Building an Honest Cost Case for a Connection Programme

A funder asks what a loneliness intervention will save the system. The honest answer is: less than the mortality statistics imply, and probably not in a way you can prove within a two-year grant cycle. Say that up front. The alternative — quoting a big epidemiological number as though it were a return-on-investment figure — is the fastest way to lose credibility with any funder who has seen this pitch before.

Where the cost case usually goes wrong

The most common move is to take a population-level mortality statistic and treat it as a service-level saving. Julianne Holt-Lunstad’s 2015 meta-analysis found that social isolation carries an odds ratio of 1.29 for early mortality, and her 2010 meta-analysis, covering 308,849 participants, found that stronger social relationships were associated with a 50% increased likelihood of survival. The American Heart Association’s 2022 scientific statement put the increased risk of heart attack, stroke, or death at around 30%. These are real, well-replicated findings. None of them tell you what a twelve-week befriending programme in one borough will save the local NHS trust this year.

Call this the epidemiology-to-invoice error: converting a population risk ratio into a claimed per-participant saving, without an intervention study anywhere in the chain. It happens because the mortality evidence is strong and the intervention evidence is thin, and a funding proposal needs a number. The mortality evidence answers “does disconnection matter?” It does not answer “does this programme change outcomes enough to justify its cost?”

The intervention evidence, where it exists, is real but modest. A 2021 systematic review of social prescribing and loneliness found that all nine included studies reported positive individual impacts, and three reported reductions in GP, emergency, social worker, or inpatient service use. A companion review that year found gains in self-esteem and self-confidence but flagged limited trial evidence and heterogeneity across programmes as the reason those gains are hard to generalise. A 2025 systematic review protocol on social prescribing for older adults noted that despite growing adoption, effectiveness remains unclear, and that only one peer-reviewed randomised controlled trial exists in the entire area. That is the honest baseline: a promising, mostly observational literature, with almost no experimental evidence at the level of cost.

What you can defensibly claim

Build the case in two separate layers and do not let them blur into one number.

Layer one: the case for why this matters at all. This is where the mortality and morbidity evidence belongs. The National Academies’ 2020 consensus report put social isolation among adults 65 and older at roughly one in four, and called on the health care system to assess it routinely. The 2023 Surgeon General’s advisory compared the mortality risk of disconnection to smoking up to 15 cigarettes a day. The CDC’s 2024 surveillance report links loneliness and lack of social and emotional support to heart disease, stroke, dementia, type 2 diabetes, depression, and anxiety. Use these to establish that the problem is large and consequential enough to warrant investment. Do not use them to size the investment.

Layer two: the case for what this specific programme is likely to change. This has to be built from absence data, service-use proxies, and comparable programme evaluations — not epidemiology. If you are working with employers, Cigna’s 2020 loneliness index found lonely workers miss work twice as often due to illness and five times as often due to stress; that is a defensible basis for an absenteeism-based cost estimate in a workplace programme, because it is measuring the same population you are serving, not a national cohort study forty years long. If you are working in health and care settings, the social prescribing service-use reductions from the 2021 systematic review are the right comparator — smaller, less dramatic, but actually about the intervention type you are running.

What this means in practice: Never quote a mortality risk ratio as a savings figure. Use mortality and morbidity evidence to justify why the problem deserves funding, and use service-use and absenteeism data — ideally from your own population or a closely comparable one — to estimate what the programme itself might save. If you cannot get comparable service-use data, say your cost case is provisional and commit to measuring it, rather than manufacturing a number.

Evidence status for the claims that usually appear in a bid

Claim commonly used in funding bids	Evidence status
Social disconnection raises mortality risk substantially	Strong — multiple large meta-analyses, consistent direction
Social disconnection raises cardiovascular event risk	Strong — AHA scientific statement, explicit about the gap in intervention evidence
Lonely workers cost employers more through absence	Moderate — single large industry survey, not independently replicated
Social prescribing reduces GP or A&E use	Weak-to-moderate — a minority of included studies report it, and heterogeneity is high
This specific programme will reduce NHS or care costs by £X per participant	Almost never supportable — requires a cost-effectiveness study this field mostly lacks

What to put in the bid instead of a savings figure

State the size of the problem using the strongest available population evidence. State what the programme does, and why the mechanism is plausible given the intervention literature that does exist. Propose a modest, specific service-use or absence metric you will track, and commit to reporting it whether or not it moves. Funders who work in this space regularly enough to have seen inflated bids before will trust a proposal that names its own evidential ceiling more than one that doesn’t.

What this does not solve

None of this produces a bulletproof return-on-investment figure, because that figure largely doesn’t exist yet in the peer-reviewed literature for programme-level connection work. It also does not help you reach commissioners who require a savings number before they will fund anything — some will still say no. And it says nothing about reach: an honest cost case for a well-evaluated programme still only covers the people who enrolled, which is rarely the population with the least existing support.