He Had The Numbers In 1847. He Published In 1861.
Mortality in his own ward fell from 11.4 per cent to under two in a year. He did not write it up for fourteen years, and when he did he called his readers murderers. Both halves of that killed women.
A months-long reporting project. Documents cited below are held in The Vault and available to readers.
VIENNA — In 1846, in the doctors' ward of a Vienna hospital, 459 women of 4,010 died of childbed fever. That is 11.4 per cent.
In the midwives' ward, across the corridor, taking women on alternating days: 105 of 3,754. Two point seven.
In May 1847 he put a basin of chlorinated lime at the door and made his students wash before they touched a patient. Within a year the doctors' ward was under two per cent.
He published in 1861.
What the two wards differed by
One thing.
The doctors' ward was staffed by physicians and students, who arrived in the morning from the dissecting room. The midwives' ward was staffed by midwives, who did not.
Everything else — the building, the beds, the weather, the women, the days of the week — was the same or alternated.
This desk has read a great many natural experiments and there are not many cleaner than that one.
The fourteen years
He had the effect by the end of 1847 and the confirmation by the end of 1848.
Colleagues asked him to write it up. We have dated every recorded occasion on which somebody did. Others published on his behalf, from his lectures and his letters, which is not the same thing and which he knew was not the same thing.
He did not write it up until 1861.
What was happening in those fourteen years is partly documented and partly not: illness, grief, a move to another country, and a difficulty with writing that people around him described at the time. Which of those was decisive, nobody can say, and the accounts that claim to are reconstructions.
What he published
The book is right about everything.
It is also, from its opening pages, an accusation. It tells the men reading it that they are murderers. When the reviews were unfavourable he answered them with open letters, published as pamphlets across 1861 and 1862, and those letters are the hardest reading in this story.
Our correspondent read them in full over two evenings and would like to record that they are the writing of a man who has watched women die for fourteen years knowing why, and that this is an explanation and not a defence.
They did not persuade anybody. They were not built to.
What the profession did
Behaved disgracefully, and this newspaper is not going to soften it to make room for the other half.
He was right, he was demonstrably right, the numbers were available to anybody who wanted them, and he was mocked, dismissed, excluded, and eventually committed. He died in 1865, at forty-seven, of an infected wound on his right hand.
Both halves of this story killed women. That is the sentence this desk has been circling for three weeks and it is the only honest way to hold it.
What the case is used for now
Dr. Anselm Rutherford-Diallo works in infection control and hears this story invoked roughly once a month.
Almost always, he says, by somebody who has not published their own data either.
"It is the most flattering story in medicine," he told us. "Everybody who tells it is Semmelweis. Nobody who tells it is the committee."
The ending our historian wanted instead
Professor Ottavia Lindqvist-Bassong's full reply runs at the foot of this page and it is better than our ending, so we are pointing at it.
Her argument: in 1847 he could not say why it worked, because there was no germ theory to appeal to. He had an intervention, a gesture at a mechanism, and a mountain of numbers — and what happened to him is what reliably happens to that combination. The numbers were not enough.
That is not a nineteenth-century failing. It is a permanent one, and she says it is still true in her own hospital.
Then she asks the question this desk cannot answer and is printing anyway: name the intervention sitting in front of us right now with good numbers and no mechanism.
She can name two. Neither is in general use. Neither has anybody obviously to blame.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
Sources & Method
We read the hospital's own monthly returns rather than the figures as they appear in later accounts, because effect sizes shrink on the way through a summary and this desk has a rule about it. Then we read the 1861 book and the open letters in full, which almost nobody does, and dated every recorded occasion on which a colleague asked him to write his results up.
Who we spoke to
- The hospital's monthly mortality returns, 1841–1849, Institutional record. Read in facsimile; the figures in this story are taken from them and not from later summaries 2026 Effect sizes shrink on the way through a summary. This desk goes to the returns.
- The 1861 book and the open letters of 1861–62, Published. Read in translation, in full, including the letters 2026 The letters are the hardest reading in this story and are the reason the piece exists.
- Prof. Ottavia Lindqvist-Bassong, Historian of medicine. Interviewed by video call three times April–June 2026 Made this desk take the germ-theory chronology seriously: in 1847 there was no mechanism to offer, and that matters more than we had allowed.
- Dr. Anselm Rutherford-Diallo, Infection control physician. Interviewed on the record about what the case is used for now May 2026 Says the story is invoked in his field roughly monthly, almost always by somebody who has not published their own data either.
Documents
- PX-2119 — Monthly mortality returns, two obstetric clinics of a Vienna hospital, 1841–1849 accepted
What we could not confirm
- What he was thinking in those fourteen years. He did not say, and the accounts that claim to know are reconstructions. Illness, grief, a move to another country and a documented difficulty with writing are all in the record; which of them was decisive is not.
- How many died as a result of the delay. The counterfactual requires assumptions about adoption that nobody can supply, and any figure you have seen is a construction.
- Whether earlier publication would have been believed. Our historian's view is that without a mechanism it might well not have been, and she asked us to print that this is her judgement rather than a finding.
How Others Covered This
The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.
-
ApertureThe Asylum
A piece about how he died — the committal, the beating, the wound — and what it says about nineteenth-century medicine's treatment of its own.
Had that we did not
The end of his life, in detail we could not match and did not attempt.
Left out
The fourteen years.
-
The Hollow PostTHEY DESTROYED HIM FOR TELLING THE TRUTH
Reported the case as a pure suppression: a man with proof, silenced by a profession protecting itself.
Had that we did not
That he was right, that he was treated appallingly, and that women died who need not have. All true and all important.
Left out
That he did not publish for fourteen years; that colleagues asked him to and he did not; and that his eventual book opens by accusing its readers of murder.
-
They Buriedthis newspaperHe Had The Numbers In 1847. He Published In 1861.
Held both facts at once — that the profession behaved disgracefully, and that he made it as hard as he possibly could.
Had that we did not
The monthly returns, and the dates of every request that he write it up.
Left out
Our headline puts the burden on him. A newspaper choosing which half of a story goes in the headline is choosing who the reader blames, and we chose. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Prof. Ottavia Lindqvist-Bassong, historian of medicine read the full draft; no deadline. Asked to reply at length because she thinks the piece is fair and will still be misread. Printed unedited.
I have no complaint about your facts and one about your shape.
You have written a piece that says: he was right, they were cruel, and he did not publish. All three are true. The trouble is that a reader finishing it will take away a lesson about him, and the lesson worth taking is about us.
Here is the thing I would put in place of your ending.
In 1847 he could not say why it worked. There was no germ theory to appeal to. He had an intervention, a mechanism he could only gesture at, and a mountain of numbers. And what happened to him is what happens, reliably, to that combination: the numbers were not enough.
That is not a nineteenth-century failing. It is a permanent one. A profession that will not act on an intervention until it has a mechanism will always be slower than the deaths, and it is still true in my hospital this year.
So when you write that he waited fourteen years, please also write that the fourteen years were not the only delay in the story. There was the delay before he wrote, and there was the delay after — decades of it — and only the first one is his.
And I would ask your readers who feel the familiar warmth of hindsight to name the intervention currently sitting in front of them with good numbers and no mechanism. I can name two. Neither is in general use. Neither has anybody obviously to blame.
Published unedited under our right-of-reply guarantee.
How was this story?
We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.
Readers' Letters 0
Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.