Determinants of tobacco use by students

Estimate the prevalence and determinants of tobacco use by students. This cross-sectional study, carried out between 2013 and 2014, evaluates 701 public school students between 10 and 79 years of age living in a city in the countryside of the State of Goias, Midwest of Brazil. A structured questionnaire collected the data and the predictor variables were demographic data, family nucleus, religion, physical activity practice, family functionality and parental smoking. Two multivariable models were implemented, the first for occasional tobacco use and the second for regular use, acquiring the measure of prevalence ratio (PR) and their respective 95%CI. Variables with p < 0.10 were included in Poisson regression models with robust variance to obtain the adjusted PR (adPR) and 95%CI. The Wald Chi-Squared test examined the differences between proportions, and values with p < 0.05 were statistically significant. The prevalence of occasional and regular tobacco use were 33.4% (95%CI 29.8-36.9) and 6.7% (95%CI 5.0-8.8), respectively. The factors associated with occasional tobacco consumption were age of 15 to 17 years (adPR = 1.98) and above 18 years (adPR = 3.87), male gender (adPR = 1.23), moderate family dysfunction (adPR = 1.30), high family dysfunction (adPR = 1.97) and parental smoking (adPR = 1.60). In regards to regular consumption of tobacco, age above 18 years (adPR = 4.63), lack of religion (adPR = 2.08), high family dysfunction (adPR = 2.35) and parental smoking (adPR = 2.89) remained associated. Students exhibit a high prevalence of occasional and regular tobacco use. This consumption relates to sociodemographic variables and family dysfunction. Estimar a prevalência e determinantes do consumo de tabaco por estudantes. Neste estudo de corte transversal, realizado entre 2013 e 2014, avaliamos 701 estudantes entre 10 e 79 anos de escolas públicas de um município do interior do estado de Goiás, Centro-Oeste do Brasil. Os dados foram coletados por meio de um questionário estruturado, cujas variáveis preditoras foram: dados sociodemográficos, núcleo familiar, religião, prática de atividade física, funcionalidade familiar e pai ou mãe fumante. Foram implementados dois modelos multivariáveis, o primeiro para uso de tabaco na vida e o segundo para uso regular, obtendo-se como medida de efeito razão de prevalência (RP) e seus respectivos IC95%. Variáveis com p < 0,10 foram incluídas em modelos de regressão de Poisson com variância robusta, para obtenção da RP