What a Volunteering Trial Actually Found About Loneliness

Most advice about loneliness rests on studies that can’t actually prove the advice works. A survey finds that people who volunteer are less lonely, and the conclusion gets written up as “volunteering reduces loneliness.” But that’s a correlation. Maybe less-lonely people are simply more likely to sign up to volunteer in the first place. The study can’t tell you which way the arrow points.

This is why the HEAL-HOA trial, published in The Lancet Healthy Longevity in November 2024, is worth pausing on. It’s a randomised controlled trial — the kind of study where researchers actually assign people to different conditions and compare outcomes, rather than just observing who does what. That design is rare in loneliness research. The National Academies noted in its 2020 report on isolation in older adults that the field is short on exactly this kind of evidence, and mostly relies on small programme evaluations that can’t separate cause from coincidence.

What the trial actually did

The researchers recruited lonely older adults in Hong Kong and randomly split them into groups. One set was supported into volunteering and other prosocial engagement. The other served as a control, without that structured push. Because assignment was random, any difference in loneliness afterward is much harder to explain away as “the people who joined in were just the more social ones to begin with.”

This matters for a simple reason: it’s the difference between “lonely people who volunteer tend to feel better” and “putting a lonely person into a volunteering role tends to make them feel better.” The second claim is the one that’s actually useful if you’re deciding what to do with your Tuesday evening.

Why “volunteering” is doing more work than it sounds like

The word volunteering can sound like a moral instruction — go and be useful, and your loneliness is a side issue. That’s not really what’s happening here, and it’s not why the trial is interesting.

What volunteering actually supplies is a role. You show up at a fixed time, to a fixed place, and you have a job to do that isn’t “make conversation with strangers.” The task carries the interaction. This lines up with something researchers looking at social prescribing — the practice of a doctor or link worker connecting someone to a community activity instead of, or alongside, medication — have found repeatedly: a 2022 qualitative synthesis of social prescribing programmes reported that the benefit people describe isn’t just having more contact with people. It’s having a reason to be there, and a sense of purpose that survives the walk home. Structured, purposeful activity outperforms unstructured contact.

That tracks with a point Eric Klinenberg makes in Palaces for the People: the shared physical spaces that generate social contact — libraries, parks, community centres — work partly because they give people a reason to occupy the same room repeatedly, not because they’re designed for mingling. A volunteering shift is a stripped-down version of the same mechanism. You don’t need to be interesting. You need to turn up and sort the donations, or read to the kids, or hand out the soup.

The low bar this actually sets

If you’ve been told to “get more involved,” it can sound like a demand for a whole new identity — the volunteering type, the joiner. That’s not what the evidence points to. The HEAL-HOA trial tested engagement that was arranged and supported, not self-generated from a standing start of motivation. Someone helped set it up. The role was made available; the person didn’t have to invent it.

That’s a smaller ask than it first appears. It means the useful move isn’t “become a more civic-minded person.” It’s “find one recurring shift, at one place, on one day of the week, that someone else has already structured for you.” A food bank with a Thursday afternoon slot. A library reading programme with a standing weekly session. A charity shop till that needs covering on Saturdays. The organisation exists already; you’re filling a gap in it, not creating a new social life from raw material.

Why this beats trying to make friends directly

There’s a reason volunteering shows up as a good route into connection when direct attempts at friend-making often stall. When you go somewhere specifically to meet people, every interaction carries the weight of being evaluated as a potential friendship. It’s effortful in a way that’s hard to sustain, and it’s also a little exposing — you’re visibly there to be liked.

Volunteering removes that framing. Nobody at a Saturday shift is there to make friends; they’re there to sort donations. The social contact happens as a byproduct of doing the task next to the same people, week after week. That repetition, not the volunteering itself, is probably the load-bearing part. Seeing the same handful of people on a fixed schedule, month after month, is what turns “people I recognise” into “people I know.” A single volunteering day does little. Standing weekly attendance does a lot more, simply by accumulating.

What the trial doesn’t tell you

Worth being honest about the limits. The trial was conducted in Hong Kong with older adults who were already identified as lonely — it doesn’t automatically generalise to a 28-year-old who’s just moved cities, or a new parent who can barely leave the house on a fixed schedule. The American Heart Association’s 2022 scientific statement on isolation and cardiovascular risk flagged this as the field’s central gap: there’s decent evidence connecting isolation to health harm, but comparatively little trial evidence on which interventions reliably fix it, for whom. HEAL-HOA is one data point in a mostly empty column, not a settled answer for every life stage.

It’s also a support structure, not just a task. The trial paired people with organised volunteering opportunities and helped them get placed. Simply telling yourself to “go volunteer somewhere” and hunting for an opening on your own reintroduces some of the friction the trial design removed. The value of a good social-prescribing model, per the 2022 research, is partly the person or system that does the matching. If you’re doing this without a link worker or a programme, that matching work falls on you — which is worth naming honestly, because it’s the part most likely to make the plan quietly not happen.

What to actually do with this, this week

If you want to test this without waiting for a doctor’s referral or a formal programme, the practical version looks like this: pick a fixed day and time, not “when I have a spare evening.” Find one organisation with a standing need — most food banks, animal shelters, museums, and hospital volunteer offices post weekly or biweekly slots rather than one-off events. Commit to the recurring slot rather than a single session. A single afternoon of volunteering will not move the needle on loneliness; a Thursday afternoon, held for eight weeks, plausibly will, based on what the repetition research keeps showing across settings.

Treat the first few sessions as pure logistics, not social auditions. You are learning where things go and what the job is. The social part accumulates on its own, later, once you’re not thinking about the task. This is the opposite of what most advice implies — that the friend-making is the point and the task is incidental. In the version the evidence actually supports, it runs the other way around.

None of this requires becoming a different kind of person. It requires finding one place that already has a schedule, and putting yourself on it.