UN MODELO PREDICTIVO DE LA PUNTUACIÓN DE CONCIENCIA EN ADOLESCENTES CHILENOS SOBRE LAS ETIQUETAS FRONTALES DE ALIMENTOS ULTRAPROCESADOS

The growing concern about obesity has led governments to implement warning label policies on ultra-processed foods in order to reduce their consumption. However, individual responses to these strategies have proven to be diverse. Therefore, it has become necessary to develop new methods to assess knowledge, understanding, and effective use of these labels. This study presents a model designed to predict the level of awareness of Chilean adolescents regarding front-of-package warning labels (FOPWL), through an illustrative case study. To this end, an innovative Robust Compromise Multicriteria Decision-Making (RC-MCDM) model was developed to calculate the degree of awareness of FOPWL. In addition, demographic, socioeconomic, health-related, dietary, environmental, and lifestyle variables were associated with the awareness score using Generalized Additive Models for Location, Scale, and Shape (GAMLSS). These models allow the analysis of variables with different statistical distributions. The results show a positive correlation between the FOPWL awareness score and age, female gender, the presence of diet-related diseases, adherence to a vegetarian diet, knowledge of healthy eating, and availability of pocket money. Conversely, negative associations were observed for variables such as body weight perception, engagement in weight-loss dieting, and screen time, while the relationship with the province of residence was found to be ambiguous. This predictive model may be useful for public health policymakers seeking to modify eating behaviors in specific population groups.