Loneliness is a health risk. What the research actually shows

You will have seen the statistic. Social isolation is as bad for you as smoking fifteen cigarettes a day. It appears in newspaper columns, public health campaigns and ministerial speeches, usually without a source.

The source is a 2010 meta-analysis by Julianne Holt-Lunstad and colleagues, pooling 148 studies covering some 300,000 people. It found that individuals with stronger social relationships had a 50 per cent greater likelihood of survival over the follow-up periods studied. The comparison to smoking came from lining that effect size up against the mortality effects of other established risk factors, where it sat in similar territory.

It is a legitimate piece of work and a slightly misleading soundbite. Effect sizes for mortality are not directly interchangeable across risk factors with different mechanisms and time courses, and the direction of causation is genuinely hard to pin down — illness causes withdrawal at least as readily as withdrawal causes illness. But the association survives adjustment for baseline health in most of the better cohorts, and it has been replicated repeatedly since. Something real is being measured.

Loneliness and isolation are not the same variable

This distinction does most of the useful work and is routinely collapsed in press coverage.

Social isolation is objective: how many people you interact with, how often, whether you live alone, whether you belong to anything. Loneliness is subjective: the distress of a gap between the relationships you have and the ones you want.

They correlate only moderately. Plenty of people live alone, see few people, and are perfectly content — the sociological literature on this is quite firm, and it tends to annoy campaigners. Plenty of others are surrounded by family and profoundly lonely. When researchers measure both, they often turn out to predict different outcomes: objective isolation shows a stronger association with mortality, while loneliness shows a stronger association with depression and with cognitive decline.

The practical implication is that interventions aimed at putting lonely people in rooms with other people have a mixed record. A coffee morning solves isolation. It does not necessarily solve loneliness, and it can occasionally make it worse.

What the biology looks like

Several plausible pathways have been described, and they are not mutually exclusive.

Chronically lonely people show, on average, higher circulating markers of inflammation, altered cortisol patterns, higher blood pressure and poorer sleep quality. Steve Cole's work on what he calls the conserved transcriptional response to adversity found a distinctive pattern of gene expression in lonely individuals — upregulated inflammatory genes, downregulated antiviral genes — which is biologically coherent if you think of loneliness as an evolved threat signal.

There are also unromantic behavioural pathways. People who live alone eat worse, drink more, exercise less, notice symptoms later, and have nobody to insist they go to the doctor. Some of the mortality signal is almost certainly this rather than anything about inflammation.

What actually helps

The intervention literature is weaker than the epidemiology, which is a familiar pattern in this field. A frequently cited review by Masi and colleagues categorised interventions into four types and found that the most effective were not the ones that increase social contact, provide social support, or improve social skills — but those addressing maladaptive social cognition. That is, the thought patterns that make a lonely person expect rejection, read neutral behaviour as hostile, and withdraw pre-emptively.

This is a genuinely useful finding and an uncomfortable one, because it is easier to fund a lunch club than cognitive therapy. It also should not be over-read; the trials in question were small and heterogeneous.

What else has reasonable support:

• Activity built around a shared purpose rather than around socialising. Choirs, volunteering, walking groups, allotments, classes. The interaction is a by-product, which appears to lower the stakes.
• Regularity. A weekly fixture beats an occasional gathering, partly because it survives a bad week.
• Contributing rather than receiving. Volunteering shows some of the more consistent wellbeing associations in older adults, and there is a plausible reason: being needed is different from being visited.
• Treating the hearing loss. Unaddressed hearing loss is one of the commonest routes into social withdrawal — conversation in a group becomes exhausting, so people stop going. This is a fixable problem misfiled as a personality change.

What to be sceptical of

Befriending schemes have a thinner evidence base than their prominence suggests. So do technology interventions — tablets and video calling — which help people who already have relationships to maintain them, and do relatively little for those who do not.

And "loneliness epidemic" framing deserves some scrutiny. The best longitudinal data does not show a straightforward rise in loneliness among older adults over recent decades; some analyses find it stable or slightly falling in the over-sixties, with the sharper increases among young adults. The problem is serious. It is not obviously new.

The one thing worth doing

If you are reading this on behalf of a parent: the single most useful intervention is usually not more visits. It is helping them acquire a recurring commitment that involves other people and does not depend on you — a Tuesday. Visits fill an afternoon. A Tuesday changes a week.