Who Owns the Referral After It Leaves the Clinic

A GP or link worker refers someone to a walking group, a men’s shed, a choir. The referral is logged as complete. Three weeks later, nobody at the community organisation has heard from that person, and nobody at the clinic knows either. This is not a rare failure. It is the default outcome of most health-community partnerships, because almost none of them specify who is accountable for the gap between referral and arrival.

That gap is the actual site of the partnership. Everything either side does before and after it is comparatively easy. The referral itself is a form. The group activity, once someone shows up, is something the community organisation already knows how to run. The problem is the handoff in between, and most partnership agreements simply don’t mention it.

Why “we’ll refer to community groups” is not a plan

Social prescribing has grown into UK health policy on the strength of a real idea: that a large share of what brings people into primary care is social, not medical, and that a clinical response alone will not fix it. The UK’s 2018 loneliness strategy formalised funding for exactly this. But the evidence on social prescribing’s actual effect is thinner than the policy enthusiasm suggests. A 2021 systematic review found consistent increases in self-esteem and confidence among participants, but noted limited trial evidence and wide variation across programmes. A separate review of loneliness-specific social prescribing found that all nine included studies reported positive impacts, and three found reductions in service use — a genuinely useful finding for anyone building a funding case — but nine studies is not a base for firm claims about what works and for whom.

Underneath the aggregate numbers, a 2022 qualitative synthesis found something more specific and more useful: people who benefited from social prescribing describe the benefit as going beyond social contact itself, toward restored participation and purpose. Structured, purposeful activity did more than the fact of contact alone. That has a direct implication for design — a woodworking group with a task in front of it will likely outperform an open drop-in with the same number of chairs — but it says nothing about whether people get there in the first place. The literature is largely silent on that step, because most evaluations start counting from the point of attendance, not from the point of referral.

The National Academies’ 2020 consensus report on isolation in older adults is blunt about where the accountability gap sits: it calls on the health care system to routinely assess isolation and loneliness, but the clinical commentary that followed points out what that would actually require — workflow changes, defined follow-up responsibility, and a receiving side capable of closing the loop. Health systems are good at referral. They are not, by design, good at knowing whether someone showed up to a woodworking group three weeks later. Community organisations are the reverse: they know exactly who showed up, but have no visibility into who was supposed to and didn’t.

The four things a partnership agreement needs to specify

Most “partnership” between a clinic and a community organisation is really a one-way pipe: referrals go out, nothing comes back. A working partnership needs an explicit answer to four questions, and the answer to each should be a name or a role, not a department.

1. Who confirms the person actually made contact with the group — a phone call, not an assumption. This is usually the link worker, if one exists, but it needs to be someone’s job description, not a hope.
2. What happens if the person doesn’t show up to the first session. Is there a second attempt? Whose budget covers the time it takes?
3. What information flows back to the referrer, and in what form. A spreadsheet nobody reads is not a feedback loop.
4. Who is responsible when the group itself can’t take more referrals. Community organisations run on volunteer time and physical space; a men’s shed has a finite number of benches. If the clinic keeps referring past capacity, the relationship degrades from both directions — the organisation starts declining referrals quietly, and the clinic never finds out why uptake dropped.

None of this requires new infrastructure. It requires the two sides to agree, in writing, who does what when the routine case doesn’t go smoothly — because the routine case rarely goes smoothly on the first attempt.

The three-week silence

Call it the three-week silence: the referral is made, the clinic marks it complete, and the community organisation either never hears from the person or hears from them once and never again. Nobody notices, because nobody on either side was assigned to notice. The clinic’s data shows a successful referral. The organisation’s data shows nothing, because it never had the person’s name to check against.

This is not a training problem or a motivation problem. It is a design problem: the referral event and the attendance event are recorded in two different systems that don’t talk to each other, held by two organisations with no shared record and, usually, no shared point of contact below the level of a strategic partnership meeting that happens twice a year. Fixing it doesn’t require integrated IT systems. It requires one named person on each side whose job includes closing that loop, and a shared, simple log — even a spreadsheet with two columns — that both check.

What this means in practice: before signing any partnership agreement, ask each side to name, out loud, the specific person who will phone a non-attending referral back. If neither side can answer without pausing, the partnership has a pipe but no plumbing, and the three-week silence will happen on schedule.

Evidence status: what this partnership model can and can’t claim

Claim	Evidence status
Social prescribing increases self-esteem and confidence in participants	Reasonably supported, across multiple reviews, though study quality varies widely
Structured, purposeful activity produces more benefit than social contact alone	Supported by qualitative synthesis; not yet tested against contact-only groups in controlled trials
Social prescribing reduces GP or emergency service use	Suggestive only — reported in a minority of included studies in one systematic review; not a settled finding
Closing the referral-to-attendance loop improves outcomes	Not directly tested. This is an operational inference from partnership failure patterns, not a cited finding
Community physical infrastructure (libraries, parks, community centres) shapes rates of informal contact	Well supported conceptually, drawing on social infrastructure research, though it is a different claim from the referral-pathway question above

The last row matters because it is easy to conflate the two problems. A social prescribing pathway assumes the destination — the shed, the choir, the walking group — already exists and has capacity. Whether that physical and organisational infrastructure exists at all is a separate, prior question, and one that referral partnerships often take for granted rather than fund.

What this does not solve

A well-designed handoff fixes the accountability gap between referral and arrival. It does nothing about the deeper capacity problem: most community organisations that receive social prescribing referrals are unfunded or underfunded relative to demand, run substantially on volunteer time, and have no mechanism to expand when referral volume increases. A partnership agreement that specifies exactly who calls a non-attender back is still a partnership feeding referrals into a group that may fold in eighteen months because nobody funded the room hire.

It also does not solve reach. Social prescribing, by construction, only reaches people who are already in contact with a health service and whose clinician thought to make the referral. It says nothing about the person who is isolated and never goes to the GP at all — who, by most of the survey evidence on loneliness, is disproportionately likely to be older, living alone, or already disconnected from exactly the kind of social network that would have nudged them toward care in the first place. A tighter handoff is worth building. It is not a substitute for asking who never enters the pipe to begin with.