Building a Cost Case for Workplace Connection Without Overselling It

You are asked for a number. A finance director wants to know what loneliness is costing the organisation before signing off a connection programme, and the instinct is to reach for the biggest available statistic: social isolation carries mortality risk on par with smoking, or a 29% increased risk of heart attack. Do not use those numbers in a budget request. They answer a different question than the one you are being asked.

Why the mortality statistics don’t belong in this document

Julianne Holt-Lunstad’s 2015 meta-analysis found that social isolation carries an odds ratio of 1.29 for early mortality, and loneliness 1.26. The American Heart Association’s 2022 scientific statement found isolation and loneliness associated with roughly a 30% increased risk of heart attack, stroke, or death from either. These are real, well-supported findings, and they are the reason loneliness belongs on a public health agenda at all.

They are also population-level epidemiological associations built from cohort studies tracking people over years or decades, not effects of a specific twelve-week programme, and the AHA statement itself names the absence of intervention evidence as the central research gap. Citing a mortality odds ratio to argue that a lunchtime walking group will reduce absenteeism is a category error, and finance directors who have sat through enough of these pitches will spot it. It also invites the obvious question back: if this is a mortality risk factor, why is the ask so small?

The honest cost case uses a narrower, less dramatic set of numbers, and argues from cost rather than from mortality.

The numbers that actually belong in a budget request

Cigna’s 2020 workplace report found that lonely workers miss work twice as often due to illness and five times as often due to stress, compared with workers who are not lonely. That is an absenteeism and productivity claim, measured in the same units a finance director already tracks. It sits alongside the finding that 61% of U.S. adults report loneliness at least sometimes, up seven points from the prior year, and that loneliness is markedly higher among younger workers — 73% of those aged 18 to 22, and over 80% of employed Gen Z. If your workforce skews young, this is the number that predicts your exposure, not the mortality figures.

Two limits on this data matter and should be stated in the same document that cites it. First, it is a single survey report from a health insurer with a commercial interest in the subject, not an independent peer-reviewed study — cite it as what it is. Second, “misses work twice as often” is a correlation from self-reported data, not a causal estimate of what an intervention would recover. Treat it as a plausible order of magnitude, not a number you multiply by headcount and present as a savings figure.

The failure mode: the borrowed statistic

Call this the borrowed statistic problem. It happens when a programme proposal cites a study measuring one thing — mortality risk in older adults, say, or cardiovascular outcomes in a general population — to justify spending on a different thing entirely, such as a peer-support group for a 200-person engineering team. The statistic is real. The application is not what the statistic supports. This is the single most common way connection programmes lose credibility with finance and leadership, because it invites scrutiny of the entire proposal once one number is shown to be doing work it cannot do.

The fix is not to drop numbers. It is to match the grain of the statistic to the grain of the claim: use workforce data for workforce claims, and hold the epidemiology in reserve for the paragraph that explains why the issue matters at all, clearly labelled as background rather than as the basis for a savings estimate.

What this means in practice: Build the cost case in two layers. Layer one, background: cite Holt-Lunstad or the American Heart Association once, briefly, to establish that social isolation is a recognised health risk — this earns the topic a place on the agenda. Layer two, the actual ask: use absenteeism, turnover, and engagement data specific to your workforce, or, absent that, the Cigna workplace figures, and be explicit that these are associational, not causal.

What to measure before you ask for money

A cost case built entirely on external statistics is weaker than one built on even a small amount of internal data. Before writing the proposal:

1. Pull existing absenteeism and turnover figures for the team or division in question, segmented by tenure if possible — loneliness risk is not evenly distributed, and younger and newer employees carry more of it.
2. Check whether your engagement survey already asks anything adjacent to connection or belonging. Most do, even if not labelled as loneliness. Report the actual figure, not an inferred one.
3. If no internal data exists, say so in the proposal rather than filling the gap with a national statistic dressed as a local one. “We do not currently measure this” is a more credible sentence than a misapplied citation, and it doubles as the case for the very small measurement step the programme should include from day one.

The evidence-status table

Claim	Evidence status
Lonely workers miss more work due to illness and stress	Moderate — single survey (Cigna), self-reported, not independently replicated
Social isolation raises mortality and cardiovascular risk	Strong — large meta-analyses and an AHA scientific statement, but population-level, not programme-specific
A workplace connection programme reduces absenteeism	Weak — plausible mechanism, no direct intervention trial cited here
Structured, purposeful group activity works better than casual social contact alone	Emerging — qualitative synthesis of social prescribing suggests this, but the evidence base is thin and mostly outside workplace settings
Younger workers carry disproportionate loneliness risk	Moderate-to-strong — consistent across Cigna’s workforce data and broader friendship surveys

The last row draws on a wider pattern than the workplace alone: the American Enterprise Institute’s 2021 survey found that 12% of Americans report no close friends at all, up from 3% in 1990, with the decline sharpest among men. That is not a workplace statistic, but it explains why a programme aimed only at “team bonding” may be solving the wrong layer of the problem for employees whose isolation predates the job.

A qualitative meta-synthesis of social prescribing found that participants describe benefit extending beyond social contact itself to a sense of restored purpose and participation, and that structured, purposeful activity appears to work better than unstructured contact. That is a useful design principle. It is not, on its own, a savings number, and should not be presented as one.

What this does not solve

A well-built cost case gets you budget. It does not tell you whether the programme you fund will reach the employees carrying the most risk — those most isolated tend to be the least likely to sign up for a voluntary lunch club, and no amount of rigor in the proposal changes that reach problem. Nor does it substitute for the internal measurement work of actually tracking whether the numbers move once the programme runs. Treat the proposal as the start of a measurement commitment, not the end of one.