The Drop-In Group Referral: Why It Fails at Week Four

A link worker refers a patient to a weekly community drop-in — a coffee morning, a walking group, an open craft session. The patient attends once, maybe twice. By week four they have stopped. The referral is marked complete in the record. Nobody follows up, because nobody was tasked with following up. This is the most common social prescribing referral in the country, and it is also the one most likely to produce nothing measurable.

Call it the week four problem. It is not a failure of the group itself, and it is not usually a failure of the patient. It is a failure of the referral model: a single handoff to an open, unstructured space, with no mechanism to notice or respond when the person doesn’t come back.

Why the model looks sound and isn’t

The logic behind the drop-in referral is reasonable on paper. Isolated people need contact with other people; a community group provides contact with other people; therefore refer them. Robert Putnam’s distinction between bonding and bridging social capital is often invoked to justify this — a drop-in group is a low-stakes bridging opportunity, easier to enter than a tight-knit club. Ray Oldenburg’s concept of the third place is invoked too: informal, low-commitment public gathering spots are supposed to be exactly where casual social contact happens.

Both ideas are right about what third places and bridging ties do for people who are already inside them. Neither says anything about how a person gets inside them in the first place, especially a person who was referred precisely because they weren’t managing that on their own. A third place works because of repeated, low-friction, self-sustaining visits. A referral produces one visit, engineered from outside. Those are different problems, and the drop-in model quietly assumes the first solves the second.

The qualitative evidence on social prescribing backs this up more specifically than the loneliness literature does in general. A 2022 meta-synthesis of how people actually experience social prescribing found that the benefit participants described was not “contact” in the abstract — it was restored participation and purpose, arrived at through activity that had some structure and continuity to it. Contact without structure did not register as valuable to the people receiving it. A systematic review the previous year, looking specifically at loneliness outcomes from social prescribing, found that all nine included studies reported some positive individual impact — but the interventions that produced it tended to have identifiable structure and a defined arc, not open attendance at an unstructured space.

What actually predicts whether someone shows up again

The AARP Foundation’s 2018 survey of adults 45 and older is useful here, even though it isn’t a social prescribing study. It found that the strongest predictors of loneliness were the size and diversity of a person’s social network and their degree of physical isolation — not simply whether they had somewhere to go. Only 33% of people who had spoken to a neighbour in the past week were lonely, against 61% of those who never had. The gap is about relationship, not attendance. A drop-in referral gets someone through a door once. It does nothing to build the small, repeated exchange — a name remembered, a seat saved, a “see you next week” — that turns a room full of strangers into a network.

The National Academies’ 2020 consensus report on isolation in older adults makes a related point from the clinical side: it calls for health systems to routinely assess isolation, not just refer once and close the file. Assessment implies a return visit, a check on whether the connection held. Most drop-in referrals have no equivalent step.

The evidence problem underneath the practice problem

None of this is settled science, and it is worth saying plainly. The main systematic review of social prescribing and wellbeing outcomes, from 2021, reports gains in self-esteem and confidence but flags limited trial evidence and heavy heterogeneity across programmes — meaning most of what’s known comes from small, differently-designed studies that are hard to compare or generalise from. Anyone commissioning at scale should treat “social prescribing works” as a directional finding, not a guarantee for any specific referral pathway.

Claim	Evidence status
Social prescribing produces self-reported wellbeing gains	Supported, but from a heterogeneous, largely small-trial evidence base
Structured, purposeful activity outperforms unstructured contact	Supported by qualitative synthesis; not yet quantified
A single referral with no follow-up predicts drop-off	Plausible and consistent with network-based loneliness research; not directly tested as a causal claim
Network size and physical isolation predict loneliness better than group attendance alone	Supported by AARP’s national survey data

What this means in practice: if your programme’s model is “refer once, to an open group, and close the case,” redesign the referral to include a scheduled check at week two and week four — not a survey, a person asking whether they went and what happened when they got there. Where possible, pair the referral with a named contact inside the group, not just an address and a time. That single change turns a one-time handoff into the beginning of a relationship, which is the thing the evidence actually says predicts whether the isolation improves.

What this does not solve

This fix helps people who already accepted a referral and made it through the door at least once. It does nothing for the larger group who never take up the referral at all — social prescribing pathways consistently reach people already willing to engage with a health or community service, which is a minority of anyone actually isolated. It also assumes the group itself has capacity to absorb a supported new member and a link worker with time to make a second call. Neither is guaranteed, and no amount of better referral design fixes a group that’s already full or a caseload that leaves no time for a week-four check-in.