Loneliness

Treat relationships as a health habit: being woven into other people's lives in several ways (a partner, friends and family, groups) predicted survival better than how many people you know. If loneliness persists, a psychological approach appeared to ease it more than simply adding social contact, though the evidence is low certainty.

Strong relationships go with longer survival

The reference study is a 2010 pooling of 148 studies: adults with strong social relationships had about 50% greater odds of survival than those with weak ones, an effect the authors judged comparable to quitting smoking and larger than obesity or physical inactivity ^[1].

Newer data covering far more people look separately at isolation and loneliness rather than at overall relationship strength, so their figures measure something different. Across 90 cohorts and 2.2 million people, social isolation went with about 32% higher mortality and loneliness with about 14% higher ^[2].

Depth counts more than headcount

What predicted survival best in the 2010 pooling was how deeply someone is woven into a network, not how many people are around. In the table, a figure of 1.50 means 50% greater odds of survival; higher is better ^[3].

The best-known of those combined measures, the Social Network Index, comes from a study of adults in Alameda County, California. It combines marriage, contact with friends and family, and church or group membership, and the most isolated adults were 2.3 (men) to 2.8 (women) times as likely to die over nine years as the most connected ^[4].

Living alone is not harmless, though. The 2010 result for it was imprecise rather than null, which the authors put partly down to crude measurement, since someone can live alone and still have a large, supportive network. In later pooled data adjusted for confounders, living alone carried about 32% higher odds of death, similar to loneliness and social isolation in the same pooling (26% and 29%) ^[5], and in a study of about 490,000 UK adults it was the isolation item most strongly tied to earlier death ^[6].

Why to doubt that isolation itself is the cause

All of these studies are observational, and lonely or isolated people differ from others in many ways. Genetic analyses (Mendelian randomization, which uses inherited differences to get around confounding) in the same UK adults found little evidence that isolation itself causes the diseases it tracks with, and blood-protein signatures, health habits and socioeconomic position each explained a sizeable share of the links ^[7].

What it may do in the body

Feeling isolated seems to register as a threat. Chronically lonely adults show raised inflammation and stress-system activity, closely mirroring the pattern of long-term stress. In blood samples from about 42,000 UK adults, proteins linked to loneliness and isolation clustered in inflammation and immune pathways, and several of them partly explained the extra heart disease, stroke and deaths seen in lonely people over 14 years ^[8].

Positive social contact engages the brain's reward circuits; in animals, social isolation switches on the cortisol stress system, and in people, support from others can ease acute stress, and recalling good memories shared with close people can blunt the cortisol response to an acute stressor ^[9].

Loneliness and dementia

Loneliness goes with about 31% higher dementia risk across 21 cohorts of more than 600,000 people, even after accounting for depression and social isolation ^[10]. Like the mortality data, this is observational. See Dementia prevention for the wider risk picture.

The "15 cigarettes a day" claim

The U.S. Surgeon General's 2023 advisory called loneliness and isolation an "epidemic" and an urgent public-health issue, put their added risk of early death at about 26% and 29%, and said that lacking social connection raises that risk about as much as smoking up to 15 cigarettes a day ^[11].

Read the cigarette figure as a loose visual comparison, not a measurement. It comes from setting the social-ties effect size beside a separate smoking estimate in one chart ^[12]. The same goes for the World Health Organization's 2025 estimate that about 1 in 6 people worldwide are affected by loneliness, linked to more than 871,000 deaths a year: a modelled figure, not a count ^[13].

Loneliness can be reduced; the health payoff is untested

Across 122 randomised trials, programs cut loneliness by a small-to-moderate amount, with benefits lasting 1 to 6 months, and psychological approaches appeared to work better than simply adding social contact. The evidence is low certainty, and the trials measured loneliness, not survival, so it is not yet shown that easing loneliness lengthens life ^[14].

What to do

1. Keep regular contact with the people who matter most, weekly rather than an annual catch-up. Combined measures of how woven in you are predicted survival better than network size alone.
2. Belong to something. Group membership is part of the combined measures that predict survival best. Religious groups are the best studied; secular, athletic or creative groups have been studied far less but fit the same measure. Volunteering adds a role on top.
3. If you live alone, build contact deliberately. Living alone carried higher odds of death in pooled data.
4. If loneliness persists, try a psychological approach rather than just more social contact; in trials it appeared to work best, though the evidence is weak.

Further reading

• Holt-Lunstad J et al. Social relationships and mortality risk: a meta-analytic review. PLoS Med 2010.^[15]
• Wang F et al. A systematic review and meta-analysis of 90 cohort studies of social isolation, loneliness and mortality. Nat Hum Behav 2023.^[16]
• Holt-Lunstad J et al. Loneliness and social isolation as risk factors for mortality: a meta-analytic review. Perspect Psychol Sci 2015.^[17]
• Ai S et al. Social isolation and 59 common health conditions: insights from observational and genetics analyses. BMC Med 2026.^[18]
• Luchetti M et al. A Meta-analysis of Loneliness and Risk of Dementia using Longitudinal Data from >600,000 Individuals. Nat Ment Health 2024.^[19]
• Lasgaard M et al. Are loneliness interventions effective for reducing loneliness? A meta-analytic review of 280 studies. Am Psychol 2026.^[20]