In 2010, the psychologist Julianne Holt-Lunstad and her colleagues Timothy Smith and J. Bradley Layton set out to answer a narrower question than the one their findings ended up being remembered for: how much does the strength of a person’s social relationships actually predict how long they live, once the effect is measured across as much existing research as possible rather than in any single study. The result, published in PLOS Medicine, pooled data from 148 independent studies covering 308,849 participants followed for an average of seven and a half years. It remains one of the largest analyses ever conducted on the relationship between social connection and mortality.
What the numbers actually show
The headline figure is stark on its own terms: people with stronger social relationships had a 50 percent greater likelihood of survival over the study period than those with weaker ones, an effect that grew even larger — a 91 percent increase in survival odds — when researchers used more detailed measures of social integration rather than simple yes-or-no questions about whether someone lived alone. For comparison, the researchers found that the protective effect of strong relationships was comparable in size to well-established risk factors like smoking and heavy alcohol use, and larger than the effects of physical inactivity or obesity — two risk factors that receive vastly more public health attention and funding.
The comparison to smoking up to fifteen cigarettes a day, which has circulated widely since the study’s publication, comes directly from Holt-Lunstad’s own framing of the work. On a page addressing the figure directly, she has clarified that the comparison reflects an aggregate across multiple measures of social connection, not a single measure like loneliness in isolation — the fifteen-cigarette figure describes what happens when researchers look at social relationships broadly, including both the number and quality of close relationships a person maintains, rather than any one narrower slice of the concept.
An effect that doesn’t discriminate
What makes the finding harder to dismiss as a quirk of one particular population is how consistently it held up once the researchers broke the data apart. The association between social relationships and survival remained stable across age groups, held for both men and women, showed up regardless of a person’s health status at the start of the study, and persisted whether the eventual cause of death was cardiovascular, or something else entirely. It also held whether researchers followed participants for a few months or for decades. A risk factor that behaves this consistently across so many different subgroups is unusual — most health effects are stronger in some populations than others, or fade out over longer time horizons, or apply mainly to people who were already unwell. This one did none of that.
Why the effect might be this large
The researchers themselves were cautious about attributing the effect to any single cause, and the honest answer is that it likely runs through several overlapping pathways at once. Close relationships appear to buffer the physiological effects of chronic stress, moderating the cardiovascular and immune wear that stress produces when it goes unmanaged. They also function as an informal health-monitoring system: people with close, regular contact with others are more likely to have someone notice early symptoms, encourage a doctor’s visit, or simply catch a problem before it becomes a crisis. And they appear to shape health behavior directly — people with strong social ties are, on average, more likely to eat regularly, sleep on a consistent schedule, and avoid the kind of prolonged unstructured stress that erodes physical health over years rather than days.
Each of those mechanisms is unremarkable on its own. What’s notable is that they appear to compound, producing an effect on mortality that rivals or exceeds risk factors most public health campaigns treat as far more urgent. Smoking cessation programs, workplace wellness initiatives, and national health guidelines are all built around risk factors that, by this analysis, carry a comparable or smaller effect on how long a person is likely to live than the strength of their close relationships does.
A risk factor without a warning label
Part of what makes the finding difficult to act on, more than a decade after its publication, is that it doesn’t fit neatly into how health risk is normally communicated. Cigarette packaging carries an explicit warning. Nutrition labels quantify fat, sugar, and sodium. There is no equivalent label for a thin social life, no number that shows up on an annual physical the way blood pressure or cholesterol does, and no single behavior to quit the way a person can quit smoking. Having few close relationships isn’t a habit in the same sense — it’s often the byproduct of a move, a demanding job, a divorce, or simply years passing without anyone noticing the erosion happening in real time.
That may be exactly why the finding has had less practical impact than its size would predict. A risk factor this large, if it behaved like any other measurable health metric, would likely be treated as a standard part of preventive care — screened for, discussed, treated as seriously as blood pressure. Instead it remains something closer to common knowledge that hasn’t yet been translated into common practice: widely cited, rarely acted on, and easy to underestimate precisely because its damage accumulates quietly, over years, with nothing as visible as a cigarette to mark that it’s happening at all. Unlike smoking, there’s no single habit to quit here — just relationships worth building or rebuilding, at any point.
