Do friends help you stay alive?

Meta-analysis of observational studies. 308,849 people. average follow-up 7.5 years (range 3 months to 58 years) across the included studies; literature searched January 1900 to January 2007.

Many studies, added up. People were watched, not told what to do. A clue, not proof.

Who: 308,849 participants in 148 English-language prospective studies; average age at intake 63.9; 49% female; 51% North America, 37% Europe, 11% Asia, 1% Australia; 60% of studies were community samples, 24% outpatients, 16% inpatients

• Many kinds of peopleMoststill alive
• Felt lonelyFewerstill alive

Stronger social relationships: 50% higher odds of survival (148 studies, 308,849 people)

Nobody can make people lonely for a test.

More to know, from our read of the paper:

• Meta-analysis of observational studies: association, not causation. The authors write 'causality is not easily established' because people cannot be randomly assigned to be isolated
• The 1.50 is an ODDS RATIO for survival, not '50% longer life' and not '50% lower death rate'
• Where studies reported several models, the authors deliberately took the estimate with the FEWEST statistical controls; for structural measures, controlled estimates were smaller than raw ones (metaregression B = -0.147, p = 0.01)
• Very high heterogeneity between studies (I2 = 81%); individual study ORs ranged from 0.77 to 6.50
• 'Social relationships' was measured in very different ways (marital status, network size, perceived support, loneliness, living alone...), mostly self-reported questionnaires; the comparison is generally 'stronger' vs 'weaker' within each study, not a fixed number of friends

Scientists added up 148 studies of 308,849 people (average age 64, mostly North America and Europe), followed for 7.5 years on average. People with stronger social relationships had 50% higher odds of still being alive (that's odds, not '50% longer life'). Living alone showed no clear link. These were observational studies, mostly with self-reported relationships, so it's linked, not proven: healthier people may simply have more friends.

We read the abstract and the full text of this paper and checked every number on this page against it (20 facts, checked on 2026-10-01). The study design is named above. Cohort and other observational studies are described as linked, not proven. Found an error? Email theis@onestudyaday.com.