Work can hurt a person in two ways. One is sudden: a fall from a ladder, a hand in a machine. The other is slow: dust breathed in every day for twenty years, a noise that wears the hearing down. The sudden one makes the news. The slow one kills more people.
Died of an illness from work2.6 million · 89%
Died in an accident at work0.33 million · 11%
Work-related deaths in a year, worldwide, 2019.
Counted across the world, about 2.9 million workers die every year because of their work. That is the count of the International Labour Organization, the UN's body for work, for 2019. Only about 330,000 died in an accident. The other 2.6 million died of an illness their work gave them.
Made ill by the work
vsvs
Hurt in an accident
Of the workers who die because of their work, which is the bigger group?
The two groups are about the same size.
Not yet. They are not close. Illness kills about eight times as many workers as accidents do, and most of it takes years to show.
Those hurt in an accident.
Not yet. Accidents are about one death in nine. They are the ones with a date, a place and a headline, which is why they feel like the bigger group.
Those made ill by their work.
Right. About nine in every ten. The accident is what people picture when they hear the words work and death, and it is the small slice.
126 a yearKilled in an accident
++
about 12,000 a yearDied of lung disease from work
++
1.9 millionIll because of work
Britain, latest counts.
Britain counts the same way. In the year to March 2026, 126 workers were killed in accidents. About 12,000 people a year die of lung disease from dust and fumes they breathed at work years before. And about 1.9 million working people have an illness their work caused or made worse.
Slide the marker to how many workers die each year across the whole world because of their work, in millions.
Slide the marker to your guess, then lock it. Nothing is scored.
123
you said · 0 m
actually · 2.9 m
0 m2.5 m5 m
1One million1 m
2Two and a half2.5 m
3Four million4 m
Workers who die each year because of their work, in millions: —
Nearly three million, and nine in ten of them from an illness rather than an accident. Guesses usually land low because the deaths people picture are the sudden ones, and the sudden ones are the small slice.
Dust breathed in→Years pass→Illness arrives→Not on the record
Slow harm is easy to miss. A fall has a day, a place and a name. A disease from dust arrives twenty years later, often after the worker has retired, and nobody writes it on the factory's record. So a workplace can look safe by its own count and still be making people ill.
Each of these is harm from work. Did it arrive in a moment, or build up over years?
One card at a time. Tap the pile it belongs to.
Card 1 of 6
A fall from a ladder.
Deafness after twenty years beside a stamping press.
A back worn down by years of lifting sacks.
A burn from a hot pan.
A lung cancer from asbestos dust.
A finger caught in a machine.
A fall from a ladder. → Sudden
Not yet. Sudden. The fall has a second it happened in, and the harm is all there by the time the worker is on the ground.
Deafness after twenty years beside a stamping press. → Slow
Not yet. Slow. No single day of noise took the hearing; every day took a little, and the loss showed only when most of it was gone.
A back worn down by years of lifting sacks. → Slow
Not yet. Slow. Each sack did a small amount; the worker felt nothing to report until the damage had built up for years.
A burn from a hot pan. → Sudden
Not yet. Sudden. The pan, the hand and the burn are all in one moment, and the kitchen can write it down that afternoon.
A lung cancer from asbestos dust. → Slow
Not yet. Slow, and the slowest of all. The dust was breathed decades before the illness showed, often at a job the person had long since left.
A finger caught in a machine. → Sudden
Not yet. Sudden. There is a day, a machine and a reason, which is exactly why sudden harm gets counted and slow harm does not.
Suddenhas a day
vsvs
Slowhas a cause, every day
Right. Sudden harm has a day. Slow harm has a cause that was there every day and a result that comes years later. Both come from the work. Only one of them is written down on the day.
The factory
↓→
The board400 days
↓→
Not on the board
A factory's board at the gate says: 400 days since the last accident. What does the board leave out?
Any harm still building up, like hearing loss or lung damage, which has no day to count from.
Right. The board counts what has a date. The dust in the air on all 400 of those days is doing its work and will never appear on it.
Nothing. 400 days without an accident means 400 days without harm.
Not yet. It means 400 days without a sudden harm. Slow harm does not stop because nobody fell, and it is the larger kind.
Accidents at the company's other factories.
Not yet. The board is about this factory, and that is fair. What it misses is not somewhere else; it is in this building, arriving slowly.
Warehouse0 accidents
vsvs
Office2 accidents
vsvs
Warehouse stafflifting all day
A company calls its warehouse its safest site: no accidents in three years, while the office had two. Warehouse staff lift heavy boxes all day. Which harm does the comparison miss?
Nothing. Fewer accidents means a safer site, by definition.
Not yet. Fewer accidents means fewer sudden harms. Whether the site is safer depends on the slow harm too, and lifting all day is where slow harm lives.
Back and joint damage building up slowly in the warehouse, which no accident count can see.
Right. Three years of daily lifting is three years of slow harm, and the count that calls the warehouse safest is the count that is blind to it.
The two office accidents, which should not count.
Not yet. They happened and they count. The problem is not that the office's harm was counted; it is that the warehouse's was not.
Lesson complete
Most people hurt by work are made ill slowly, and slow harm has no day to be counted on.