Designing the Referral Handoff So People Actually Show Up

A GP flags loneliness. A link worker takes the referral. A community group gets a name on a list. Somewhere in that chain, most people disappear. Not at the point of referral — at the point of handoff, the moment one person’s responsibility for a case becomes another person’s responsibility, or nobody’s.

Programme evaluations rarely measure this moment directly, which is part of the problem. They report enrolment and they report outcomes, and the gap between them — who was referred but never attended a single session — gets absorbed into “engagement,” a word that hides more than it reveals. If you are designing or commissioning a referral pathway, the handoff is the part of the system most worth engineering deliberately, because it is the part everyone else treats as administrative.

Why the pathway, not the intervention, is the design problem

The UK’s 2018 national loneliness strategy embedded social prescribing into mainstream health commissioning specifically because informal, non-clinical referral into community activity was judged to be under-resourced infrastructure, not because any single intervention had strong trial evidence behind it. That distinction still holds. A 2021 systematic review of social prescribing and loneliness found that all nine included studies reported positive individual impacts, and three reported reductions in service use — but the review sits on thin, heterogeneous evidence rather than trials with control groups. A parallel systematic review on wellbeing found similar gains in self-esteem and confidence, again while flagging limited trial evidence. A 2025 protocol reviewing social prescribing specifically for older adults notes that only one peer-reviewed randomised controlled trial exists in this space, despite the pathway being widely adopted across health systems.

None of this means social prescribing doesn’t work. It means the evidence base is currently better at describing what happens to people who complete the pathway than at explaining why so many don’t. A 2022 qualitative meta-synthesis of how people experience social prescribing found that the benefit people describe goes beyond social contact itself — restored participation, purpose, a sense of mattering. That’s a strong argument for the destination. It says nothing about whether people reliably arrive there, and the mechanics of arrival are where most programmes are silent.

The cold handoff

Call it the cold handoff: a referral is logged, a name and phone number are passed from clinician to link worker to activity provider, and at each transfer the receiving party has slightly less context, slightly less relationship, and slightly less obligation to follow up than the previous one. The referring GP feels the loop is closed once the referral is made. The link worker, managing a caseload in the hundreds, treats the community group as the endpoint. The community group treats a name on a list as an expression of interest, not a commitment, and has no mechanism to notice when someone doesn’t show up twice.

Nobody in that chain owns the failure. Everyone did their part correctly. The person who was lonely enough to be flagged by a clinician in the first place is now also managing the logistics of self-referral into an unfamiliar group, alone — which is precisely the task loneliness makes hardest.

Evidence status of common pathway design claims

Claim	Evidence status
Social prescribing improves self-reported wellbeing and confidence for people who engage	Reasonably supported by qualitative and observational reviews; not yet by controlled trials
Social prescribing reduces downstream GP or emergency service use	Suggested by a minority of included studies in systematic reviews; too few studies to generalise
Structured, purposeful activity outperforms unstructured social contact	Supported by qualitative synthesis and consistent with trial data on volunteering and behavioural activation
A warm introduction (a named person, a specific first session) improves attendance versus a cold referral	Plausible and consistent with clinical common sense; not directly tested at scale in the cited literature

• | Follow-up contact after a missed first session recovers a meaningful share of no-shows | Not directly evidenced here; treat as an operational hypothesis to test locally |

That last row matters. Much of what this article recommends is inference from adjacent evidence, not a directly cited finding. Say so plainly rather than dressing up a sensible operational guess as proven practice.

What the intervention trials actually tell you about design

The two best-controlled trials available on loneliness interventions both concern older adults and both compare something structured against befriending. The HEAL-HOA trial on volunteering in Hong Kong tested prosocial engagement against a control in a randomised design — one of the few loneliness interventions tested this way rather than evaluated as an uncontrolled programme. A later HEAL-HOA trial went further: telephone-delivered behavioural activation and mindfulness, delivered in eight 30-minute sessions by trained laypeople who were themselves older adults with lived experience of loneliness, reduced loneliness at twelve months more than a befriending comparison group, in a trial of 1,151 older adults who were poor, living alone, and digitally excluded. Separately, a 2025 randomised trial of befriending in residential aged care found real reductions in UCLA Loneliness Scale scores at eight and sixteen weeks — befriending works, it’s simply the weaker arm when tested head-to-head against something more structured.

The design lesson isn’t about which activity to prescribe. It’s that every trial that worked had a defined person delivering a defined number of contacts on a defined schedule, and someone was accountable for whether those contacts happened. That is a description of a handoff done correctly, not just an intervention done correctly.

What this means in practice: Build the pathway with a named owner at every transition, not just a named owner for the case. A referral should specify who confirms the first session happened, by what date, and what happens if it didn’t. If no one in your pathway can answer “who finds out if this person never shows up,” you have a program design gap, not an engagement problem.

Four things to fix before you fix the activity roster

1. Name a single point of continuity. The National Academies’ 2020 consensus report on isolation in older adults calls for the health system to assess isolation routinely, and a 2020 clinical commentary on that report pushes further, arguing that assessment without a designated person to act on it changes nothing. Whoever does the assessment should not be the last person who owns the case.
2. Specify the first contact, not just the referral. “Referred to X group” is not a plan. “First session confirmed for Tuesday, contacted by Priya” is. The gap between those two sentences is most of the attrition.
3. Build a missed-first-session trigger. Treat a no-show at the first session as a signal requiring a follow-up call, not as evidence of disinterest. AARP’s 2018 survey of adults 45 and older found that network size and physical isolation are the strongest predictors of loneliness — exactly the people least likely to self-initiate a second attempt after a missed first one.
4. Feed attendance data back to the referrer. Most pathways are one-directional. The GP or link worker who made the referral rarely learns whether it worked, which means the system never learns from its own failures at scale.

What this does not solve

A well-designed pathway still only reaches the people who reach a referrer in the first place — a GP visit, a link worker conversation, a self-referral form. It does nothing for the isolated person who never enters a system capable of noticing them, which by most estimates is a larger group than the one being referred. Fixing the handoff makes existing pathways less wasteful. It is not a substitute for the harder, unsolved problem of finding people who are isolated enough that no institution currently sees them.