One of the major goals of drug use prevention programs is to delay the age of onset of substance use. What is called early initiation, usually occurring in adolescents under the age of 15, is a salient predictor of Substance Use Disorders later in adulthood. The causes of early initiation are complex and multifaceted and this has led to the identification of a rich set of risk and protective factors that influence age of onset. Nonetheless, there is little knowledge about the interdependence of these factors in their impact on early initiation.

This paper addresses this question by applying Multiple Correspondence Analysis to data on family, community and social risk and protective factors from over 1200 adolescents. We find that community and to a lesser extent social factors are the most clearly associated to early initiation and we compare our results to those obtained from linear regression analyses of the same data that do not incorporate interdependence and find opposite results. We discuss the differences between linear regressions and MCA to evaluate the interplay of risk and protective factors and the implications of our findings for health policy and the design of prevention interventions aimed at delaying age of onset.