Prediction of smoking initiation in junior high school students

Abstract

Introduction: predictive instruments contribute to the stratification of the risk of the appearance of diseases and harmful habits, which is why they are very useful as a complement to the clinical-epidemiological method.

Objective: to estimate the capacity of some factors as predictors of smoking initiation in students of the “Paco Cabrera” Junior High School of Puerto Padre.

Methods: cross-sectional analytical study in 257 basic secondary school students. Multivariate binary logistic regression was used to identify the set of independent variables with predictive capacity. The ethical principles of scientific research were complied with.

Results: the combination of six explanatory variables was identified. The predictor that exhibited the highest adjusted odds ratio value was: “smoking is a good option to spend free time” (28.96). The Hosmer-Lemeshow X2 test showed that the model fits the data (X2=5.615; p=0.585). After executing the model, the value of y=0.9968 was obtained as maximum probability and the value of Y=0.0011 was obtained as minimum probability. In the analysis of the discriminative capacity of the model, a ROC curve and its area under the curve (AUC=0.917; asymptotic significance=0.000) were used, which was considered acceptable.

Conclusions: the probability of starting smoking in the lower secondary students in this study can be predicted by the synergistic combination of the predictors: close friends who smoke, smoking eliminates stress and anxiety, it is good for losing weight, it is a good option to spend time, age over 12 years and living with smoking family members.