Effects of social isolation on all-cause mortality in older adults: do age, gender, and measurement matter?

BACKGROUND AND OBJECTIVES: Research has suggested that social isolation increases mortality risk. However, findings have been mixed regarding gender differences, and few studies have explored age differences among older adults. We examined whether the association of social isolation and all-cause mortality varied by gender and age and compared two established measures of social isolation. RESEARCH DESIGN AND METHODS: Participants (N = 7,026, age 65+) in the National Health and Aging Trends Study (NHATS) were followed up to 12 years. 2 NHATS-based social isolation measures (Pohl and Cudjoe), both with 3 categories (socially integrated, somewhat isolated, severely isolated), were used. Cox proportional hazards models, adjusted for sociodemographic, health, and health behavior factors, were applied. Interaction terms tested the moderation by gender and age. RESULTS: During follow-up, 2,457 (35.0%) participants died. The 2 social isolation measures had substantial misalignment, with only 36.6% of the sample being classified in the same category. Using Pohl's measure, both somewhat and severely isolated groups had significantly higher mortality risks than socially integrated individuals, and effects were stronger for men than for women. No age interaction was observed. Associations using Cudjoe's measure were non-significant after adjustment. DISCUSSION AND IMPLICATIONS: Social isolation hastens death in older adults, particularly among men, but findings depend on how isolation is measured. Measures emphasizing family/friend contact may better capture mortality-related risk. Healthcare providers should consider screening for, and using social prescribing to address, social isolation when treating older patients. Reducing social isolation in older men requires particular attention.