To identify combinations of loneliness and social isolation and their associations with demographic, health, and psychosocial variables in a nationally representative sample of Brazilians aged 50 and older.
This cross-sectional study included 6139 participants from the baseline (2015-2016) of the Brazilian Longitudinal Study of Aging. Clusters were identified using an exploratory non-hierarchical analysis (k-means), and the odds ratios of the occurrence of variables were estimated by multinomial logistic regression.
36% were classified as lonely/not isolated, 19.4% as lonely/isolated, and 44.5% as not lonely/not isolated. The odds of belonging to the ‘lonely/not isolated’ cluster were lower for males and higher for those with more depressive symptoms, lower life satisfaction and quality of life scores, and poor mobility. The odds of being in the ‘lonely/isolated’ cluster were higher for individuals with more depressive symptoms, aged 80 years or older, and with lower levels of social participation, education, expectations of care, life satisfaction, and perceived quality of life.
Measures aimed at reducing social isolation and loneliness should be tailored to the characteristics of each cluster. The ‘lonely/isolated’ cluster would benefit from measures that reduce socioeconomic disadvantages throughout life and create opportunities for quality social interaction in old age.