Every autumn the figures arrive.

964,000 workers in Great Britain with work-related stress, depression or anxiety. 22.1 million working days lost. Fifty-two per cent of all work-related ill health, and 62% of all working days lost to it. Published on 20 November 2025, accredited official statistics, confidence intervals on every chart.

Then look at what the same document says about what is causing it. That data is an average over 2009/10 to 2011/12.

We count the damage every year. We last asked what was doing it fifteen years ago.

The counting is very good indeed

I want to be fair to the HSE statistics here, because the measurement is genuinely careful. The prevalence rate is 2,770 per 100,000 workers, and HSE says plainly that this is significantly higher than the year before. It breaks the figures down by industry, occupation, age, gender and workplace size. It tells you the three-year averages so you are not misled by a single year’s noise. It tells you the rates in each of the last three years sit above the 2018/19 pre-pandemic level.

It even tells you which sectors carry the most of it: public administration and defence, human health and social work, education. Three parts of the public sector, which is worth sitting with for a moment.

None of this is sloppy work. It is a serious annual account of how many people are being made unwell by their jobs, and roughly where.

The causes data stops in 2012

Further into the same publication there is a section on how the stress, depression or anxiety was caused or made worse by work. It is the section that would tell you what to change.

Its source line reads: average estimate over 2009/10 to 2011/12.

Underneath it sits a second dataset, from a network of GPs reporting newly diagnosed cases of work-related ill health. That one runs to 2015, and there is a sentence in the notes explaining why it stops there. HSE funding ended in 2016, so 2015 is the last year of data available.

So the position is this. We have an annual, accredited, statistically rigorous count of how many people work is making unwell. Our national picture of what work is doing to make them unwell is fifteen years old, and the medical surveillance that backed it up was switched off a decade ago.

You cannot fix what you have decided to stop measuring. And defunding the question is a decision, in a way that a rising number is not.

We already know the answer, which makes it stranger

Here is the part I keep turning over. It is not as though the answer is a mystery.

HSE’s fifteen-year-old data found workload pressures the predominant factor, with interpersonal relationships at work and changes at work also significant. Mental Health UK’s Burnout Report 2026 went and asked again this year, independently, and found a high or increased workload is still the biggest driver of stress at work at 42%, followed by unpaid overtime and fears about job security.

Fifteen years apart, two different organisations, two different methods, the same answer.

So this is not an open question that needs funding to settle. It is a settled question that stopped being asked officially, while the thing it describes carried on getting worse. Nine in ten UK adults now report high or extreme levels of pressure or stress in the past year. One in five workers took time off in the last year because of it.

Workload, unpaid hours, insecurity. Every one of those is a decision somebody made about how to run a workplace. None of them is a medical problem, and none of them is fixed by an app.

The number nobody quotes

The Burnout Report is usually reported through one figure: only 27% of people say mental health is genuinely prioritised and supported at their workplace through visible action and resources. Just over one in four. It is a good headline and it deserves the attention.

But the more useful number is the one directly above it, and I have not seen anybody quote it.

Twenty-nine per cent said their employer does raise awareness about mental health, but their managers lack the time, training or resources to provide meaningful support.

That is the largest single group in the whole question. Bigger than the 27% who say it is done properly. Bigger than the 18% who call it a tick-box exercise, and far bigger than the 10% who say it is not prioritised at all.

Read it again, because it is not what the usual telling of this story assumes. That 29% is not indifference. It is not cynicism or cost-cutting or a company that does not care. It is employers who have decided mental health matters, said so out loud, put the posters up, marked the awareness days, and then sent managers into those conversations with no time, no training and nothing to offer.

Awareness without capability. It is the most common state of workplace mental health in Britain, and we barely talk about it.

What happens when somebody actually goes off

Follow that same gap to its conclusion and it gets worse.

Of the people in that report who took time off and came back, 17% had a formal return-to-work or recovery plan. Sixteen per cent were offered a different working environment. Around one in ten, 11%, were offered regular check-ins with their manager about how they were doing.

Just over one in four got flexible or reduced hours, and that is the most common form of support anybody received.

Dame Carol Black made the same point from the other direction at this year’s Work and Health Summit. Employers routinely wait far too long before speaking to someone who is off sick, sometimes three months. By six weeks that person is already on a downward slope. By six months the way back is longer for everybody, including the employer.

None of that is expensive. A check-in is not a budget line. A phased return is not a capital project. What it takes is a manager who has been given the time and the words, which brings us back to the 29%.

What this looks like from a training room

This is the bit I see directly, so I will say what I see.

Organisations arrive having already done the awareness. They have run the campaigns. Some have trained champions. And they cannot work out why nothing has shifted, because on paper they have done everything the sector told them to.

What they usually have not done is give anybody the practical capability to sit in a difficult conversation, notice what a colleague is carrying, and respond without either panicking or handing over a leaflet. That is a learnable skill. It is not a personality trait and it is not a poster.

And it is not industry knowledge either, which is the request I get most often. Construction asks loudest. But knowing somebody’s sector was never the missing piece, because what is missing is the noticing and the response, and those look the same everywhere.

I would rather be honest about the limits of this. Training a manager well does not make a contract secure or a workload survivable. If the job is built to break people, a trained manager watches it happen with better language for it. I have written before that knowing your non-negotiables is not the same as being able to use them, and the same applies to organisations. Capability without conditions runs out.

But capability is the part inside an employer’s control this quarter, without waiting for legislation, and most of them have skipped straight past it to the posters.

Eighteen years is the polite number

The Work Foundation went back to Dame Carol Black this summer, eighteen years after her review of the health of Britain’s working-age population, and asked what had changed. Her answer was that the problems she identified in 2008 are still with us and the structural problems have not gone away. We have fiddled at the edges, she said. We have wonderful initiatives, so many of them, so many fragmented bits.

Eight Prime Ministers have been through Number 10 since that review. Nearly three million people are economically inactive because of long-term sickness. 6.8 million people are in severely insecure work.

And eighteen years is generous, because the things she described in 2008 did not begin in 2008. Insecure work, wages that do not cover a life, public services cut back, risk moved quietly off employers and onto individuals. Most of us have worked our whole lives inside it.

Which is why “delivery has fallen short” is the wrong reading. Falling short suggests we tried to build something and did not manage it. A fragmented initiative is what you fund when you do not intend to change anything. And a causes dataset left at 2012, with the GP reporting switched off in 2016, is what that looks like in a spreadsheet.

We are not short of numbers. We publish an excellent one every November, and it goes up.

What we stopped funding is the question that would tell us what to do about it. That was not an oversight, and it can be chosen differently.

Get new posts by email

New posts most weeks. No spam, unsubscribe any time.

Subscribe