Being a bit isolated

The single most important thing to understand about this evidence is that it does not grade how isolated you are. Every study pooled here took a countable thing - how many people are in someone's social network, how often they see or telephone anyone, whether they belong to a club or a congregation, whether they have a partner - and drew a line through it, putting each participant on one side or the other. Some studies used a formal instrument such as a social network index; others used whatever their survey happened to ask. Then they followed everybody for years and counted who died. So the figures on this page compare a group that met a definition of isolated with everybody else. There is no curve here running from "very connected" through "a bit isolated" to "completely alone", of the sort that exists for alcohol or for steps per day, and anyone who offers you one is inventing it. That has a consequence we take seriously: the question at the top of this page is a fair thing to ask yourself, but the numbers below it cannot be attached to the answer. They describe a contrast between groups, and they are quiet on where you personally sit. One further gap is worth naming, because it is the first thing people ask. None of this evidence compares seeing someone in person with telephoning them or messaging them. The reviews measured contact and network size, not the medium, so this page makes no claim that one kind of contact counts and another does not.

Nakou A et al. 2025

Loneliness, social isolation and living alone get used as synonyms in almost every article written about this subject, and they are three different measurements with three different results. The 2025 review this page is built on separated them deliberately, pooling 86 prospective cohort studies of older adults. Social isolation - the countable version - came out at a hazard ratio of 1.35 (95% CI 1.27 to 1.43), from 37 studies contributing 61 estimates and 233,279 participants. Loneliness - the reported feeling - came out at 1.14 (95% CI 1.10 to 1.18), from 52 studies contributing 82 estimates and 308,948 participants. Living alone, a simple fact about a household, came out at 1.21 (95% CI 1.13 to 1.30), from 17 studies contributing 29 estimates and 86,553 participants. Read them together and the shape is clear: the objective measure carries the largest association, the feeling the smallest. This matters because it inverts the usual story. Loneliness is the word that gets the headlines, and it is the one that is easiest to feel ashamed of; but the thing most strongly associated with dying sooner, in this body of work, is the unglamorous business of how many people are actually in your week. There is a caveat in the other direction, and it belongs here rather than in a footnote. When the reviewers split loneliness into emotional loneliness - missing one close confidant - and social loneliness - missing a wider circle - neither half showed a clear association on its own. Only four and three studies respectively could be pooled, which is far too few to conclude anything, but it is in the same table as the headline and you should know it is there. This is also why the page prints no schedule. The advice you will have heard is to see someone once a week; we went looking for where that came from and could not find it. No study in this page's source set tests one frequency of contact against another, none compares meeting in person with a phone call, and none of the mortality reviews measures a dose at all - they compare an isolated group with everybody else. There is a further complication worth being honest about: the largest review of things that actually reduce loneliness, covering 280 studies, found the most effective strategies were psychological rather than logistical - working on how someone reads social situations beat simply creating more opportunities for contact. The reviewers graded their own confidence as low to very low, and none of those studies measured how long anyone lived. And no study has tested naming a day either. It costs nothing and buys nothing; it is the smallest available move in the direction the evidence points, not a claim that it will make you live longer.

Nakou A et al. 2025

Anyone who has watched an illness unfold knows the direction can run the other way. People stop going out because they are unwell, not the reverse; a diagnosis shrinks a social life; grief and frailty and pain all keep people at home. If that explains what these studies see, then the association is real and the interpretation is upside down. This is the single strongest objection to everything on this page and it cannot be settled, because you cannot randomly assign people to be isolated. What can be done is to look at whether the association survives the checks that would weaken it, and here the answer is mixed in an interesting way. When the reviewers compared studies that statistically adjusted for participants' physical health against those that did not, the adjusted studies did not show a smaller association - social isolation came out at 1.37 (95% CI 1.28 to 1.47) among the 49 estimates that adjusted, against 1.24 (95% CI 1.11 to 1.39) among the 12 that did not. The same held for adjustment for mental symptoms. That is not what you would expect if poor health at the outset were generating the whole finding. Pulling the other way: studies with less than five years of follow-up produced larger figures than studies following people for five years or more, for isolation (1.46 versus 1.31) and for loneliness (1.22 versus 1.10) alike. Short follow-up is exactly where illness that was already present but unrecorded would do its damage, so that pattern is consistent with some of the association running backwards. The honest summary is that reverse causation almost certainly inflates these numbers and almost certainly does not account for all of them, and that no observational study can tell you the split.

