What the Loneliness Literature Does Not Cost Out

The strongest number in this literature is a mortality effect size. The weakest is anything denominated in dollars or pounds. That asymmetry is worth sitting with, because a great deal of policy writing on loneliness leans on cost figures that the underlying research does not actually support.

Start with what is solid. Julianne Holt-Lunstad’s 2010 meta-analysis, pooling 148 studies and 308,849 participants, found that stronger social relationships were associated with a 50% increased likelihood of survival — an effect size the authors compared to established risk factors like smoking. Her 2015 follow-up, focused specifically on isolation and loneliness rather than relationship quality broadly, put social isolation at an odds ratio of 1.29 for early mortality, loneliness at 1.26, and living alone at 1.32, holding after adjustment for health status. The American Heart Association’s 2022 scientific statement, led by Crystal W. Cene, converted similar evidence into cardiovascular terms: roughly 29% increased risk of heart attack or death from heart disease, 32% increased risk of stroke. These are large, multiply-replicated, cross-population effects. Nobody serious disputes that isolation and loneliness carry health cost in the epidemiological sense.

None of that is a dollar figure. And the gap between “carries mortality risk comparable to smoking” and “costs the health system $X billion a year” is where the literature gets thin fast.

Where the cost claims actually come from

Cost estimates that circulate in policy documents and press coverage tend to originate from one of three places, and each has a specific weakness.

The first is extrapolation from healthcare utilization studies. The National Academies’ 2020 consensus report on social isolation in older adults notes that roughly one quarter of adults 65 and older are socially isolated, and recommends the health care system routinely assess for it — but the report is built around a clinical-practice argument, not a costed one. Its accompanying commentary in the American Journal of Geriatric Psychiatry pushes further on what routine assessment would require operationally, but stops short of putting a figure on what isolation costs a health system to leave unaddressed, or what screening would cost to implement at scale. The absence is not an oversight; it reflects that the underlying utilization data — emergency visits, readmissions, GP contacts attributable specifically to isolation rather than to the comorbidities that often travel with it — is not well isolated in the source studies the report draws on.

The second source is employer survey data, which is more concrete but narrower in scope. Cigna’s 2020 workplace report found that lonely workers miss work roughly twice as often due to illness and five times as often due to stress, and that 61% of U.S. adults reported feeling lonely sometimes or always, up seven points year over year. That is a genuine, quantified productivity signal — but it is self-reported absenteeism from a single corporate survey instrument, not a national accounting exercise, and it says nothing about direct healthcare spending. Extending it into an economy-wide cost figure requires assumptions about wage levels, replacement costs, and causal direction that the survey itself does not test.

The third source is intervention evidence, and this is where the AHA statement is most useful precisely because of what it declines to claim. The 2022 statement explicitly identifies the absence of intervention evidence as the central research gap in this field. That matters for cost estimation specifically: a cost-of-inattention figure implies a counterfactual — what would be saved if isolation were addressed — and that counterfactual requires trials showing an intervention reduces isolation, sustains the reduction, and reduces downstream health utilization as a result. The AHA statement’s own summary in the association’s newsroom repeats the epidemiological risk figures but does not attach a savings estimate, which is a reasonable choice given the gap the statement itself names.

The social prescribing literature illustrates the problem well

Social prescribing — referring patients to community activities, groups, or services rather than, or alongside, clinical treatment — is the intervention most often invoked as loneliness’s return-on-investment case. Two systematic reviews published in 2021 assessed the evidence base directly. One, in the International Journal of Environmental Research and Public Health, reported increases in self-esteem and self-confidence as the consistent outcomes across included studies, while explicitly noting limited trial evidence and heterogeneity across programmes. The other, in Perspectives in Public Health, reviewed nine studies and found all nine reported positive individual impacts, with three reporting reductions in GP, emergency, social worker, or inpatient service use.

Three studies reporting reduced service use, out of nine total, in a field the reviewers themselves describe as heterogeneous, is not a basis for a national savings figure. It is a basis for cautious optimism and a call for better-controlled trials. The UK’s 2018 loneliness strategy, the first national strategy of its kind, funded social prescribing partly on this logic and embedded loneliness measurement into the Office for National Statistics — a genuinely useful step for future cost analysis, because it creates a consistent measurement baseline. But a funding decision made in anticipation of stronger future evidence is a different thing from evidence that the funding pays for itself, and the strategy does not claim otherwise.

What a defensible cost estimate would require

Holt-Lunstad’s 2021 review in the American Journal of Lifestyle Medicine argues that social connection belongs in preventive health frameworks alongside diet, exercise, and smoking cessation. That comparison is instructive for cost estimation, because the field’s understanding of smoking’s economic cost rests on decades of attributable-risk calculations tied to specific, measurable disease pathways and validated by natural experiments — tax changes, workplace bans — that isolated causal effects on utilization and spending. Loneliness research has the epidemiological half of that architecture. It largely lacks the intervention and natural-experiment half.

A credible cost figure for loneliness or isolation would need three things the current literature does not yet supply together: a validated attributable-risk fraction linking isolation to specific downstream conditions rather than all-cause mortality broadly; utilization data that separates isolation’s marginal contribution from comorbid conditions that commonly accompany it, such as depression or mobility limitation; and intervention trials large and long enough to establish that a given program changes utilization, not merely self-reported wellbeing. Until those exist, any dollar figure attached to loneliness should be read as an illustrative extrapolation from health-risk data, not as a costed finding — a distinction that gets lost reliably between the research and the press release.