Who decides what Keep Britain Working counts?

The government’s Keep Britain Working programme has a line in it that I’ve been waiting years for someone official to say. “Currently, sickness absence is tracked inconsistently, and return-to-work outcomes are rarely measured.”¹ It comes from the July 2026 announcement of a Workplace Health Intelligence Unit. The government says the proposed Unit will track sickness absence, return-to-work outcomes and disability participation.¹

I’ve spent fifteen years in employee benefits, so I read that as a data problem before I read it as a policy. That’s a diagnosis, and there’s no treatment. Here is what the programme says it will do, the question I think that leaves, and the draft I’ve built to put a possible answer in front of people.

What the programme has said

Keep Britain Working began as a review led by Sir Charlie Mayfield, with its final report published on 5 November 2025. It proposed employer-led Vanguards and a healthy working standard.²

The June 2026 update lists seven jobs for the Intelligence Unit. The first is standard setting: “Define standardised approaches to measuring and collecting performance data, and health and work-ability data.” The others cover collecting data from employers and providers; holding individual health and work-ability profiles securely, with consent for onward sharing; aggregation; confidential benchmarking for employers and providers; evaluating interventions; and informing policy.³

BSI is developing the employer-facing Healthy Working Lifecycle standard. The June update names Valerie Todd as chair of its drafting panel and says the standard should be workable and proportionate for employers, including smaller businesses.³ BSI’s March announcement expected a draft for public comment in early 2027 and described the goal of a voluntary standard by 2029.⁴

The bit in the middle

Those dates describe the BSI standard. They do not tell us when the Intelligence Unit’s data definitions will be ready. The June update separately lists work on its data requirements and architecture.³

The unglamorous question never makes the conference slide, because it can’t be answered with a stock photo of a smiling nurse. What will an employer or provider actually record? What counts as a return to work? Does it mean full duties, a phased return, or a return with adjustments? And when do we check whether it lasted? I want those choices tested against the records people use while the programme develops its approach.

So I wrote one and gave it away

Nobody asked me to. Nobody ever does. That turns out to be the whole story of standards.

The Open Workplace Health Standard is an open data standard for workplace health: sickness absence, return to work, occupational health, wellbeing measurement and benefit provision, defined once, in plain language and machine-readable schemas.⁵ It’s at version 0.2, an early open draft, published for review.⁶

A few things about it that matter for anyone reading this with a notebook.

The draft uses the ONS sickness absence categories, the HSE Management Standards domains and the statutory fit note as reference points.⁵ Its return-to-work outcome list proposes five categories: returned to full previous duties, phased return, returned with adjustments, did not return, and an ill-health retirement or medical capability exit. It marks these as provisional pending WHIU definitions.⁷

It also reserves a whiu: namespace for future official definitions. That leaves a place to refer to them; whether today’s records fit those definitions will depend on what the Unit decides.⁵

OWHS prohibits direct identifiers in its payloads. Its schemas reject undeclared fields and check other structural rules, but cannot recognise a name or clinical detail hidden inside an allowed text value. Passing the validator therefore does not establish that a record meets the privacy rules.⁵ Pseudonymised data remain personal data for someone who can identify the person using additional information.⁸

It’s free. The specification is CC-BY 4.0, the schemas and tools Apache 2.0, and anyone can implement it commercially without asking me.⁹

What it isn’t

I wrote the draft because I wanted a common record I could use. Publishing it gives people something concrete to test and challenge. I want that work to be useful to the programme as it develops its definitions. The draft is independently maintained by Alltoogether, and the governance document commits to adopting official definitions as they emerge.⁹

It’s early. I’m not aware of anyone outside Alltoogether implementing it yet, which is the least impressive sentence on this site. I’ve left it in, partly for honesty and partly because it’ll be funny later, one way or the other. The instrument registry alongside it has its grades frozen until two named psychometric raters who don’t work for the steward have joined.⁵ The steward is Alltoogether, my company, which develops commercial products in the same field. The governance document acknowledges that conflict and commits to logging material decisions publicly. It sets out triggers for sharing control, including independent use, recurring contributions and independent review.⁹ If I steward it badly, the licences let someone else take a copy and continue the work.⁹

What I’d ask

If you’re a journalist covering Keep Britain Working, I’d ask which absence, return-to-work and participation measures employers and providers will record, who will define them, and how individual records become safe comparisons. Which details will stay with the employer or provider, and which will reach the Intelligence Unit? Those are the questions I’d want answered before treating two numbers as comparable.

If you run a Vanguard, or you’re a provider, try mapping a realistic, made-up record and tell me which fields or definitions don’t fit. I want an example I can improve the draft against. Praise is lovely and does nothing. A record that doesn’t fit is worth an afternoon of my life, and I don’t say that about much.

Footnotes

1. Department for Work and Pensions (2026). Keep Britain Working continues drive to stop people falling out of the workforce, 3 July 2026. gov.uk. The quotation and the three measures are in the section on the Workplace Health Intelligence Unit. ↩ ↩²

2. Department for Work and Pensions and Department for Business and Trade (2025). Keep Britain Working Review: Final report, published 5 November 2025. Publication page and executive summary, paragraph 33. ↩

3. Department for Work and Pensions (2026). Keep Britain Working: the story so far (June 2026), first published 3 July 2026; publication page updated 4 August 2026. Report and publication history. See the WHIU function table under Infrastructure and market frameworks, and the Healthy Working Lifecycle Standard and Data subsections under Next steps. ↩ ↩² ↩³

4. BSI (2026). New BSI standard to be developed to address UK’s economic inactivity crisis, 31 March 2026. Press release, opening four paragraphs: the employer-facing standard, the voluntary-standard goal for 2029 and the expected public-comment draft in early 2027. ↩

5. Open Workplace Health Standard, README and v0.2 draft specification, checked 16 September 2026. See the specification’s primary-source anchors, code lists, validation and conformance sections, and the honesty pass on future WHIU definitions. The README’s instrument-registry section states the frozen-grades condition. These are the project’s account of its design and limits. ↩ ↩² ↩³ ↩⁴ ↩⁵

6. Open Workplace Health Standard, openworkplacehealth.org, checked 16 September 2026. The home page labels the specification v0.2 draft and open to review; the pinned repository sources above document that version. ↩

7. OWHS, rtw-outcome codelist, version 0.1.0, five values and provisional-status note. This codelist version is used by the v0.2 ReturnToWorkOutcome schema. It is OWHS’s proposed vocabulary. ↩

8. Information Commissioner’s Office, Pseudonymisation, Is pseudonymised data still personal data?, checked 16 September 2026. ↩

9. OWHS, GOVERNANCE.md, checked 16 September 2026: Current stewardship, The conflict of interest, How governance evolves, Continuity, Licences, Change process, and Relationship to government and official definitions. These are published commitments, not an independent audit of compliance. ↩ ↩² ↩³ ↩⁴