The lowest point of loneliness in an adult life tends to fall in middle age. It rises again afterwards, and by the oldest ages it can sit higher than it did in a person’s twenties. That is the pattern set out in a coordinated analysis of nine long-running studies, published in Psychological Science in 2024 by Eileen Graham and colleagues, drawing on more than 128,000 people aged 13 to 103 across over twenty countries. You can read the paper in Psychological Science.

The version of this that travels is shorter: loneliness peaks after 75. That is close to the truth, but not the whole of it, and the distance between the two is where the useful reading sits.

We are writers, not clinicians. What follows is a reading of the research, not medical or psychological advice.

What the analysis actually measured

The Graham team did not run a new survey. They pooled and harmonised data from nine existing longitudinal studies, among them the English Longitudinal Study of Ageing, the German Socio-Economic Panel, the Health and Retirement Study in the United States, the Household, Income and Labour Dynamics in Australia survey, and the Survey of Health, Ageing and Retirement in Europe. The point of pulling them together was to see whether loneliness moves in a consistent shape across very different national settings.

It does.

Across the datasets, loneliness followed a U-shaped curve: relatively high in young adulthood, falling towards a midlife low, then climbing through older age. Only one of the nine studies was based in the United States, which the authors read as a sign the pattern is not simply an American story.

Two things are worth holding in view before the pattern hardens into a rule. All nine studies were conducted before the COVID-19 pandemic, and the underlying loneliness scales differed from one study to the next, which the authors list as a limitation. This is a strong, carefully assembled piece of work. The finding is worth taking seriously, but it should not be read as the final word.

Why “after 75” is the wrong number to fixate on

The specific age comes from older research, not from the 2024 analysis. Louise Hawkley and colleagues, writing in the journal Psychology and Aging, examined US adults over 50 and found that loneliness tended to fall after 50 until roughly 75, after which it began to rise. That is the origin of the figure that now gets attached to the broader cross-country finding.

The part that usually gets dropped is the reason. In that analysis, the rise after 75 was largely accounted for by two things: declining health and the loss of a spouse or partner. When the researchers adjusted for those factors, loneliness kept falling into the oldest ages rather than climbing. A summary of this appears in Our World in Data’s overview of the loneliness research.

So the number describes a clustering of events, not a property of the birthday. People do not become lonely because a calendar turns. Widowhood, reduced mobility, hearing loss, the thinning of a social circle through death rather than choice: these arrive more often in the late seventies, and they are what the curve is tracking. What looks like an effect of age is closer to a tally of the losses that tend to land around it.

Where the picture is contested

Not every study finds the same late-life spike. A 2021 systematic review and meta-analysis in PLOS One, pooling 39 studies covering around 120,000 older people across 29 countries, put the prevalence of loneliness in later life at roughly 28.5 per cent and reported no clear evidence that loneliness was more common in people over 75 than in those aged 65 to 75. That review is available in PLOS One.

Some of the disagreement is a measurement problem. Loneliness gets captured by different instruments, from the multi-item UCLA scale to single blunt questions about feeling left out, and the wording shapes the answer. Some of it is cohort effects. Separate work by Bianca Suanet and others has suggested that today’s older adults report less loneliness than older adults of earlier generations, which complicates any tidy story about ageing and isolation moving in lockstep.

There is also the matter of what most people assume. The common expectation is that loneliness belongs to the young, or to the pressured middle of life when work and caring for others crowd out friendship. The cross-country data pushes back on the second half of that. Midlife, on this evidence, is where loneliness is often at its lowest, not its highest.

What this leaves us with

The durable takeaway is narrower than the headline and more useful for it. Loneliness is not a fixed trait that some people carry and others escape. It moves across a life, and the direction it moves in later years is shaped heavily by circumstances that are, in part, addressable: bereavement support, help staying mobile, treatment for the hearing loss that quietly shrinks conversation, transport that keeps someone in contact with other people.

That reframing matters more than the exact inflection point. If the late-life rise were simply what age does to a person, there would be little to do about it. Because it tracks specific losses, there is.

If loneliness in later life becomes persistent and heavy rather than an ordinary reaction to a hard stretch, a GP is a reasonable first person to raise it with, and often a route to practical support. That is a low bar to clear.

What should stay with a reader is less the number than what sits behind it. The loneliest stretch of many lives is not the one people brace for in youth, and when it arrives later, it usually arrives attached to something with a name.