Who Pays for the Group at the End of the Referral

In 2018, the UK government published the first national loneliness strategy of any country and used it to fund something specific: link workers, embedded in primary care, whose job is to connect patients to non-clinical activity in the community. That decision has shaped how social prescribing has been funded almost everywhere it has spread since. Health systems pay for the referral. Community organisations are expected to supply the destination.

This is the funding decision most commissioners inherit rather than choose, and it deserves more scrutiny than it gets.

The model, and where the money actually goes

Social prescribing works, in its simplest form, like this: a GP, nurse, or other primary care staff member identifies a patient whose problem is partly social — isolation, low mood, an unmanaged chronic condition tangled up with loneliness — and refers them to a link worker. The link worker has time the clinician doesn’t, and knowledge of local activity: walking groups, art classes, men’s sheds, befriending schemes, debt advice. The link worker’s salary sits inside the health or care budget. The walking group, the art class, the shed — those sit inside a community or voluntary sector organisation’s budget, funded however that organisation happens to be funded, which is usually short-term grants, local authority contracts under constant renegotiation, or nothing formal at all.

The UK strategy funded the connector. It did not, as a rule, fund the thing being connected to.

That asymmetry is not a design flaw anyone hid. It is a consequence of where the money originates. Health budgets can justify spending on a role that reduces demand on clinical services. They have a harder time justifying core funding for a community choir, even when that choir is doing the actual work of reducing the loneliness that was driving the GP visits in the first place. The 2020 National Academies report on isolation in older adults, addressed specifically to health systems, calls for routine assessment of isolation and loneliness in clinical settings — a call about identification, not about who funds the response once someone is identified. The clinician-facing commentary that followed makes the same emphasis: what routine assessment would require of care teams, not what it would require of the organisations patients get sent to.

What the evaluations actually show

A 2021 systematic review of social prescribing and wellbeing found consistent increases in self-esteem and self-confidence among participants, but noted limited trial evidence and substantial heterogeneity across programmes — different populations, different activities, different outcome measures, making it hard to say which version of the model is doing the work. A separate 2021 systematic review focused specifically on loneliness found all nine included studies reported positive individual impacts, and three of them reported reductions in GP visits, emergency attendances, social worker contact, or inpatient stays. Three out of nine is not nothing, and it is not a robust evidence base either.

A 2022 qualitative meta-synthesis adds a detail that matters for the funding question specifically: participants described benefit that went beyond social contact itself, toward restored meaningful participation and purpose. Structured, purposeful group activity appeared to do more than unstructured contact. That finding puts weight on the quality and continuity of the community-side offer — the choir needs to still be running, well-run, and able to take new members, not folded six months after the referral pathway that feeds it was announced with fanfare.

None of the published evaluations directly test different funding architectures against each other. There is no study comparing “fund the link worker only” against “fund the link worker plus multi-year core grants to receiving organisations.” That comparison would be worth running. In its absence, the practical read of the existing findings is that community capacity is not a free input health systems can assume will always be there.

The referral cliff

Call it the referral cliff: the point at which a well-funded, well-staffed referral pathway feeds into a community organisation that has no corresponding increase in capacity, and either turns people away, runs a waitlist that erodes the entire point of a quick non-clinical referral, or simply closes the activity when its own funding cycle ends — often unrelated to and unsynchronised with the referral scheme that depends on it.

The failure is invisible from the health system side. The link worker role, the referral numbers, the training — all of that shows up in a commissioner’s dashboard as activity. What doesn’t show up is the community centre that stopped running its Tuesday group because its three-year grant ended in year two of the referral pathway’s life, or the walking group whose volunteer coordinator burned out absorbing referral volume nobody funded them to absorb.

What this means in practice: if you are commissioning or funding a social prescribing pathway, do not fund the link worker role alone and assume the community sector will supply capacity for free. Build a funding line — even a modest one, multi-year rather than annual — for the organisations receiving referrals, tied to the volume the pathway is expected to generate, and set it up before the pathway goes live rather than as a retrofit once organisations start turning people away.

Evidence status

Claim	Evidence status
Social prescribing improves self-esteem and confidence for participants	Supported by systematic review, but trial evidence is limited and programmes are heterogeneous
Social prescribing reduces GP, ER, or inpatient use	Reported in a minority of included studies (3 of 9); not a general finding
Structured, purposeful activity produces more benefit than contact alone	Supported by qualitative synthesis; not tested against funding structure
Funding the referral pathway without funding receiving capacity undermines outcomes	Plausible from the funding architecture and the qualitative literature on sustained participation; not directly tested in any published evaluation
Health systems should routinely assess isolation in clinical settings	Explicit recommendation of the National Academies consensus report

What this does not solve

This approach addresses funding architecture, not reach. Social prescribing pathways depend on a patient encountering a clinician who thinks to ask, and on a link worker having capacity to follow up — both of which favour people already in regular contact with primary care. People who avoid the health system, who lack a GP registration, or who are isolated in ways that never surface in a clinical encounter are not touched by any funding fix to the pathway itself. Fixing the referral cliff makes the pathway more durable for the people already inside it. It does not widen who gets referred in the first place.