The Overlooked Age Finding in the Mortality Literature

Julianne Holt-Lunstad’s 2015 meta-analysis in Perspectives on Psychological Science is cited constantly for its headline odds ratios: 1.29 for social isolation, 1.26 for loneliness, 1.32 for living alone, each a measurable increase in risk of early death. Less often repeated is a secondary finding buried in the same paper: these associations were stronger, not weaker, in samples with an average age under 65. That result cuts against the way loneliness policy has been built for the past five years, almost all of which treats social disconnection primarily as a problem of aging.

The National Academies’ 2020 consensus report is explicit about this framing. It estimates that roughly one quarter of adults 65 and older are socially isolated, and it calls on the health care system to screen for isolation and loneliness in geriatric settings specifically. The UK’s 2018 loneliness strategy, the first of its kind from any government, opens with the same emphasis: older people living alone, bereavement, retirement, reduced mobility. AARP’s 2018 survey of adults 45 and older, using the 20-item UCLA Loneliness Scale, put the figure at one in three lonely, and became one of the most widely cited prevalence numbers in the field. None of this is wrong. But it has produced a public conversation in which loneliness is substantially an older-adult health issue, and the mortality data on which the whole field rests does not obviously support that emphasis.

What the meta-analysis actually found

Holt-Lunstad’s 2015 review pooled studies with wide variation in participant age and found that when a study’s average age was under 65, the association between social deficits and death was more pronounced than in studies of predominantly older samples. This is not a finding about young people dying of loneliness at higher absolute rates than older people; older adults have higher absolute mortality regardless of social connection. It is a finding about the relative predictive strength of isolation and loneliness — how much extra risk they add on top of a person’s baseline. Among younger people, whose baseline mortality risk is lower, social disconnection appears to do proportionally more of the explanatory work.

This tracks with the earlier 2010 meta-analysis from the same research group, covering 148 studies and over 300,000 participants, which found that stronger social relationships corresponded to a 50% increased likelihood of survival, an effect comparable to smoking cessation. That paper did not disaggregate cleanly by age in the same way, but it established the baseline effect size against which the 2015 age finding stands out.

Neither paper offers a causal mechanism for why the relative effect would be stronger at younger ages. It could reflect that isolation in midlife is less normative and therefore more physiologically or socially costly than isolation in old age, where reduced social contact is more expected and perhaps better buffered by other adaptations. It could reflect selection effects in which studies of older cohorts already exclude the frailest and most isolated people, who died before enrollment. The sources available do not adjudicate between these explanations, and this article does not attempt to.

The prevalence data tells a consistent story, if anyone looked

Separately from mortality, the prevalence surveys have been pointing the same direction for years without much cross-reference to the mortality literature. Harvard’s Making Caring Common project found in 2021 that 61% of young adults aged 18 to 25 reported serious loneliness, against 36% of Americans overall. Cigna’s 2020 workplace survey found 73% of workers aged 18 to 22 reporting loneliness, and over 80% of employed Gen Z respondents. The American Enterprise Institute’s 2021 survey on friendship found that the share of men reporting six or more close friends fell from 55% in 1990 to 27% in 2021, with the friendship decline concentrated among younger cohorts rather than confined to retirement-age men losing colleagues and spouses.

Put the two literatures side by side and a pattern emerges that neither, on its own, states outright: younger adults report more loneliness on prevalence surveys, and the mortality literature suggests social deficits carry more relative weight in younger samples. Policy attention has gone almost entirely to older adults.

The 2023 Surgeon General’s advisory, the most comprehensive federal document on the subject to date, does broaden the frame somewhat, noting that roughly half of U.S. adults report loneliness and framing connection as a population-wide determinant of health rather than a geriatric concern. That is progress on the framing question, though the advisory’s six-pillar strategy does not single out the age pattern in the mortality data as a specific rationale.

Why this is hard to settle

None of the mortality studies were designed to test an age-by-isolation interaction as a primary hypothesis; the finding emerged from subgroup analysis across heterogeneous studies with different follow-up periods, different measures of isolation, and different baseline health. That is a real limitation. A subgroup pattern across pooled observational studies is suggestive, not confirmatory, and it says nothing about whether intervening on isolation in midlife would change mortality outcomes, since none of these are trials.

What would strengthen this would be a single longitudinal cohort followed from early adulthood through old age, using one consistent isolation and loneliness instrument throughout, with mortality tracked at every stage rather than assembled after the fact from studies that started measurement at different ages. Until that exists, the honest statement is narrower than either the older-adult narrative or the young-adult counter-narrative: the relative mortality cost of disconnection appears higher earlier in life than current policy attention would suggest, and nobody has yet built the study that would say why.