Loneliness and health: what the research shows and what it doesn't

It is Sunday evening. Your phone has been silent all day, apart from a notification from the bank. Scrolling, you pass a headline: “Loneliness is as bad for your health as fifteen cigarettes a day.” Something tightens. If you have not exchanged more than two sentences with anyone this weekend, other than at the till, a headline like that does not read as news. It reads as a verdict.

There are more of these headlines every year, and serious research stands behind them. But a good deal gets lost between what researchers actually measured and what reaches the headline. In this post I would like to do two things: give a fair summary of what the two most frequently cited analyses, both led by the American psychologist Julianne Holt-Lunstad, actually show, and say what you cannot conclude from them. Above all, that one quiet weekend is not a danger to your heart.

First, though, we need to separate three words that everyday conversation throws into the same basket.

Solitude, loneliness and social isolation are not the same

Solitude is a state: I am alone, without company. It can be chosen and pleasant, such as a walk without your phone or a morning coffee before the house wakes up. There is more about it in the knowledge base under solitude.

Loneliness is a feeling. In the early 1980s the psychologists Daniel Perlman and Letitia Anne Peplau described it as the unpleasant experience that arises when there is a gap between the relationships you have and the relationships you want, whether in number or in depth. That is why a person can be lonely in a full office, in a marriage or at a festive table. And it is why someone who lives alone and sees two friends a month may not be lonely at all. A short explanation is also in the entry on loneliness.

Social isolation is something that can be counted from the outside: few contacts, a small network of people, rarely taking part in anything shared. It says nothing about how the person feels about it.

The distinction is not hair-splitting. Studies sometimes measure one and sometimes another, and the results do not always agree. Besides, each of the three calls for something different: solitude does not need “treating”, isolation is helped by more opportunities for contact, and loneliness is often helped not by more people but by a different kind of contact with them.

What the 2010 analysis showed

In 2010 Julianne Holt-Lunstad, Timothy Smith and Bradley Layton published a meta-analysis in the journal PLoS Medicine, that is, a study that pools the results of many earlier studies. They gathered 148 studies involving more than 308,000 people in total. Each of those studies had measured something about participants’ relationships at the outset and then tracked who was still alive years later. People were followed for seven and a half years on average.

The main finding: people with stronger social relationships were about 50 per cent more likely to be alive at the end of follow-up than people with weak ones. The association did not differ meaningfully by sex, age, initial health status or cause of death.

There is an interesting detail that rarely makes the headlines. The association was strongest where researchers had measured social integration in several ways at once: whether a person has a partner, friends, groups they belong to, someone they can rely on. It was weakest for the simplest question of all, whether a person lives alone or with someone. The number of people in a flat, taken by itself, tells you very little.

And what the 2015 one added

The first analysis looked at relationships in general. Five years later the same researcher and her colleagues asked a narrower question in the journal Perspectives on Psychological Science: what about loneliness, isolation and living alone specifically, each taken separately? This time they pooled 70 studies with more than 3.4 million participants, followed for seven years on average.

After other factors were taken into account, the likelihood of dying during follow-up was about 26 per cent higher among people who reported loneliness, 29 per cent higher among the socially isolated and 32 per cent higher among those living alone. The authors found no meaningful difference in the size of the effect between the feeling (loneliness) and the countable state (isolation).

One more detail is surprising: the association was stronger in samples with an average age under 65 than among older people. We tend to picture loneliness as a problem of late old age, yet the data suggest the body registers it in mid-life too.

Where the cigarette comparison comes from

The fifteen cigarettes in the headline are not invented, but the sentence is a long way from what was measured. The authors of the 2010 analysis set their result alongside effect sizes that other research had calculated for well-known risk factors. Weak relationships landed in the same order of magnitude as smoking, and above physical inactivity and obesity.

That is a comparison of statistical figures from different studies, not an experiment in which anyone measured the harm of a single lonely evening. The purpose of the comparison was a public health one: to tell doctors and decision-makers that relationships belong on the same list as blood pressure and exercise. It was not meant as a message to an individual that every quiet evening shortens their life.

It also matters what “26 per cent more likely” means. It is a relative difference between groups, not your personal forecast. If the chance that a healthy forty-year-old will die in the next seven years is very small, then a chance that is a quarter larger is still very small. The numbers describe large groups of people over many years. They know nothing about you on a Sunday evening.

