A Referral Is Not a Relationship

A link worker refers someone to a walking group. The person attends once, is thanked for coming, and does not come back. The referral gets logged as a successful outcome. Nobody checks whether the person is any less isolated three months later.

This is the gap between a service and a community, and social prescribing pathways routinely collapse it without noticing. If your commissioning documents describe “connecting people to community assets,” ask a harder question first: are you connecting people to a place that will still recognise them in six months, or to a single transaction that happens to occur in a room with other people?

What actually distinguishes the two

A service delivers a defined activity to a person and ends. A GP appointment, a benefits check, a one-off cookery class: these are services. They can be excellent services. They are not, by themselves, community.

A community is a standing group with continuity — people who show up expecting to see some of the same faces, where membership accrues (you become a regular, not just an attendee) and where the relationship outlasts any single visit. Ray Oldenburg’s concept of the third place captures this: a pub, café, or barbershop works as social infrastructure not because of what happens on any one visit but because of the accumulated, unscheduled contact over years. Eric Klinenberg’s argument about libraries and parks makes the same point about physical spaces — their value is in repeated, low-stakes encounter, not in any single programmed event.

Social prescribing referrals are usually built to deliver the first kind and are then evaluated as though they produced the second. The systematic review of social prescribing and wellbeing found consistent gains in self-esteem and confidence, but also flagged how thin and heterogeneous the underlying trial evidence is — a sign that “did the person attend” is being counted as success far more often than “did the person’s isolation actually change.” A separate systematic review of social prescribing and loneliness found that all nine included studies reported positive individual impacts, and three reported reduced use of GP, emergency, social worker, or inpatient services. That is a real signal. But look at what was being measured: attendance and short-term self-report, not whether people built a durable social network they’d still have access to a year later.

Why this matters for what you build

The AARP Foundation’s national survey of adults 45 and older is useful here because it did not just ask whether people felt lonely — it asked what predicted it. The strongest predictors were the size and diversity of a person’s social network and physical isolation, not a single missing activity. Among people who had spoken to their neighbours, 33% were lonely; among those who never had, 61% were. That is a network effect, accumulated through repeated low-stakes contact, not something a one-off session produces.

Robert Putnam’s distinction between bonding and bridging social capital sharpens this further. A service can occasionally produce a bridging contact — a new acquaintance, a new piece of information — but it rarely produces bonding, the thicker, reciprocal tie that people actually mean when they say they feel less alone. Bonding takes repetition. A referral pathway that ends at the first session has structurally ruled out the thing most likely to reduce isolation.

The qualitative meta-synthesis of what people say they get from social prescribing backs this up from the participant side: people describe benefit that goes beyond social contact itself, toward restored meaningful participation and purpose. That describes belonging to something, not attending something. The same review notes that structured, purposeful group activity appears to work better than contact alone — which is a case for building groups people can join and stay in, not for maximising the number of referrals made.

The one-and-done failure mode

Call this the one-and-done problem: a referral is completed, the person attends once, the outcome is recorded as positive, and no one checks what happens next. It is the social prescribing equivalent of measuring a gym’s success by counting people who walk through the door on their first visit.

One-and-done is attractive to commission because it is easy to count. Referrals made, sessions attended, satisfaction scores collected on day one — all of these are cheap to capture and look good in a quarterly report. Retention at twelve weeks, whether someone has made a reciprocal tie, whether they’d notice if the group stopped meeting — these are harder to measure and routinely go unmeasured. The gap between what’s easy to count and what actually indicates community is where a good pathway quietly turns into a bad one.

A quick diagnostic

Run your current pathway through these five questions. If you answer no to more than one, you are likely running a service dressed up as a community-building intervention.

1. Does the activity exist independent of any single referral? A community group that would keep meeting whether or not your service referred anyone into it this month is a real asset. A group convened specifically to receive referrals, with no life outside that pipeline, is a service wearing a community’s clothing.
2. Is there a mechanism for a second, third, and tenth visit — not just the first? Check whether anyone follows up after week one, or whether the referral is marked complete on arrival.
3. Would a regular notice if a specific person stopped coming? If the honest answer is no, the group has not yet formed enough continuity to function as social infrastructure in Oldenburg’s sense.
4. Is anything measured after the first session? If your only data point is attendance on day one, you are measuring the service, not the community outcome you are claiming to produce.
5. Could the person plausibly attend without a referral at all? Open, walk-in-able groups — the kind Klinenberg describes forming around libraries and parks — build community precisely because they don’t depend on a gatekeeping referral. A pathway that only exists through professional referral has a structural ceiling on how communal it can become.

Claim	Evidence status
Social prescribing referrals produce short-term gains in self-esteem and confidence	Reasonably supported, though trial evidence is thin and heterogeneous
Social prescribing reduces use of GP, emergency, or inpatient services	Supported in a minority of included studies; not the default finding
Network size and diversity, not single activities, predict lower loneliness	Well supported by survey data
Structured, purposeful group participation outperforms unstructured contact	Supported by qualitative synthesis; not yet by controlled trials
One-off attendance at a referred activity durably reduces isolation	Not supported — most designs don’t measure past the first visit

What this means in practice: before commissioning or renewing a social prescribing pathway, separate two budget lines — one for the referral and assessment process, and one for sustaining the standing groups people are referred into. If the second line doesn’t exist, or exists only as “community assets we point to,” you are funding a directory, not an intervention. The pathway will keep looking successful on attendance data while doing almost nothing to the underlying isolation it was commissioned to address.

Building toward continuity, not just access

None of this argues against referral pathways. Getting someone from a GP’s office to a room full of other people is a real and non-trivial piece of infrastructure, and the reduction in service use found in some studies is worth having on its own. The point is narrower: don’t let the ease of measuring referrals substitute for measuring whether anyone stayed. Fund the follow-up call at week four, not just the warm handoff at week one. Ask community groups whether they have capacity for regulars, not just capacity for one-off visitors, before routing people into them at volume. And when a funder asks for outcomes, resist reporting attendance as though it were belonging — the qualitative and survey evidence above says fairly clearly that those are different things, achieved by different design choices.

What this does not solve

This framing does not tell you how to fund the standing groups themselves, which is a harder and more chronic commissioning problem than funding a referral pathway — third places rarely fit neatly into a grant cycle. It also does nothing for the people who never reach a link worker in the first place: social prescribing, however well designed, reaches people who are already inside a system that can refer them. The isolated person who avoids their GP, or who has no link worker to encounter, is untouched by any distinction drawn here between a good pathway and a bad one.