The Discharge Cliff: Why Time-Limited Social Prescribing Referrals Lose Their Gains

The standard social prescribing model for loneliness runs like this: a GP or link worker identifies someone isolated, refers them to a community group — a walking group, a choir, an arts session — for a fixed course of six to twelve weeks, then discharges them. The referral is the intervention. The clock starts at intake and stops at the final session.

This is not a fringe design. It is close to the default across UK social prescribing schemes, and versions of it appear anywhere a health or social care system funds “connection” as a time-boxed service. It is also the design most likely to show a measurable win at week six and a fade by week twenty. Call it the discharge cliff: the point at which the structure that was holding someone’s social contact together — the referral, the named group, the light institutional push — is withdrawn, and the contact does not survive on its own.

What the evidence actually supports

A 2021 systematic review of social prescribing and wellbeing found consistent increases in self-esteem and self-confidence across studies. That is a real finding, but it is not the same as a reduction in loneliness, and the review’s authors flagged limited trial evidence and heavy heterogeneity across programmes — different populations, different group types, different follow-up windows. A separate systematic review focused specifically on loneliness found all nine included studies reported positive individual impacts, and three found reductions in GP, emergency, or inpatient service use. Again real, but “positive individual impacts” in nine heterogeneous studies is a low bar, and none of the included evidence tracked whether social contact persisted once the programme ended.

Here is the evidence status, claim by claim:

Claim	Evidence status
Social prescribing referral increases self-esteem/confidence during the programme	Supported, across multiple studies
Referral reduces loneliness scores during the programme	Weakly supported, small heterogeneous studies
Gains persist after discharge	Not measured in the available reviews
Referral reduces downstream service use	Suggestive, three of nine studies in one review
The model builds durable social networks, not just attendance	Not tested

That last row is the whole problem. Nobody is measuring the thing that would tell you whether the discharge cliff is real, because almost no programme follows participants past the end of the referral.

Why the design produces the cliff, not just fails to prevent it

Robert Putnam’s distinction between bonding and bridging social capital is useful here. A six-week referral to a group of strangers is, at best, an attempt to generate bridging capital — loose ties across otherwise unconnected people. Bridging ties are real and useful, but they are also the most fragile kind: they depend on continued proximity and repeated low-stakes contact, which is exactly what a fixed-term programme removes at the moment it ends.

Ray Oldenburg’s concept of the third place makes the same point from a different angle. A third place works because it is durable and low-commitment — you can drop in indefinitely, at no administrative cost, and the tie renews itself through routine. A twelve-week course is the opposite of a third place. It has a start date, an end date, and usually a facilitator whose job ends with the contract. When the facilitator and the structure disappear, so does the reason the group kept meeting at the same time in the same room. Attendance is not habit yet. Twelve weeks is rarely enough to convert a scheduled activity into a self-sustaining one, particularly for people whose isolation was severe enough to warrant referral in the first place.

The AARP’s 2018 national survey of adults 45 and older adds a mechanism to this. It found the strongest predictors of loneliness were the size and diversity of a person’s social network, not simply whether they had some contact — and that only 33% of people who had spoken to their neighbours were lonely, against 61% of people who never had. Loneliness tracks network breadth and depth, not attendance at a single course. A referral programme that produces one narrow, temporary tie has moved the needle on almost nothing that the AARP data identifies as protective.

What this means in practice: if you are commissioning or running a fixed-term referral programme, build in a mechanism for the tie to outlast the referral — a standing weekly slot at the same venue that continues after the funded course ends, a peer host who takes over facilitation, or referral into an existing open-ended group rather than a bespoke cohort that dissolves on schedule. Measure attendance at week twenty, not just week six. If nobody is still going, the programme did not fail gently — it did exactly what its structure was built to do.

What to check before funding another one

Ask whether the group being referred into already exists independently of the programme, or was created for it. Groups that pre-exist the referral (an established choir, a running club) have a survival mechanism the referral doesn’t need to supply. Groups created specifically to receive referrals usually depend entirely on the funding cycle and disband when it ends.

Ask what happens on day one after discharge. If the answer is “nothing, formally,” the discharge cliff is built into the design, whatever the six-week outcome data shows.

What this does not solve

None of the available systematic reviews tracked participants long enough to confirm the discharge cliff empirically — this is a mechanism argued from social capital theory and the third-place literature, not a measured finding, and it deserves to be tested directly rather than assumed. It also does not address the reach problem common to all referral-based models: they depend on someone already in contact with a GP or link worker identifying the isolation in the first place, which misses people who never present to a service at all.