of three interventions. Adolescents were administered questionnaires on their intrapersonal and ecodevelopmental risk and protective factors as well as on their substance use and sexual behavior at baseline, 6, 12, and 24-months. K-means cluster analysis identified four adolescent subgroups: (1) moderate ecodevelopmental, but high intrapersonal risk, (2) high ecodevelopmental but low intrapersonal risk, (3) low ecodevelopmental but high intrapersonal risk, and (4) low ecodevelopmental but low intrapersonal risk. Chi-square analysis showed that substance use at baseline differed by adolescent subgroup, with those adolescents reporting high ecodevelopmental, but low intrapersonal risk (Subgroup 2) reporting the highest levels of substance use. Survival analysis showed that adolescent subgroup was an important risk factor for substance use and sex initiation. More specifically, adolescents with moderate ecodevelopmental but high intrapersonal risk (Subgroup 1) and adolescents with high ecodevelopmental but low intrapersonal risk (Subgroup 2) were at increased risk of initiating substance use compared to adolescents with low-intrapersonal and low ecodevelopmental risk (Subgroup 4). Adolescents with high ecodevelopmental but low intrapersonal risk (Subgroup 2) were also at increased risk of initiating sex, relative to adolescents with low ecodevelopmental and low intrapersonal risk (Subgroup 4). The results also suggest that adolescents randomized to Familias Unidas + PATH were at less risk of initiating substance use than adolescents randomized to either one of two other interventions. Post-hoc analyses also suggest that for those adolescents with high ecodevelopmental risk (Subgroup 2), Familias Unidas + PATH was most efficacious in preventing substance use. The implications of the findings for epidemiological and prevention research are discussed