How Loneliness Became 'Equivalent to Smoking 15 Cigarettes a Day'

The single most-repeated statistic in loneliness policy is a comparison, not a measurement: that lacking social connection carries a mortality risk “comparable to smoking up to 15 cigarettes a day.” It appears in the 2023 U.S. Surgeon General’s advisory on the epidemic of loneliness and isolation, and from there it has propagated through press coverage, advocacy materials, and policy briefings, including some published well after the advisory itself. It is a good example of how a number can travel a long way from its source study while gaining certainty it never had.

Where the number comes from

The comparison traces back to Julianne Holt-Lunstad’s 2010 meta-analysis in PLoS Medicine, which pooled 148 studies covering 308,849 participants and found that stronger social relationships were associated with a 50% increased likelihood of survival over the study follow-up periods. The paper’s contribution was not a single hazard ratio but a benchmarking exercise: it placed the mortality effect of weak social ties alongside other well-studied risk factors, including smoking, obesity, and physical inactivity, to argue that social connection deserved comparable public health attention. It was, from the outset, an argument about relative magnitude, not a claim that a specific quantity of loneliness produces the same biological harm as a specific quantity of tobacco smoke.

Holt-Lunstad’s 2015 follow-up in Perspectives on Psychological Science sharpened the picture and complicated the smoking analogy further. That meta-analysis separated three related but distinct exposures: social isolation, loneliness, and living alone, reporting odds ratios of 1.29, 1.26, and 1.32 respectively for early mortality. These are meaningfully different numbers from a “50% increased likelihood of survival,” and they are also different in kind. An odds ratio of 1.29 for isolation is not the same statistical object as a hazard ratio comparing high versus low social integration, and neither is directly convertible into a cigarette count. The 2015 paper also found that these effects were stronger in samples averaging under 65, a detail that gets lost whenever the top-line comparison is quoted without qualification.

The advisory compresses two meta-analyses into one line

By the time the Surgeon General’s advisory was published in 2023, the smoking comparison had become a settled rhetorical device rather than a reported effect size. The advisory states plainly that the mortality risk of social disconnection is comparable to smoking up to 15 cigarettes a day. It does not present this as a new calculation; it is a restatement, compressed for a policy audience, of the general magnitude established across Holt-Lunstad’s two meta-analyses roughly a decade apart. The advisory’s real contribution was elsewhere: the finding that roughly half of U.S. adults report experiencing loneliness, and the six-pillar National Strategy to Advance Social Connection that the advisory used the smoking comparison to justify.

That is the pivot point. A benchmarking exercise designed to argue for public health parity between social connection and other risk factors became, in the advisory’s framing, a fixed number attached to a specific behavior. Press coverage of the advisory largely reproduced the cigarette figure as if it were a directly observed dose-response relationship, because it is more quotable than an odds ratio.

A parallel, more modest number gets less attention

It is worth setting the smoking comparison against the American Heart Association’s 2022 scientific statement, published the year before the advisory, which took a narrower and more cautious approach. That statement, led by Crystal Cene on behalf of multiple AHA councils, put the increased risk of heart attack, stroke, or death from either at roughly 30% associated with social isolation and loneliness, breaking it down further into a 29% increased risk of heart attack or death from heart disease and a 32% increased risk of stroke. These are the kind of numbers that survive close reading: specific outcomes, specific risk ratios, and an explicit acknowledgment, stated directly in the AHA statement, that the absence of intervention evidence is the central gap in the field. The AHA’s press materials repeated the 30% figure rather than reaching for a smoking analogy, and it did not travel nearly as far in subsequent coverage. A precise, modest number apparently spreads less than an imprecise, dramatic one.

What the WHO report added, and what it did not

The World Health Organization’s 2025 Commission on Social Connection report introduced a different headline figure: an estimated 871,000 deaths annually attributable to loneliness worldwide, or roughly 100 an hour. This is a fresh estimate rather than a restatement of the smoking comparison, and it has begun to circulate alongside rather than instead of it. The two statistics are not in tension, but they are not interchangeable either; one is a relative risk analogy rooted in two specific meta-analyses, the other is an absolute mortality estimate applied globally. Reports that cite both without distinguishing them risk implying they are the same kind of evidence.

The policy uptake outran the precision

The UK’s 2018 loneliness strategy, the first national strategy of its kind, predates the cigarette comparison in its current form and built its case primarily on prevalence rather than mortality analogy. Later policy documents have had the smoking figure available and have used it because it survives translation into a single sentence a minister or a journalist can repeat. A 2023 review in BMC Public Health flagged inconsistent measurement as one of the field’s core methodological problems; the smoking comparison is a case study in that inconsistency, not because either underlying meta-analysis was flawed, but because a magnitude comparison, once it enters advocacy language, tends to be treated as a measurement.

A more useful public figure would separate the isolation, loneliness, and living-alone estimates the way Holt-Lunstad’s 2015 paper did, rather than compressing them into a single smoking-equivalent number that no single study actually produced.