How loneliness gets under the skin

An association alone does not tell us how a feeling might affect the heart and blood vessels. In a 2010 review article, Louise Hawkley and John Cacioppo gathered the pathways with the most evidence behind them. None of them is mysterious.

The body stays on alert. Cacioppo compared loneliness to hunger: it is a signal that something essential is missing. For a species that survived in groups, an individual at the edge of the group was in danger, so when we feel cut off, the body prepares for threat. In the short term that makes sense. If it lasts for months or years, it means more stress hormones and, over time, higher blood pressure; in one study by the same group, loneliness predicted a greater rise in blood pressure over the following years. How this kind of long-running overdrive wears the body down is described in the entry on chronic stress.

Sleep becomes shallower. Lonely people in studies do not necessarily sleep fewer hours, but their sleep is more fragmented, with more brief awakenings, and they wake less rested. The explanation is similar: someone who does not feel safe sleeps with one eye open.

The immune system shifts. In people who have been lonely for a long time, researchers have observed more signs of quiet, chronic inflammation, a state linked to cardiovascular disease. These studies are smaller and the findings less solid than for blood pressure, so I treat them as a direction, not as proof.

Habits change. Someone who is alone a lot cooks less often, moves less, reaches for alcohol more easily and notices later that something is wrong. On top of that, there is no one to say “go and see a doctor”, to drive them to an appointment or to notice that they have been different these past few weeks. This last pathway is the most prosaic and probably one of the most important.

A British meta-analysis led by Nicole Valtorta in 2016 focused on the heart and brain alone: weak social relationships were associated with a roughly 29 per cent higher risk of coronary heart disease and a 32 per cent higher risk of stroke. Here too we are talking about an association in large groups, not a forecast. I describe a similar bodily story after the loss of someone close in the post Grief and the body: why we fall ill more often after a loss.

A loop that turns by itself

The most useful part of Cacioppo’s work, to my mind, is not the biology but the description of what loneliness does to attention. When it lasts, the brain becomes more alert to signs of rejection. A neutral face seems cold, a late reply to a message reads as lack of interest, an invitation as mere politeness. So the person protects themselves: they suggest less, cancel sooner, say less about themselves in conversation. Others read this as disinterest and step back. The expectation is confirmed.

A conversation between two acquaintances might sound like this (a composite, not a real person):

“We’re going up Šmarna gora on Saturday. Coming?”
“I don’t know, you’ve probably got your own crowd already.”
“That’s why I’m inviting you.”
“We’ll see. Last time I felt like I was in the way.”
“Last time you left after half an hour. We thought you were bored with us.”

Neither of them is lying. One left because he felt in the way; the others concluded that he did not like them. This loop is especially stubborn alongside depression, which I write about in the post Why depression makes you withdraw just when you need people most.

There is no blame in this. Heightened alertness to rejection is not a character flaw but a predictable response of a nervous system that has gone too long without a sign that it is among its own. It is useful to know, though, that during such a period your own interpretations of other people’s behaviour cannot be fully trusted.

What helps, according to research

In 2011 Christopher Masi and colleagues reviewed studies of programmes for reducing loneliness and sorted them into four groups: teaching social skills, increasing social support, creating more opportunities for contact, and working with the thoughts loneliness brings with it. In randomised studies the last group did best, in other words, checking precisely those automatic interpretations from the previous section.

That does not mean opportunities for contact are unimportant. Someone who works from home all week in a town where they know nobody needs people first, not work on their thoughts. What is lost when chance encounters disappear is the subject of the post Working from home and quiet loneliness: what is lost when the corridor disappears. It does mean that the advice “get out among people more” is often not enough if a person reads every encounter through the expectation that they are not wanted.

In Slovenia we have something for regular, undemanding contact that many places lack: local societies and clubs. A choir, a mountaineering club, the volunteer fire brigade, allotment gardeners, a reading group at the library. Their advantage is that people meet around an activity, not around talking about themselves, and that you are expected again next week. The feeling of having somewhere you fit is described in the knowledge base under belonging. How to go about new ties in mid-life, when school and university no longer do it for you, is the subject of the post How to make friends at forty: why it is harder and what helps.

The research points to one more thing that is easy to forget: quality counts, not the total. In the 2010 analysis the most predictive measures were those that captured several kinds of ties at once. One person you can tell how you really are, and one place where you are expected, are worth more to begin with than ten new numbers in your phone.