Nakou A et al. 2025

Solid enough to take seriously, and softer than the headline figures make it look. The strongest thing that can be said is replication. A second review, published in Nature Human Behaviour in 2023 by a different team with a different search, pooled 90 prospective cohort studies covering 2,205,199 adults aged eighteen and over and reported a pooled effect size for all-cause mortality of 1.32 (95% CI 1.26 to 1.39) for social isolation and 1.14 (95% CI 1.08 to 1.20) for loneliness. Those are close to the older-adult figures, from a largely different pile of studies, and they also confirm that this is not only a finding about old age. Now the softness. In the older-adult review the studies disagreed with one another a great deal, and the reviewers report prediction intervals - their estimate of where a future study would land - of 0.94 to 1.94 for social isolation and 0.89 to 1.46 for loneliness. Both include the possibility of no difference at all. They also found statistical signs that studies with unimpressive results are missing from the published record, which would push the pooled figures upward. And the lowest-quality studies of social isolation - the eight estimates scoring six or below on the quality scale, covering 41,479 people - produced by far the largest figure, 1.81, while the higher-quality ones produced 1.31. None of this makes the association go away - the effect survives leaving out any single study, and it survives in a second independent review - but it does mean the right posture is that something real is being measured imprecisely, rather than that a number has been nailed down.

Nakou A et al. 2025 · Wang F et al. 2023

If you are in crisis right now, contact your local emergency number or a crisis line in your country. Nothing on this page is for that moment. This is written for someone whose contact with other people has quietly thinned out without any decision being made about it. It is not written for, and is not adequate for, anyone whose loneliness has become distress. If you feel persistently low or hopeless, if you have stopped wanting things you used to want, if the isolation is not a drift but something you cannot get out of, or if you have had thoughts of harming yourself or of not wanting to be here, that is a matter for a doctor and not for a page about scheduling. It is an ordinary reason to make an appointment. Nothing on this page diagnoses anything and nothing on it is treatment. The research quoted here counted deaths in large groups of people over years; it cannot say anything about any individual, and it was not designed to. The same applies if you are housebound, caring for someone, recently bereaved, or living with a condition that limits how much you can get out: those are real constraints, the advice below is not built for them, and a clinician or a carers' organisation will have more useful things to say than we do.

Lasgaard M et al. 2026

No, and the studies cannot tell the difference, which is a limit rather than a reassurance. What was measured was countable contact - the size of someone's network, how often they saw or spoke to anyone, whether they lived alone - not whether they wanted more of it or were content with what they had. Someone who lives alone by choice, sees three people a week and finds that plenty is scored the same way as someone in identical circumstances who is miserable about it. This is also why nothing here should be read as an instruction about how to live: these are associations observed across large groups over years, and observational research of this kind cannot establish that contact causes anything or that solitude harms anyone. Notably, the one measure closest to plain solitude - simply living alone - produced the weakest evidence of the three on this page, and in this review it did not show a clear association in women at all.

Then this page is the wrong size of answer and it is worth saying so plainly. Loneliness that has become persistent distress - low mood that does not lift, losing interest in things that used to matter, being unable to face people rather than simply not seeing them, and above all any thoughts of harming yourself or of not wanting to be here - is a reason to speak to a doctor. That is an ordinary appointment to make, not an extreme one, and if things are acute right now a crisis line or your local emergency number comes first. Nothing on this page is treatment, nothing on it diagnoses anything, and none of the research quoted here studied whether any behaviour helps depression. What can be said, from a 2026 meta-analytic review of 280 intervention studies, is that loneliness is not a fixed trait: interventions do move it, and the approaches that worked best were psychological ones rather than simply arranging more contact - which is another reason the person to talk to may be a clinician rather than a calendar.