When a Commissioner Asks for Loneliness Outcomes You Cannot Promise

A commissioner writes a target into the contract: reduce loneliness among referred patients by a set margin over twelve months, measured against a validated scale. The link worker service signs it. Eighteen months later, someone has to explain why the number did not move the way the contract implied it would.

This is a predictable failure, not a bad-luck one. The evidence base for social prescribing does not currently support that kind of promise, and it is worth being precise about why, so you can negotiate a different ask before you sign, not explain a miss after.

What the evidence actually supports

A 2025 systematic review protocol makes the state of play blunt: despite growing adoption of social prescribing for older adults, its effectiveness remains unclear, and only one peer-reviewed randomised controlled trial exists in this specific area. One trial. Almost everything else practitioners cite is observational, uncontrolled, or qualitative.

That is not nothing. A 2021 systematic review found consistent increases in self-esteem and self-confidence across social prescribing studies, though it flagged limited trial evidence and heavy heterogeneity between programmes — different referral routes, different activities, different measures, making pooled effect sizes close to meaningless. A separate 2021 review of loneliness-specific outcomes found all nine included studies reported positive individual impacts, and three reported reductions in service use such as GP visits or emergency attendance. A 2022 qualitative synthesis found something more specific and more useful: participants describe benefit that extends beyond social contact itself, toward restored purpose and meaningful participation, and that structured, purposeful activity appears to work better than contact alone.

Put together, this is a reasonable case that social prescribing helps some people feel differently about their lives. It is not a case that it reliably moves a loneliness score by a defined amount, on a defined timeline, at population scale. Those are different claims, and commissioners routinely ask for the second while the literature can only support the first.

The one randomised trial worth naming — a 2024 dual RCT of a volunteering intervention among lonely older adults in Hong Kong — exists precisely because trials in this space are rare enough that a single one stands out. Its existence is a reminder of how thin the comparison group is everywhere else.

What this means in practice: if a contract specifies a numeric loneliness reduction target for a social prescribing pathway, that target was not derived from evidence about what social prescribing reliably does. It was derived from what looks fundable. Renegotiate the outcome before you renegotiate anything else.

The named failure mode: the borrowed number

Call it the borrowed number problem. A commissioner needs an outcome for a business case. They find a big, credible-sounding figure — the National Academies’ estimate that roughly a quarter of adults 65 and older are socially isolated, or the American Heart Association’s finding that isolation and loneliness carry something like a 30% increased risk of heart attack or stroke — and treats it as if it describes what your specific twelve-week programme, for forty referred patients, in one borough, should achieve. It does neither. Those are population-level associational findings, not intervention effect sizes, and the American Heart Association’s own scientific statement is explicit that the absence of intervention evidence is the central gap in this research. The number was borrowed from epidemiology to answer a programme-design question it was never built to answer.

The National Academies’ 2020 report and the clinical commentary that followed it push in the opposite, more useful direction: they call for routine assessment of isolation in health care settings, not for a promised reduction. Assessment is achievable and measurable. A defined reduction, absent a controlled evidence base, is not.

What to do instead

1. Separate the assessment claim from the outcome claim. You can commit to systematically screening for isolation and loneliness at referral, and to tracking who was screened and referred. That claim is supportable. A committed percentage reduction in loneliness scores is not.
2. Report mechanism-level outcomes, not the headline construct. Track participation, retention, and self-reported change in purpose or confidence, which the qualitative literature does support, rather than a UCLA Loneliness Scale delta with a target attached.
3. Distinguish isolation from loneliness in the target itself. A 2024 study on the interplay between isolation, age, and loneliness during the pandemic found the relationship between the two varies by age group — they are not interchangeable, and a target that conflates them will be met or missed for reasons that have nothing to do with the programme.
4. Put the evidence gap in writing to the funder before signing, not after reporting. State that the literature supports plausible benefit and does not support a defined effect size, and propose what you will measure instead. Funders who cannot accept that framing are asking for a number nobody can deliver honestly.
5. If a numeric target is contractually unavoidable, scope it to something you actually control — referrals completed, sessions attended, follow-up contact made — rather than to a downstream psychological construct that many things outside the programme also influence.

Evidence status

Claim	Status
Social prescribing improves self-esteem and confidence	Supported by systematic review, but heterogeneous evidence
Social prescribing reduces loneliness by a specific margin	Not supported — no controlled evidence at this level of specificity
Structured, purposeful activity outperforms unstructured contact	Supported by qualitative synthesis, not yet by trials
Population isolation prevalence figures apply to your local cohort	Not supportable — these are national or global estimates, not programme effect sizes
Routine screening for isolation in health settings is worthwhile	Supported by consensus report and clinical commentary

What this does not solve

None of this tells you how to help a commissioner who has already built a business case around a number they cannot get evidence for, or how to fund a service when funders reward confident overstatement more than honest scoping. It also says nothing about reach: this entire negotiation happens only for people already inside a referral pathway, which is a small and self-selecting slice of anyone who is isolated or lonely.