Designing the Handoff: A Framework for Referral Pathways Into Connection Services

The referral gets made. The link worker sends the email. The community group has a slot. And then, somewhere in that gap, a large share of people who were identified as isolated never show up anywhere. This is not a motivation problem in most cases. It is a design problem, and it sits in the handoff — the point where responsibility for a person passes from one party to another, and where, too often, nobody is actually holding it.

Referral pathways for loneliness and isolation get built the way most clinical pathways get built: as a flowchart from identification to intervention. That flowchart is usually fine on paper. It falls apart at the joints — the specific moments where one actor’s job ends and another’s has not yet started. This framework treats those joints as the design problem, not the intervention itself.

Why the joint is where pathways fail

A GP, care coordinator, or employer identifies someone as isolated. A social prescribing link worker, a community organisation, or a programme takes over. The National Academies’ 2020 consensus report was blunt about the first half of this problem: the health care system rarely assesses isolation in any structured way, so a lot of referrals originate from an ad hoc judgment rather than a repeatable process. The clinician-facing commentary on that report pushed further, arguing that routine assessment requires clarity about what happens after a positive finding — otherwise clinicians reasonably ask why they should screen for something they cannot act on.

That “what happens after” is the handoff. Systematic reviews of social prescribing find real reported benefits — increases in self-esteem and confidence, and in some cohorts reduced use of GP, emergency, and inpatient services. But the same reviews are consistent about heterogeneity: the studies included describe very different pathways, staffing models, and follow-through, which is a polite way of saying nobody has isolated which parts of the pathway are doing the work.

A qualitative meta-synthesis on how people experience social prescribing offers the clearest clue. Benefit, in participants’ own accounts, extends beyond the moment of social contact to a restored sense of meaningful participation and purpose — and structured, purposeful group activity appears to do more than contact alone. That has a design implication most pathways ignore: a referral that ends at “here is a phone number for a walking group” is handing someone a much weaker product than a referral that ends at “someone is expecting you on Thursday and will notice if you don’t come.”

The four joints, and who owns each one

Map any referral pathway onto these four transitions. Almost all attrition happens at one of them.

1. Identification to referral. Someone notices isolation and decides to act on it. Weak point: no shared threshold for what counts as isolated enough to refer, so referral rates vary wildly by individual clinician or manager judgment.
2. Referral to first contact. The referral is made; someone from the receiving side has to reach the person. Weak point: this step is frequently unowned — the referrer assumes the receiving organisation will follow up, and the receiving organisation assumes the person will call.
3. First contact to first attendance. The person has spoken to someone but has not yet shown up anywhere. Weak point: too much time elapses, or the offer doesn’t match what was promised.
4. First attendance to sustained participation. The person comes once. Whether they come back depends on what happened in the room, not on anything upstream.

The most common failure is at joint two. A referral is logged as “complete” the moment it is sent, when nothing has actually happened for the person yet. Call this the completed-referral illusion: a pathway that measures itself on referrals made rather than contacts landed will always look healthier than it is.

What this means in practice: name an owner for joint two explicitly, in writing, before you launch anything. If the answer is “the community organisation will call them,” write down the maximum number of days that can pass before someone escalates a referral that has gone quiet. Silence should trigger an action, not a default closure.

Building the pathway backwards

Design pathways starting from joint four and working back, not the reverse.

• Start with what sustained participation requires. Klinenberg’s account of social infrastructure argues that shared physical spaces shape rates of contact partly through repeat, low-friction exposure. A referral pathway that dumps someone into a one-off event has recreated the opposite of that: a high-friction, single-exposure design.
• Then design first attendance around that. If the destination is a recurring group, the first attendance should be framed to the person as “the first of several,” with a specific next date already named, not “come along and see.”
• Then design first contact to make that credible. Whoever makes first contact needs to be able to say something specific and true about what will happen — not a general description of a service.
• Then design the referral trigger to match capacity at the other end. A pathway that refers faster than the receiving organisation can make first contact will produce a growing backlog that quietly becomes joint two’s weak point, no matter how well the rest is designed.

Evidence status of the underlying claims

Claim	Evidence status
Social prescribing produces reported gains in confidence and self-esteem	Supported by systematic review, but heterogeneous programme designs limit generalisation
Structured, purposeful activity outperforms unstructured social contact	Suggested by qualitative synthesis; not tested against alternatives in controlled comparison
Routine isolation screening improves outcomes	Recommended by consensus report; the report itself notes evidence gaps on downstream effect
Reduced use of emergency and inpatient services follows social prescribing	Reported in a minority of included studies in a systematic review; not the median finding
National loneliness strategy improves pathway design	The UK strategy funded social prescribing and embedded measurement, but did not itself evaluate handoff quality

What this does not solve

None of this addresses who gets identified in the first place. Pathways built on clinician judgment inherit whatever bias that judgment carries, and a beautifully designed handoff still only serves people who were noticed, referred, and had a phone number that worked. The people most isolated are often the ones least visible to any system with a referral form.