Objectives: This study aims to identify and analyze clusters formed by risk and protective factors for obesity among Brazilian adolescents.
Methods: Cross-sectional data collected in a national representative sample of adolescents by the Brazilian National School Health Survey of 2012 were used (n = 109,104). Cluster analysis was applied to identify clusters involving dietary intake, eating behavior, physical activity and sedentary behaviors. Logistic regression was employed to contextualize the clusters according to students' sociodemographics and schools' characteristics.
Results: Two clusters were identified. Cluster 1 was predominantly characterized by the low frequency of both risk and protective factors for obesity; Cluster 2 was characterized by high frequency of both of these factors. None was essentially healthy or unhealthy. Adolescents from less developed regions and with mothers with lower education level were associated to Cluster 1.
Conclusions: The identification of two mixed clusters indicate wide spread obesity risk among scholars in the country. The association between the clusters and sociodemographic characteristics of the population allows a greater refinement of health promotion and obesity prevention and combat actions in Brazil.
Keywords: Adolescent; Health behavior; Multivariate analysis; Obesity; Public health; Risk factors.