Loneliness is not merely an emotional discomfort — it is a measurable state that consumes a predictable portion of later life and correlates with mortality risk. Quantifying exactly how many years older adults actually live in loneliness, and which structural and individual factors accelerate or buffer that transition, reframes loneliness as a public health metric comparable to years lived with disability.
Drawing on five harmonized multinational cohort datasets comprising 61,511 adults aged 60 and older across 24 countries, researchers applied multistate Markov modeling to estimate both total life expectancy (TLE) and lonely life expectancy (LLE) — the expected duration spent in a lonely state. At age 60, the average TLE was 23.14 years, of which 4.31 years, or roughly 19%, were spent lonely. The transition from non-lonely to lonely was significantly more likely among women, those with low educational attainment, individuals with multimorbidity, and those with psychological disorders. Physical activity emerged as the primary individual-level factor associated with recovery from loneliness back to a non-lonely state. At the country level, higher GDP correlated with 1.25 additional years of total life and 2.40 fewer lonely years. Civil society participation was associated with a 2.78-year reduction in LLE, and public transport accessibility with a 1.97-year reduction — effects of considerable magnitude. Urbanization, counterintuitively, was linked to modestly elevated LLE despite slightly longer TLE.
These findings carry several important implications. The LLE framework itself is analytically novel — translating an abstract psychosocial variable into life-table years makes comparisons across interventions far more tractable. The magnitude of the transport accessibility effect is striking and likely reflects mobility-linked social participation rather than transport per se, suggesting that physical access to community functions as a loneliness buffer. The civil society effect aligns with a growing literature on civic engagement and healthy aging, though the cross-national design introduces unmeasured confounding between GDP, institutional quality, and social infrastructure. Because this is observational and relies on self-reported loneliness, causal inference is limited. Nonetheless, the scale — over 61,000 participants across diverse economies — lends this study unusual generalizability. The finding that nearly one in five post-60 years is spent lonely should be treated as a baseline against which population-level interventions can be benchmarked.
