Social isolation in the oldest-old: determinants and the differential role of family and friends

Purpose To examine the association of sociodemographic and health-related determinants with social isolation in relation to family and friends in the oldest-old. Methods Database was the multi-center prospective AgeCoDe/AgeQualiDe cohort study assessed at follow-up wave 5 ( N = 1148; mean age 86.6 years (SD 3.0); 67% female). Social isolation was assessed using the short form of the Lubben Social Network Scale (LSNS-6). The LSNS-6 contains two sets of items establishing psychometrically separable subscales for isolation from family and friends (ranges 0–15 points), with lower scores indicating higher isolation. Cross-sectional linear (OLS) regression analyses were used to examine multivariate associations of sociodemographic and health-related determinants with social isolation from family and friends. Results Overall, n = 395 participants (34.6%) were considered socially isolated. On average, isolation was higher from friends (mean 6.0, SD 3.8) than from family (mean 8.0, SD 3.5). Regression results revealed that in relation to family, males were more socially isolated than females (β = − 0.68, 95% CI − 1.08, − 0.28). Concerning friends, increased age led to more isolation (β = − 0.12, 95% CI − 0.19, − 0.05) and functional activities of daily living to less isolation (β = 0.36, 95% CI 0.09, 0.64). Independent of the social context, depression severity was associated with more social isolation, whereas cognitive functioning was associated with less social isolation. Conclusions Different determinants unequally affect social isolation in relation to family and friends. The context of the social network should be incorporated more strongly regarding the detection and prevention of social isolation to sustain mental and physical health.