Risk Profiles of Poor Diet Quality Among University Students: A Multivariate Segmentation Analysis

BACKGROUND: University students often display unhealthy eating habits shaped by economic, cultural, and psychosocial factors. These behaviors increase risks of chronic and mental disorders. The COVID-19 pandemic further changed their diet and physical activity (PA) habits, highlighting the need to identify determinants of diet quality (DQ). OBJECTIVE: The objective of this study is to identify risk profiles of poor DQ among university students from Chile, Mexico, Spain, and Italy through multivariate segmentation analysis. METHODS: ) and DQ (HEI). Analyses included descriptive, inferential, and decision tree (CHAID and CART) models. Chilean and Mexican students had the lowest DQ. Both models achieved high overall accuracy (≈91%), but balanced accuracy was around 50%, reflecting limited discrimination of healthy diet profiles and underscoring their exploratory value for identifying at-risk subgroups rather than precise prediction. CART identified country of residence and socioeconomic status as the primary determinants of poor diet quality (DQ), while CHAID highlighted field of study and socioeconomic status, with PA and BMI contributing at secondary levels. CONCLUSIONS: The results emphasize adapting public health strategies to local contexts-promoting Mediterranean-style diets in European universities and improving access to affordable healthy foods in Latin American campuses, complemented by campus initiatives integrating nutrition education, physical activity, and psychosocial support.