What Loneliness Actually Costs: The Weakest Number in the Literature

The World Health Organization’s Commission on Social Connection put a number on it in June 2025: loneliness is linked to an estimated 871,000 deaths annually, roughly 100 an hour, worldwide. That figure has since appeared in policy briefs, opinion pieces, and conference slides as though it were a settled fact, comparable to the way public health has long quantified deaths attributable to smoking or air pollution. It is worth asking what that number is built on, because the answer says something about the state of cost evidence in this field generally: the epidemiology of loneliness is far more developed than the economics of it.

The mortality risk is real. The cost figure is a second inference on top of it.

Start with what is well established. Julianne Holt-Lunstad’s 2010 meta-analysis, covering 148 studies and 308,849 participants, found that stronger social relationships were associated with a 50% increased likelihood of survival, an effect size comparable to established risk factors like smoking or obesity. Her 2015 follow-up refined this into separate estimates: social isolation carried an odds ratio of 1.29 for early mortality, loneliness 1.26, and living alone 1.32. The American Heart Association’s 2022 scientific statement added cardiovascular specificity, putting the increased risk of heart attack or death from heart disease at roughly 29%, and of stroke at 32%.

These are association studies, and the authors are careful to say so. Deriving a global death toll like the WHO’s 871,000 requires taking a prevalence estimate — how many people are lonely — and multiplying it by an assumed excess mortality risk, then summing across populations with very different baseline mortality rates, health systems, and definitions of loneliness. Each step compounds uncertainty from the last. The WHO report itself does not claim clinical-trial-grade precision for this figure; it presents it as a population-attributable estimate, which is a legitimate epidemiological method but a different kind of claim than “loneliness caused these deaths.” The distinction rarely survives the trip into a news headline or a briefing slide.

The dollar figures are thinner still

If the mortality-attribution literature is imperfect but methodologically real, the direct cost-to-health-systems literature is comparatively undeveloped. The 2020 National Academies consensus report on social isolation in older adults, one of the most rigorous syntheses available, is notable for what it does not do: it calls for the health care system to routinely screen for isolation and loneliness, but it does not attach a firm national cost figure to inaction, because the underlying utilization data were not judged strong enough to support one.

Where cost-adjacent numbers do exist, they tend to come from employer surveys rather than health economics research. Cigna’s 2020 workplace report found that lonely workers miss work twice as often due to illness and five times as often due to stress — a striking ratio, but one drawn from self-reported absenteeism in a single commissioned survey, not from claims data or a controlled comparison. It is useful as a signal that loneliness has workplace costs. It is not a basis for a dollar figure, and Cigna does not present it as one.

The CDC’s 2024 MMWR surveillance report links social disconnection to heart disease, stroke, dementia, type 2 diabetes, depression, anxiety, and premature mortality, using 2022 survey data. This is a list of associated conditions, each with its own established treatment cost literature elsewhere. But the CDC report does not itself sum those conditions into an attributable national health expenditure figure for loneliness. Anyone who has seen a specific dollar figure — “loneliness costs the US healthcare system $X billion annually” — attributed loosely to “the CDC” or “federal researchers” should treat that figure with skepticism about its actual source and method, because the primary document does not contain such a calculation.

The 2023 Surgeon General’s advisory, which did the most of any single document to popularize the mortality comparison to smoking up to 15 cigarettes a day, is similarly disciplined about not extending that comparison into direct cost terms. The advisory’s six-pillar national strategy is framed around access, infrastructure, and culture rather than a projected savings figure, which is itself informative about how much confidence the authors placed in monetized estimates.

Why the gap persists

The 2023 BMC Public Health review of the state of loneliness research identifies inconsistent measurement as a central barrier to comparing findings across studies. That problem is worse, not better, when the question shifts from prevalence to cost. Isolation and loneliness are measured with different instruments across studies — the UCLA Loneliness Scale, single-item self-report questions, structural network measures — and attaching healthcare utilization or productivity cost to each requires linking survey data to claims or payroll data that most of these studies simply do not have. The 2020 National Academies report explicitly frames better clinical assessment as a prerequisite for better cost evidence, not a parallel track: it is difficult to cost a condition that health systems are not yet routinely screening for.

What would actually settle this

A defensible cost estimate for loneliness would need three things the field mostly lacks: a consistent case definition applied across the studies being pooled, a linkage between individual-level loneliness or isolation status and individual-level healthcare claims or employment records rather than self-reported absenteeism, and a control group or comparable counterfactual rather than a before-and-after or cross-sectional comparison. The randomized trials that do exist in this space, such as the HEAL-HOA studies of behavioral interventions in older adults, measure loneliness reduction, not cost offset, and none of the sources reviewed here report a health-economic outcome from a controlled trial. Until that kind of study exists, the honest summary is that loneliness’s mortality burden rests on a genuinely strong evidence base, while its financial burden rests mostly on extrapolation, survey self-report, and analogy to other risk factors whose own cost accounting took decades to build.