RESUMEN
BACKGROUND: Loneliness is considered a public health problem, particularly among older adults. Although risk factors for loneliness have been studied extensively, fewer studies have focused on the protected and risk groups that these factors configure. Our objective is to analyze the variables and latent factors that predict loneliness in older adults and that enable risk and protected groups to be configured. METHODS: We employed an epidemiological, cross-sectional survey that was carried out on a random sample of 2060 people over 65 years extracted from the census. A structured telephone interview was used to assess mental and physical health, habits, quality of life, and loneliness, applying the COOP-Wonca, Goldberg General Health (GHQ-12), and Barber Questionnaires. RESULTS: Predictors of loneliness were: mental health, living alone, quality of life, depressive symptoms, low educational level, and some deficiency situations such as having no one to turn to for help. The factors extracted (Factorial Analysis) were: a subjective experience of poor health, objective isolation, and psychological isolation. We established at risk and protected groups ("Decision Tree" procedure), and loneliness was referred to by 73.2% of the people living alone and with poor mental health and quality of life (risk group). By contrast, only 0.8% of people living with others, with good mental health and good quality of life felt loneliness (protected group). CONCLUSION: In a well-developed city, subjective and objective factors are associated with loneliness. These factors, especially those associated with at risk or protected groups, must be considered to develop strategies that address loneliness.