How '15 Cigarettes a Day' Became the Number Everyone Cites

Almost every policy document on loneliness published in the last five years contains some version of the same sentence: that lacking social connection carries a mortality risk comparable to smoking up to 15 cigarettes a day. It appears in the U.S. Surgeon General’s 2023 advisory, in press coverage of the American Heart Association’s 2022 scientific statement, and, more quietly, in the framing behind the World Health Organization’s 2023 Commission on Social Connection. It is one of the most-cited numbers in the field. It is also worth tracing back to where it actually came from, because the trip from statistical finding to public health talking point involved a fair amount of simplification.

The original comparison

The starting point is Julianne Holt-Lunstad’s 2010 meta-analysis in PLoS Medicine, pooling 148 studies and 308,849 participants. It found that stronger social relationships were associated with a 50% increased likelihood of survival over the follow-up periods studied. Holt-Lunstad’s point, made explicitly in that paper and repeated in her subsequent work, was that this effect size sits alongside other well-established mortality risk factors — smoking being the most recognizable comparator to a general audience. That is a comparison of relative effect sizes across different bodies of literature, not a finding from a single study that measured loneliness and cigarette consumption side by side and calculated an equivalence.

Her 2015 meta-analysis in Perspectives on Psychological Science refined the picture considerably. Isolation, loneliness, and living alone were examined as separate risk factors, with odds ratios of 1.29, 1.26, and 1.32 respectively for early mortality. These are meaningfully smaller numbers than “50% increased survival,” because they come from a differently structured analysis and separate three constructs that the 2010 paper had sometimes treated together. The 2015 paper also found something a press release would be unlikely to headline: the effects were more pronounced in samples averaging under 65, which complicates any story about loneliness as chiefly a problem of old age.

What the advisory did with it

By the time the Surgeon General’s 2023 advisory was drafted, the “comparable to smoking” framing had become a fixed phrase, cited as fact rather than as an effect-size analogy. The advisory states that the mortality risk of social disconnection is comparable to smoking up to 15 cigarettes a day, presented without the qualifications present in the original papers about which construct — isolation, loneliness, or living alone — is doing the work, or about the studies’ cross-sectional versus longitudinal composition. The advisory’s own reference base includes the 2010 and 2015 meta-analyses, so the lineage is traceable. But the number has been flattened into a single figure attached to “social disconnection” as an undifferentiated category, which is precisely the conflation Holt-Lunstad’s 2015 paper had tried to move past.

The American Heart Association’s 2022 scientific statement handled the same underlying literature with more care, reporting isolation and loneliness as carrying roughly a 30% increased risk of heart attack, stroke, or death from either — 29% for heart attack and death from heart disease, 32% for stroke. That statement also states directly, as its central finding, that there is no strong intervention evidence: none of the studies pooled show that addressing isolation reduces cardiovascular risk, only that isolation and poor outcomes travel together. The AHA newsroom summary of that statement, aimed at general readers, kept the 30% figures rather than reaching for the smoking comparison, which suggests the flattening happened somewhere between the academic literature and the advisory-writing process, not inside the cardiology literature itself.

Where it landed

The number’s final stop, so far, is the CDC’s Morbidity and Mortality Weekly Report, published in June 2024 using 2022 survey data. That report links loneliness and lack of social and emotional support to heart disease, stroke, dementia, type 2 diabetes, depression, anxiety, and premature mortality, treating the connection as sufficiently established to warrant federal surveillance infrastructure. It does not repeat the cigarette comparison, but it inherits the same underlying assumption: that the association identified in the 2010 and 2015 meta-analyses is stable enough to build population health monitoring on. Given that a 2023 review in BMC Public Health identified inconsistent measurement as one of the field’s persistent methodological gaps, that assumption is doing more work than it usually gets credit for.

None of this means the underlying association is wrong. Two well-conducted meta-analyses, examined a decade apart, both found social disconnection predicts earlier death, and the effect held after adjusting for baseline health status. What the “15 cigarettes” figure obscures is that this is an association derived from observational studies, not a measured causal pathway with a known dose-response curve the way tobacco epidemiology eventually built one. The UK’s 2018 loneliness strategy, to its credit, described loneliness as comparable to obesity and inactivity as a public health issue without reaching for the cigarette figure at all. That the comparison persists anyway, three institutional layers removed from the paper that first suggested it, says something about which numbers survive the trip from journal to policy document: not necessarily the most precise ones, but the most quotable.