Designing the Handoff: What Happens After You Refer an Employee for Loneliness or Isolation

An employee tells their manager they’re struggling. The manager does the right thing and points them to the EAP, or an internal wellbeing line, or — increasingly — a social prescribing-style referral if the organisation has built one. Three weeks later, nobody knows whether that employee ever made contact with the service they were referred to. HR has closed the ticket. The manager assumes it’s handled. The employee, if they didn’t follow through, is now worse off than before they disclosed anything, because they raised a hand and watched it drop.

This is not a training problem or an awareness problem. It’s a pathway design problem, and it’s the part of workplace loneliness programming that gets the least attention because it happens after the visible part of the intervention is over.

The referral is not the intervention

Gallup’s 2024 workplace data puts loneliness at roughly one in five employees globally, higher among people working fully remote and higher again among workers under 35. Cigna’s 2020 workplace survey found lonely workers miss work twice as often for illness and five times as often for stress — the kind of gap that makes a referral pathway a genuine cost issue, not just a wellbeing gesture. But a referral is a piece of paper, or a link, or a phone number. It only becomes an intervention if the person on the other end of it actually receives support. The qualitative evidence on social prescribing — the closest analogue we have to structured workplace referral pathways, since both route a person from one service into community-based support — is consistent on this point: a 2022 meta-synthesis found that people describe benefit coming from restored participation and purpose, not from the act of contact alone. Structured, purposeful engagement outperforms a name and a number.

That distinction matters for workplace design because most internal referral pathways are built to optimise the referral — get the disclosure recorded, get the resource named, close the case — and not the connection. The failure isn’t usually a bad resource at the other end. It’s the silence in the middle.

The referral cliff

Call this the referral cliff: the point at which responsibility for a person’s wellbeing passes from someone who knows them (a manager, an HR contact, an occupational health advisor) to a service that has never met them, with no mechanism to check the transfer actually happened.

The referral cliff has three recognisable shapes:

1. The single-handoff cliff. One person makes the referral and one link is sent. If the employee doesn’t open the email or doesn’t call, there’s no second attempt, because nobody owns follow-up.
2. The anonymity cliff. The workplace resource (often an EAP) is confidential by design, which is right for trust — but it means the organisation has no visibility into uptake at the individual level, and often not even at the aggregate level, beyond a headline utilisation rate reported once a year.
3. The construct-mismatch cliff. The employee was referred for “isolation” when what they’re experiencing is loneliness, or vice versa, and the service they land in isn’t built for that. A 2024 study of healthcare workers found isolation and loneliness behave as separate constructs with different correlates and different responses to remote work — treating them as interchangeable at the point of referral routes people to the wrong kind of help.

Most workplace pathways are vulnerable to all three at once, because they were built to satisfy a duty-of-care checklist rather than to move a specific person from disclosure to support.

What a workable pathway needs

A referral pathway that survives the handoff needs four things that most current EAP-and-link setups don’t have.

A named owner for the follow-up window, not just the referral moment. Someone — a link worker equivalent, an occupational health nurse, a trained peer contact — needs responsibility for checking, within an agreed window, whether contact was made. This does not require breaching confidentiality about content; it requires a yes/no on whether the loop closed. If your pathway cannot answer “did this person reach the service” for a given case, you don’t have a pathway, you have a signpost.

A second attempt built into the design, not left to initiative. Single-touch referrals fail for the same reason single-touch anything fails: most people don’t act on the first prompt, especially when the prompt concerns something they feel some shame about. Build in a scheduled second contact — a check-in message, a follow-up call — as a default step, not an exception someone has to remember to do.

A distinction between isolation and loneliness at the point of triage. These are different problems. Someone who is objectively isolated (works alone, remote, few structural opportunities for contact) needs a pathway toward contact — a team ritual, a co-working option, a community group. Someone who is lonely despite having contact needs something closer to the social prescribing model: purposeful activity that restores meaning, not just proximity to other people. Referring both types into the same generic “wellbeing resource” line is a common and avoidable error.

A record of what happens after the referral, even in aggregate. The National Academies’ clinician-facing commentary on isolation and loneliness argues that routine assessment in health settings requires infrastructure to track outcomes, not just to flag the problem. The same logic applies inside an organisation. If a wellbeing pathway can report referral volume but not aggregate follow-through, it cannot be improved, because nobody can see where it breaks.

Evidence status

Claim	Evidence status
Loneliness is elevated among remote and younger workers	Reasonably well supported — Gallup’s 2024 global workplace data
Lonely workers have higher absence and stress-related absence	Supported by a single large employer survey (Cigna, 2020); not independently replicated at that scale
Isolation and loneliness are distinct and need different responses	Supported for healthcare workers specifically; plausible but not established generally across occupations
Structured, purposeful referral outperforms contact alone	Supported by qualitative synthesis in social prescribing; not tested in a workplace-specific trial
Systematic follow-up improves uptake of a referral	Plausible and consistent with the broader social prescribing literature on service linkage, but not directly tested as a workplace intervention

That last row is worth sitting with. The case for building follow-up into a pathway is an inference from adjacent evidence — social prescribing research on what makes referrals land — not a workplace-specific trial. Treat it as a strong design principle, not a proven outcome.

What this means in practice: if your EAP or wellbeing referral process cannot tell you, in aggregate, what proportion of referrals result in contact with the service, that is the first thing to fix — before adding a new programme, a new provider, or a new awareness campaign. A pathway you cannot see into is a pathway you cannot improve.

Building this without breaching confidentiality

The tension programme owners raise immediately is confidentiality: EAPs are often deliberately walled off from HR precisely so employees will trust them enough to use them. That’s correct and shouldn’t be undone. But confidentiality about content is not the same as invisibility about process. Most EAP contracts can be structured to report aggregate connection rates — the percentage of referrals that result in a first contact within, say, two weeks — without identifying anyone. If your current contract doesn’t include that reporting line, it’s a negotiable term, not a fixed feature of confidential services.

For non-EAP pathways — internal peer support, community group referrals, occupational health-led routes — the same principle holds with more room to move, since these are often internal enough that a named follow-up contact is feasible without the same confidentiality wall.

What this does not solve

None of this reaches the employee who never discloses in the first place, which on the workplace loneliness numbers is most of them. A well-designed referral pathway improves outcomes for the person who already put their hand up. It does nothing for the person sitting quietly at 20% loneliness prevalence who never tells anyone, and no amount of handoff design changes that; it requires a different intervention aimed at disclosure itself, which is a separate design problem. This note assumes the disclosure has already happened and asks only what the organisation does next — because what happens next is where most current pathways quietly fail, and it is also the cheapest part of the system to fix.