Social Prescribing's First Four Weeks: What the Sequencing Looks Like

A GP or nurse refers someone into social prescribing. A link worker picks up the referral. Somewhere between that referral and the person actually turning up to a walking group, a choir, or a repair café, a lot of people disappear. The published evaluations of social prescribing schemes don’t agree on much, but they agree on this: the gap between referral and first attendance is where the model succeeds or fails, and most services have converged on a similar four-week sequence to close it.

This is not a single named programme’s approach — it’s a pattern visible across the systematic reviews of how UK-style social prescribing has been implemented since the 2018 national loneliness strategy pushed funding and standardisation into the field. Worth being explicit that it is a pattern, not a protocol, because the evidence underneath it is thinner than the enthusiasm around it.

The sequence, as services have run it

Week one: the long conversation, not the referral form. Systematic reviews of social prescribing describe the initial link worker contact as deliberately unstructured — closer to a motivational interview than an intake assessment. The point is to find out what the person actually wants, which is frequently not what the referring clinician assumed. Someone referred for “loneliness” may turn out to want debt advice, or vice versa. Reviews of the loneliness-specific evidence note that services doing this well treat week one as diagnostic in the social sense, not just the clinical one.

Week two: goal-setting with the person, and a short list, not a single option. The AARP Foundation’s 2018 national survey of adults 45 and older found that the biggest predictors of loneliness were the size and diversity of a person’s social network and physical isolation — not mood, not diagnosis. That argues against week two being “here is your one referral to the walking group.” Services that build a shortlist of two or three options, matched to what came out of the week-one conversation, are working with the grain of that finding rather than against it.

Week three: the warm handover. This is the step most often skipped and most often cited as the reason people never attend. A warm handover means the link worker does more than hand over a phone number — a phone call ahead of time to the group organiser, an offer to attend the first session together, a text reminder close to the date. The systematic review of social prescribing’s impact on loneliness found that all nine included studies reported positive individual outcomes, but the studies that described any mechanism at all tended to describe this kind of active bridging rather than passive referral.

Week four: check that attendance actually happened, and do something if it didn’t. This is the step that turns a referral list into an outcome. Reviews of social prescribing’s broader wellbeing effects report self-esteem and confidence gains as consistent outcomes — but consistently among people who attended. The link worker’s job in week four is not to close the case. It’s to find out whether the handover in week three worked, and if it didn’t, why.

The week four problem

Call it that because it is where the sequence collapses even when the first three weeks were done properly. A person is warmly handed over to a group, they don’t show up, and nobody follows up to ask why — because by week four the link worker has moved to the next referral in the queue, and the case looks closed on paper. The systematic reviews are candid that this is exactly the point where the evidence thins out: several of the included studies had no attendance data at all, only referral counts. A service that reports “200 referrals this quarter” and cannot tell you how many of those people attended anything is describing a mailing list, not an intervention.

This matters because the theoretical case for social prescribing rests on people actually connecting with something — a group, a place, another person — not on being told to. Eric Klinenberg’s account of social infrastructure argues that libraries, parks, and other shared spaces shape contact rates because people physically pass through them repeatedly, not because someone once suggested they might. A referral that ends at the door is not social infrastructure. It’s a suggestion.

Evidence status

Claim	Evidence status
Warm handovers (introduction, accompaniment, reminder) improve attendance	Plausible mechanism, described in reviewed studies but not isolated or tested against referral alone
Social prescribing improves self-esteem and confidence for those who engage	Reported consistently across included studies in the systematic reviews
Social prescribing reduces GP, A&E, or inpatient service use	Reported in three of nine studies in the loneliness-specific review; not the majority
Network size and diversity, not mood, predict loneliness in midlife and older adults	Supported by AARP’s 2018 national survey using the UCLA Loneliness Scale
A four-week check-in point improves long-term retention	No direct trial evidence; inferred from where existing services report drop-off

What programme managers should actually build

What this means in practice: if your service tracks referrals but not attendance, you cannot tell the difference between a working sequence and a broken one. Build in a mandatory contact at week four — not a survey, a phone call — with a simple binary log: attended or didn’t, and if not, why. That single data point turns a referral pipeline into something you can actually manage.

The National Academies’ 2020 consensus report on isolation in older adults, and the clinical commentary that followed it, both push toward this same discipline from the health-system side: routine, structured assessment rather than one-off screening. The same logic applies downstream of the referral. A link worker who screens well in week one but has no mechanism to know what happened in week four is running half a system.

None of this requires new funding. It requires deciding, before the first referral goes out, that “attended” is a status you will actually record.

What this does not solve

This sequencing helps people who are already inside the system — referred by a GP, engaged enough to take the first call, willing to be followed up. Holt-Lunstad’s 2015 meta-analysis found isolation and loneliness carry mortality risk comparable to other major clinical risk factors, which is the reason health systems care about this at all. But the people at the highest end of that risk are often the hardest to refer in the first place: those with no GP contact, no one flagging the isolation, no first phone call to answer. A better four-week sequence makes existing referrals count for more. It does not find the people who never get referred.