On this page
Today, washing hands between patients is routine. In 1840s Vienna, doctors and students went from autopsies to examining women in labour. Ignaz Semmelweis was the doctor who showed, with a careful comparison of two clinics, that this was killing mothers. This post covers what he found, works through his numbers, and asks what a student can take from the way he reasoned.
Who was Ignaz Semmelweis?
According to a historical review in GMS Hygiene and Infection Control (Ravel and colleagues, 2025), Semmelweis was born on 1 July 1818 in the Tabán district of what is now Budapest. He studied medicine, graduated in 1844, and turned to obstetrics. On 1 July 1846 he became assistant to Johann Klein in the maternity division of the Vienna General Hospital. His duties included examining patients each morning, teaching students and keeping the clinic's records. That last job matters: it gave him the numbers.
The hospital had two obstetric clinics. From 1840, medical students and junior doctors were taught in the First Clinic, and midwives in the Second. Semmelweis saw that deaths from puerperal fever (childbed fever, an infection after childbirth) in the First Clinic were far higher than in the Second.
How did he rule out the easy explanations?
Prevailing theories blamed "bad air" or "miasmas", the GMS review notes. Semmelweis tested the ideas he could test, one at a time, as the GMS review describes:
Birth position. Women in the Second Clinic delivered lying on their sides, those in the First on their backs. He changed the position in the First Clinic. Deaths did not fall.
Fear. A priest ringing a bell walked through the First Clinic's wards on the way to dying patients. Semmelweis had him change route and stop the bell. No effect.
Then his colleague Jakob Kolletschka died of septicaemia from a finger injury picked up during a post-mortem examination, as the AJPH article below describes. Semmelweis reasoned that doctors and students, who did autopsies, were carrying "cadaverous particles" on their hands to the women they examined. Midwives did not do autopsies. As the American Journal of Public Health (Lane, Blum and Fee, 2010) puts it, he began to insist that anyone attending autopsies scrub their hands in chloride of lime before entering the maternity wards.
What do his numbers show? A worked example
The GMS review gives the clinic figures from Semmelweis's own publication. Handwashing began in the First Clinic around mid-May 1847, so compare 1846 (before) with 1848 (a full year after).
Step 1: Turn counts into rates. A rate is deaths divided by births.
First Clinic, 1846: 459 deaths out of 4,010 mothers. 459 ÷ 4,010 = 0.1145, about 11.4%.
Second Clinic, 1846: 105 out of 3,754. 105 ÷ 3,754 = 0.0280, about 2.8% (the review rounds it to 2.7%).
Step 2: Compare the clinics. 11.4 ÷ 2.8 ≈ 4.1. A mother in the doctors' clinic was about four times as likely to die.
Step 3: Compare before and after. First Clinic, 1848: 45 out of 3,556. 45 ÷ 3,556 = 0.0127, about 1.3%. The Second Clinic that year was also about 1.3% (43 out of 3,219). The gap had closed.
Step 4: Find the relative fall. 1 − (1.27 ÷ 11.45) ≈ 0.89, a fall of about 89% in the First Clinic's death rate. The review describes it as a drop of 90%.
Step 5: Estimate lives saved. At the 1846 rate, 3,556 births would have meant about 0.1145 × 3,556 ≈ 407 deaths. There were 45, so roughly 360 fewer. This is a rough estimate, since other things can change between years, but Stang, Standl and Poole (2022) note that Semmelweis himself reasoned this way, estimating the deaths an intervention prevents.
The common mistake is to compare raw death counts. The two clinics had different numbers of births, so only rates can be compared fairly. Our post on common probability mistakes covers the same trap.
Why was he ignored?
His results were strong, yet many obstetricians rejected them. The GMS review gives several reasons. His idea implied that doctors themselves were spreading the disease. Germ theory had not yet arrived; Louis Pasteur's work came later, as our profile of Pasteur and germ theory explains. And Semmelweis had what he called an "innate aversion" to writing: his friend Ferdinand von Hebra reported the findings first, in December 1847, and his own book, Die Ätiologie, der Begriff und die Prophylaxis des Kindbettfiebers, appeared only in 1861.
When the book was ignored or attacked, Lane and colleagues write, he began to denounce leading obstetricians publicly. In 1865 he was committed to an asylum. The GMS review says he died of sepsis on 13 August that year, aged 47; Lane and colleagues trace it to a gangrenous wound to his hand. Looking back with modern methods, Stang and colleagues found that he had already handled problems such as confounding and misclassified outcomes, and that his data "provide a great pool for illustrating the logic of scientific discovery".
What can a student take from how he worked?
Keep records. His breakthrough started with the clinic ledger. A mistake log for your mock tests does the same job.
Compare like with like. He compared rates, not totals, and two clinics in the same hospital. When you judge a study method, compare your scores on similar papers.
Test one idea at a time. Birth position, then the bell: each was changed and checked. Change one thing in your routine at a time so you know what worked.
Explain as well as you test. His slow, angry writing let others dismiss correct results. Being right is not enough; you also have to show your working clearly.
If biology ideas such as infection and immunity feel thin, the Fermi Concept Library covers every concept in physics, chemistry, biology and mathematics, each with its own lesson.
Redo Step 1 to Step 4 with the Second Clinic's figures for 1846 and 1848, and see how much its rate changed.
