Why Aged-Care Befriending Models Don't Transfer Cleanly to the Workplace

An employer with a hybrid workforce wants to do something about isolation among remote staff. Someone on the wellbeing team finds a randomised trial showing that befriending reduces loneliness scores in older adults in residential care, and proposes a version for work: pair people up, have them check in regularly, call it a buddy scheme. It sounds reasonable. It is also importing an intervention built for a population that has almost nothing in common with the one it is being applied to.

This is worth stopping and checking before you brief HR on a rollout.

The evidence being borrowed

The most solid recent evidence on befriending comes from a randomised controlled trial in aged care, which found that structured befriending reduced UCLA Loneliness Scale scores by 2.39 points at eight weeks and 2.71 points at sixteen weeks compared with a control group. That is a real effect, not a null result dressed up. It is also a trial run on people in residential aged care: old, often physically frail, living in an institutional setting, with a narrow existing social world.

A separate and more recent trial, the HEAL-HOA behavioural activation study, tested something more structured against befriending itself, in a different but equally specific population — 1,151 older adults living in poverty, alone, and digitally excluded. The comparison group got standard befriending. The intervention group got eight 30-minute telephone sessions of behavioural activation and mindfulness, delivered by trained laypeople who were themselves older adults with lived experience of loneliness. At twelve months, the structured approach beat befriending. An earlier trial in the same research programme had tested prosocial engagement and volunteering against a control, again in lonely older adults, again in a tightly defined population.

Three separate rigorous trials, then, all built around the same population: old, isolated, often low-income, often digitally excluded, often living alone. That specificity is exactly what makes the results useful — and exactly what should make anyone pause before lifting the intervention wholesale into a workplace.

Why the workplace population is a different problem

Loneliness among employees is real and well documented. Gallup’s 2024 workplace data found one in five employees worldwide reporting loneliness on the previous day, with fully remote workers at 25% against 16% for fully on-site staff. Cigna’s 2025 survey put loneliness among workers at 52%, and its 2020 loneliness index found lonely workers miss twice as much work due to illness and five times as much due to stress. The scale of the problem is not in question.

But the mechanism is different. A study of healthcare workers under remote arrangements found that workplace isolation and loneliness are separate constructs with different correlates — someone can feel disconnected from colleagues while having an intact social life outside work, or vice versa. More recent nationally representative work on remote employees reinforces that remote work and loneliness are linked, but through pathways tied to structure, visibility, and incidental contact at work, not through the kind of comprehensive social deprivation that aged-care befriending trials were built to address. Research published in Science modelling the shift to remote work found it explains roughly a third of the rise in isolation and mental distress between 2011–2019 and 2022–2024, and that the effect is driven overwhelmingly by people who live alone losing incidental contact — a much narrower and more specific mechanism than the multi-domain isolation seen in institutionalised older adults.

In other words: the older-adult trials were treating people whose entire social world had contracted. The workplace problem is usually a contact deficit in one domain — the working day — layered onto a social life that may otherwise be intact. Those are not the same disease, even if both get labelled “loneliness.”

What follows for programme design

The practical question is not “does befriending work” but “what is the deficit this population actually has.” Structured, purposeful contact tends to outperform unstructured contact across the literature — a qualitative synthesis of social prescribing found that participants describe benefit coming from restored meaningful participation and purpose, not contact for its own account. That maps onto the HEAL-HOA finding: the structured, skills-based intervention beat plain befriending head-to-head.

For a workplace, that suggests a workplace peer scheme should be built around a task, project, or shared activity with a defined purpose, rather than an open-ended “check in with your buddy” instruction with no content. An unstructured buddy scheme asks two busy employees to manufacture warmth from nothing on a fixed schedule, which is a much harder ask than it sounds, and closer to what befriending alone failed to beat in the older-adult trial than to what worked.

Claim	Evidence status
Structured befriending reduces loneliness in older, isolated adults	Solid — replicated across at least two RCTs in that population
Structured behavioural activation outperforms befriending in older adults	Solid — one head-to-head RCT, large sample
Workplace isolation is distinct from general loneliness	Established as a construct, not yet extensively intervention-tested
Remote work loneliness responds to peer-buddy schemes modelled on aged-care befriending	Untested — no equivalent RCT exists for this population
Structured, purpose-built group activity beats unstructured contact generally	Reasonably consistent across social prescribing literature, though trial evidence overall remains thin

What this means in practice: before adapting an intervention from one population to another, name the actual deficit you are addressing. If your remote staff have an intact social life outside work but no informal contact during the working day, the fix is structural — shared work, visible collaboration, reasons to be in the same room or call — not a scripted friendship assignment borrowed from a trial built for people with no other social contact at all.

The borrowed buddy problem

Call this the borrowed buddy problem: taking an intervention that was validated for its ability to fix a specific, severe deficit in one population, and deploying it against a milder, differently-shaped deficit in another population, on the assumption that the label — “befriending,” “buddy scheme,” “peer support” — carries the evidence with it. The label does not. The trial evidence sits with the population it was tested on. None of the RCTs cited here were run on employed adults, remote or otherwise, and no equivalent trial currently exists for a workplace buddy scheme.

That does not mean peer schemes at work are worthless. It means they should be evaluated on their own terms, with their own outcome measures — workplace isolation scales, not the UCLA Loneliness Scale imported wholesale — and built around structure and purpose rather than an assumption that unstructured pairing will do for a hybrid worker what it partially failed to do for a homebound one.

What this does not solve

None of this tells you whether a given workplace peer scheme will work, because the trial that would answer that question has not been run. It also does not address the underlying driver identified in the remote-work research: contact lost to the shift away from shared physical space. A peer scheme can supplement that loss; it cannot substitute for redesigning when and how people are actually in the same room.