Onset of Alcohol, Cigarette and Marijuana Consumption Among Adolescents in Santiago Chile: Results of Categorical Time Event History Analysis Models

Background and Purpose One of the major predictors of substance abuse is the early onset of substance use. Although prior research has examined in detail factors related to use of cigarettes, alcohol, and marijuana, less research exists exploring their simultaneous consumption. Furthermore, despite the seminal work by Kandel on the ‘gateway' hypothesis and the subsequent research on the topic, studies on the joint onset of these substances among U.S. Hispanic and Latin American populations remains almost non-existent. In this study we expand on prior work in two important ways. First we examined the onset of cigarettes, alcohol, and marijuana use with an international sample of urban adolescents from Santiago, Chile. We then identified onset risk profiles for each of these substances by examining individual, peer, parenting, and familial factors. Methods This study used cross-sectional data from the first wave of the Santiago Longitudinal Study, a NIDA-funded study of community-dwelling adolescents in Chile. The sample consisted of 998 participants of mid-to-low socioeconomic status (mean age=14, 49% females). At the study site, in private rooms, youth completed a 2-hr interviewer administered questionnaire with comprehensive questions on substance use, individual, peer, parenting, and family variables. To incorporate information about the timing of events, data were analyzed using several Discrete-Time Event History Analysis models with STATA 10.0. Results Unadjusted results showed that 44%, 36%, and 14% of adolescents had consumed alcohol, cigarettes, and marijuana, respectively. The average age of alcohol and cigarettes initiation was 12 while that for marijuana was14. All adolescents who consumed marijuana had already consumed alcohol and cigarettes. The estimated hazard functions of substance initiation were not the same for all substances, with marijuana having the most different hazard function. Initiation of alcohol and cigarette use started around age nine and showed a clear increment around age 12. Marijuana onset was estimated to occur around age 13 and showed a clear increment between ages 14 and 15. From the adjusted model, those with individual, peer, parenting, and familial factors, three profiles of initiation were identified: low, medium, and high risk. For alcohol and cigarettes the estimated hazard rates between the high and medium profile were similar but in the case of marijuana there was a clear distinction between the three profiles. Conclusions and Implications Consistent with the gateway hypothesis, we found that all Chilean adolescents who had consumed marijuana had also consumed alcohol and cigarettes. However, unlike studies in the U.S., the onset of cigarettes and alcohol occurred essentially at the same time. With the extensive psychosocial data available, we identified three profiles of initiation; profiles that differed depending on the substance studied. Interestingly, no interactions were observed by sex. We discuss these findings in (1) relation to cross-cultural differences in consumption patterns and (2) the design of prevention interventions with Chilean youth. Finally, in our presentation we will describe in detail the use of Categorical Time Event History Analysis Models, an advanced statistical method that is generally underutilized by social work researchers.