Who Holds the Relationship When a GP Refers to a Community Group

A GP or link worker can refer as many patients as their caseload demands. A walking group, a men’s shed, or a community choir can absorb roughly the number of people its one or two volunteer coordinators can manage before the group stops being the thing that worked. Nobody designs the mismatch on purpose. It happens because the referring side and the delivery side are usually different organisations, funded by different bodies, answering to different targets, and the conversation between them stops at “yes, we’ll take referrals.”

This is the operational question that decides whether a social prescribing pathway helps anyone past month three: who is accountable for keeping referral volume inside what the community organisation can actually deliver, and who pays for the difference when volume grows.

The referral cliff

Give it a name, because programme teams keep rediscovering it as if it were new: the referral cliff. A pathway launches, a health system markets it well, referrals ramp faster than the group’s capacity, the group’s quality drops as coordinators try to onboard everyone, existing members drift off because the thing that made the group work — a small number of people who know each other — is gone, and referral numbers start looking fine on a dashboard while the actual experience on the ground has collapsed. The health side sees a successful pathway. The community side is running an entirely different, worse programme than the one it agreed to.

The UK’s 2018 national loneliness strategy pushed social prescribing into the mainstream of health commissioning, which means most readers of this piece are now operating in a system built around exactly this handoff: clinical or administrative referral into non-clinical community activity. The strategy assumed the community-sector capacity would be there to receive people. It did not fund it as a distinct, ongoing line — and that gap is where the referral cliff lives.

What the evidence actually supports

Before building a partnership structure, it’s worth being precise about what social prescribing has been shown to do, because the case you make to a commissioner should not outrun the literature.

Claim	Evidence status
Social prescribing improves self-esteem and self-confidence for many participants	Reasonably supported — a 2021 systematic review reports this as a consistent outcome, though trial quality is mixed
Social prescribing reduces use of GP, emergency, or inpatient services	Weakly supported — a 2021 systematic review found three of nine included studies reported this; not a reliable basis for a cost-avoidance pitch to a funder
Structured, purposeful activity works better than unstructured social contact	Supported by qualitative synthesis — participants describe the value as restored participation and purpose, not contact for its own sake
Older adults’ isolation is a legitimate clinical concern, not a lifestyle issue	Supported — the National Academies’ 2020 consensus report calls for routine assessment of isolation in health care settings, and clinical commentary on that report argues for building it into practice
Social infrastructure (libraries, parks, shared physical space) shapes contact rates independent of programming	Supported at a conceptual level, drawn from Eric Klinenberg’s account of social infrastructure, though it speaks to place rather than to referral pathways specifically

The pattern across the social prescribing reviews is consistent: individual-level benefit is real and reasonably well attested; system-level claims about reduced service use are thin. Build your case to funders on the first, not the second.

A structure that survives contact with real referral volume

1. Set a capacity ceiling before the pathway opens, not after it strains. Ask every partner community organisation what number of new referrals per month it can absorb without losing the quality of what already exists. Write that number into the partnership agreement. Treat it the way a clinical service treats caseload limits — a hard constraint, not an aspiration.

2. Route referrals through a link worker who tracks capacity, not just need. The link worker’s job is not finished at the point of referral. It includes knowing, in real time, which partner organisations are near their ceiling and diverting new referrals elsewhere until capacity frees up. This requires the community organisation to report capacity regularly — weekly is workable for most groups — and requires the health side to actually use that information rather than treating referral as a one-way action.

3. Fund the receiving capacity, not just the referral activity. Money that pays for link workers to make referrals but not for the community organisation to onboard, run, and sustain the group they’re referred into is funding half a pathway. If the commissioning body cannot fund community-side capacity directly, it should at minimum fund a coordinator role inside the community organisation whose job is managing referral flow — the qualitative evidence on social prescribing repeatedly points to the quality of the receiving activity as what determines whether people report benefit, not the act of referral itself.

4. Agree what happens when a group is full. A waiting list with no communication is where people disengage. Decide in advance whether the health side holds the waiting list, offers an interim alternative, or whether the community organisation opens a second cohort — and cost that decision before you need it.

5. Review the partnership against the ceiling, not against referral count. A dashboard that only shows referrals made looks good right up until the point it’s meaningless. Add a second number: proportion of referred people still attending after eight weeks. That number tells you whether the group absorbed them or whether it cracked under load.

What this means in practice: before a health service signs up a community organisation to a referral pathway, get a number from them — how many new people per month they can take without the group degrading — and write it into the agreement as a ceiling, not a target. Fund whatever it costs to manage that ceiling. A pathway that refers steadily within a known capacity will outperform one that refers enthusiastically into a group that quietly stops working.

What this does not solve

None of this fixes the underlying scarcity: most areas do not have enough community groups, clubs, and third places to absorb the volume of isolated people that health systems could plausibly identify if they screened for it systematically, as the National Academies’ 2020 report recommends. A well-structured referral pathway manages the capacity that exists. It does not create capacity that doesn’t. And it still only reaches people who make it into a GP’s office or a link worker’s caseload in the first place — the much larger group who never present to a health service at all remains outside the scope of anything this piece has described.