Abstract

Background: Adolescent alcohol and drug consumption are important public health problems in the Chilean young population. Aim: The purpose of this study was to examine the potential of a data mining approach in scaffolding policy making, using the particular case of differential risks of harmful alcohol consumption in adolescent students. Material and Methods: Index and control groups were composed by 7918 and 7138 participants respectively (drawn from a CONACE survey 2009), aged 16 +/- 2 years, 52% male. Heavy drinking at last month was the independent variable. As dependent variables parenting style, peer group influence, age and sex were used. For data analysis, a data mining approach was applied (CART, SPSS version 15). Results: The peer group influence was the main discriminant variable in males and the total sample, proving to be the only relevant variable in the case of women. The results suggest how a data mining approach may be useful in order to develop a hard data scaffolding for making and implementing policies in general and policies addressing adolescent alcohol consumption in particular.