In 2010, Julianne Holt-Lunstad, Timothy B. Smith and J. Bradley Layton published a meta-analysis of social relationships and mortality. It combined 148 studies involving 308,849 participants.
Stronger social relationships were associated with better survival. The pooled survival odds ratio was 1.50, with a 95 per cent confidence interval of 1.42 to 1.59. This is an odds comparison, not a finding that weakly connected people were exactly 50 per cent more likely to die.
What the smoking comparison means
In her explanation of the comparison, Holt-Lunstad describes lacking social connection as comparable to smoking up to 15 cigarettes daily. The comparison benchmarks effect sizes from separate bodies of research. It is not a calculation of how many cigarettes an individual person’s loneliness is equivalent to.
The wording matters. The 2010 analysis pooled several measures of social connection, including social networks, support and isolation. Loneliness was one component, not the sole exposure. The number should therefore not be presented as an exact effect of chronic loneliness.
Loneliness and isolation are different
A 2015 follow-up review examined loneliness, social isolation and living alone separately. Feeling lonely, having few social contacts and living alone can overlap, but they are not interchangeable descriptions.
A risk association, not a personal prediction
The 2010 evidence came from prospective observational studies. Pooling those studies strengthens the estimate of an association, but does not turn it into a randomised test or establish a precise causal effect for each person.
The evidence supports taking social connection seriously in public health. The responsible version of the headline preserves the study’s broader exposure, the association with survival and the limits of the comparison.
Correction, 3 October 2026: The earlier headline treated a broad social-connection finding as an exact effect of chronic loneliness. The body also misinterpreted a survival odds ratio as a 50 per cent increase in death probability. Both have been corrected, and unsupported claims about the study’s historical and policy impact have been removed.