As a therapist in training under supervision in relational family therapy, I look at loneliness mainly through relationships: how you learned how much closeness is safe, and what you do when someone comes nearer. Loneliness within a relationship or a family is often a question of exactly that, not of the number of people around you. Sessions are held in person in Ljubljana, and I also work with English-speaking clients.

An exercise: reviewing one week of contact

The exercise takes ten minutes and a sheet of paper. Its purpose is not a grade but a picture: most people do not know their own week as well as they think.

1. Write down every contact from the past seven days. Really all of them: a chat with a neighbour on the stairs, a call from your mother, a message from a friend, coffee with a colleague, the shop assistant who knows you.
2. Next to each one, mark how you felt afterwards: better (+), the same (0) or worse (–). Not how you ought to have felt, but how you did.
3. Circle the contacts in which you said something true about yourself. One sentence is enough, for example “I’m tired this week”.
4. Look at the sheet and answer: am I short of the number of contacts, of depth, or both? The first is closer to isolation, the second to loneliness.
5. Choose one small move for next week. If the number is missing, one recurring slot (a class, a choir, the market at the same time each week). If depth is missing, tell one person with a plus one sentence more than usual.

If the sheet is almost empty, that is not a judgement on you. It is information about the week you have now, and a starting point for the next one.

Where these numbers do not tell the whole story

Since I have cited research throughout this post, it is only right to say where its limits lie.

An association is not a cause. All of these studies are observational: nobody randomly assigned people to be lonely or not. It may be that loneliness harms health, but it may also be that illness, poverty or depression damage health and shrink a person’s circle at the same time. The authors took much of this into account statistically, but not everything can be. The direction probably runs both ways.

The results are not entirely consistent. In 2013 Andrew Steptoe and colleagues followed older people in England and found that social isolation predicted higher mortality even after health and other circumstances were taken into account, whereas loneliness, after those adjustments, no longer did. The feeling and the countable state are therefore not interchangeable, and researchers do not yet agree on which of the two matters more.

This is about a long-term state. The studies speak of years, not of the week after a move, the winter after a break-up or the first year in a new city. Passing loneliness is an ordinary part of life and, on Cacioppo’s account, even a useful signal, much like thirst.

Solitude is not a risk factor. Some people need a great deal of time to themselves and are content with it. If you are one of them, these numbers are not about you and nobody has the right to prescribe you more company. Living without housemates is not a problem in itself either; in the 2010 analysis that very piece of information was the weakest predictor.

The answer is not always inside you. The loneliness of an older person in a village with no bus, of a mother with a baby in a block of flats where she knows no one, or of an immigrant who does not yet speak the language is not a matter of mistaken thoughts. There the answer is transport, space, time and money. The studies I am summarising were also written so that the people who decide about such things would hear it.

When it is a good idea to seek help

Feeling lonely is not an illness in itself and does not require a professional. Talking to someone does make sense if it has lasted several months without shifting, if you notice yourself turning down invitations you actually want, if it comes with listlessness, poor sleep or a loss of pleasure in things you used to enjoy, or if you feel lonely precisely in the relationships you have. For physical problems such as high blood pressure, insomnia or chest pain, your GP is the first port of call; a conversation with a therapist does not replace that.

If loneliness comes with thoughts that nothing has any point, or that it would be better if you were not here, do not wait. In Slovenia you can call the Samarijan confidential helpline on 116 123 (24 hours a day) or the Klic v duševni stiski crisis line on 01 520 99 00 (every day from 7 pm to 7 am), and in an emergency 112.

For many people the hardest days are those when everyone else seems to be sitting round the same table. I write about them in the post Spending the holidays alone: how to get through them without pretending.

What to remember

• Solitude is a state, loneliness is a felt gap, social isolation is a small number of contacts. They can overlap, but they are not the same.
• Julianne Holt-Lunstad’s meta-analyses (2010, 2015) show that, in large groups and over years, weak relationships are associated with higher mortality, on a scale similar to well-known risk factors.
• The cigarette comparison is a comparison of statistical effects, not a forecast for you; one quiet weekend does not harm your heart.
• The most likely pathways are a body kept on long-term alert, poorer sleep, inflammation, changed habits and fewer people to notice that something is wrong.
• Long-term loneliness is often helped less by more people than by checking your own interpretations and by one regular, undemanding contact.