Where Social Prescribing Groups Actually Lose People

The instinct in most programme designs is to protect the first session. Get someone through the door once and the theory writes itself: comfort builds, habit forms, attendance holds. The published record on social prescribing and community group models suggests this instinct is aimed at the wrong point in the timeline. First sessions are usually fine — referral momentum, curiosity, or a link worker’s encouragement carries people that far. The collapse happens later, at a point that different programme models handle very differently: the moment the original reason for showing up runs out.

Call it the handover point. A person is referred to a walking group, a craft session, a befriending scheme, or a volunteering programme for a reason — a GP conversation, a bereavement, a discharge plan. That reason has a shelf life. Somewhere around the middle of a structured course, the task that justified attending is exhausted. What happens next depends on whether anything has grown underneath it: a routine, a set of relationships, a reason to return that has nothing to do with the referral. Programmes that never build that second layer lose people right at the handover. Programmes built around it don’t.

Three models, three answers to the handover problem

The case-approach worth looking at here is not one organisation but three structurally different classes of response, all visible in how they’re described in the literature.

Time-boxed structured programmes end before the handover arrives. The HEAL-HOA trial in Hong Kong tested a defined course of prosocial engagement and volunteering against a control group of lonely older adults, and a companion strand in the aged-care literature tested befriending against control and found UCLA Loneliness Scale reductions of 2.39 points at eight weeks and 2.71 points at sixteen weeks. Neither design asks the question retention forces on open-ended groups, because the programme is engineered to deliver its effect and close before the purpose that recruited people has faded. The trade-off is durability: a fixed course produces a measured result at week eight or sixteen, but says nothing about whether anyone kept meeting after the researchers stopped counting.

Open-ended interest groups face the handover problem head-on, with mixed success. Meetup’s model — built on recurring, unlimited, member-organised groups — depends entirely on a self-renewing reason to return that has nothing to do with why someone joined originally. Timeleft solves a narrower version of the same problem by removing the organiser dependency altogether: dinners are weekly, seats are confirmed once three people join, and there’s no messaging beforehand to negotiate around. Both models are, in effect, an answer to the handover point rather than a workaround of it — they try to convert an initial reason for attending into an ongoing one before the initial reason expires.

Social prescribing sits in between, and the evidence shows it. A qualitative meta-synthesis of how people experience social prescribing found something specific: benefit came less from contact itself and more from restored meaningful participation and purpose, and structured, purposeful group activity appeared to outperform contact alone. That is a direct statement about the handover point — the activity that keeps working is the one that keeps generating purpose, not the one that merely keeps people in a room together. But a systematic review of social prescribing’s effect on loneliness found all nine included studies reported positive individual impacts while a broader review of wellbeing outcomes flagged limited trial evidence and heavy heterogeneity across programmes. A 2025 protocol reviewing social prescribing specifically for older adults noted that only one peer-reviewed randomised controlled trial exists in the entire field, despite growing adoption. The enthusiasm is ahead of the proof.

What the evidence actually supports

Claim	Evidence status
Structured, purposeful activity retains people better than unstructured contact	Reasonably supported — qualitative meta-synthesis of social prescribing experience
Social prescribing reduces loneliness for the people who complete it	Supported for those who complete it — systematic review found positive impact in all nine included studies
Social prescribing is effective for older adults generally	Not established — only one RCT exists in this specific population
Fixed-length interventions (befriending, volunteering) produce measurable loneliness reduction within weeks	Supported by randomised trials, but silent on what happens after the trial ends
Open-ended interest groups sustain attendance through habit rather than purpose	Plausible from model design, not tested against a control

The third place sidesteps the handover point entirely

There’s a fourth answer worth naming, because it works by refusing the premise. Ray Oldenburg’s account of the third place — the café, the barbershop, the bar with regulars — describes places that were never structured around a task in the first place. Nobody attends a third place to complete a referral goal. Eric Klinenberg’s account of social infrastructure extends the same logic to libraries and parks: physical spaces that shape contact simply by being routinely available, with no programme logic and no handover point to fall off. The retention question doesn’t apply because there was never a defined reason to expire.

The catch is that third places are disappearing, and not evenly. Research tracking twelve categories of third place across the US found closures in every category between 2019 and 2021, concentrated in census tracts with higher social vulnerability and in rural areas — a pattern a University of Colorado Boulder summary connects directly to loneliness risk. The model that solves retention most cleanly is also the one losing physical ground fastest, and losing it fastest exactly where structured programmes are most needed to fill the gap.

What this means in practice: Don’t measure retention from session one. Measure it from the session at which the referral reason would plausibly have been satisfied — usually somewhere in the middle third of a structured course — and build something else in before you get there: a rotating task, a peer-led element, a reason to return that outlives the reason to have started. If your programme has no answer to what happens after the handover point, expect the drop-off to show up exactly there, not at the start.

What this does not solve

None of this tells you how to reach the people who never get referred, never walk into a library, and never see a Meetup listing. Every model discussed here — structured trial, open group, third place — depends on someone already being inside a system that can route them there: a GP conversation, an internet search, a physical space that still exists in their neighbourhood. The third-place closure data is a reminder that where you live increasingly determines whether any of these retention strategies are even available to try.