Dr. John Snow and the 1854 Soho Cholera Outbreak

London in the mid-19th century was a public health nightmare. Drinking water came from the Thames River, which was a sewer, or from shallow wells that were frequently fouled with trash or dead animals. Horses needed to transport goods into and around the city produced mounds of manure. When a horse died, it was unharnessed and left to lie where it fell until the dogs and maggots and crows reduced it to bones.

There was so much filth that it became a source of income. People made a living as “pure collectors” — scooping up fresh dog manure and hurrying it to tanners’ shops while it was still warm. Others combed the muddy shallows at low tide to gather up anything of slight worth that had fallen into the water.

Everything that made London vibrant and energetic and alive also made it putrid. Steam-powered looms and breweries belched smoke and coal ash into the air. In 1858, the Thames reached its filthiest in an event called “The Great Stink.” The river was cleaned up after that, but the problem of horse manure persisted until the 1890s, when city planners seriously considered abandoning London altogether. Motorcars, buses and motor lorries arrived just in time to save the city.

John Snow was a doctor in London in the mid 1800s, when the city was foul and stinking, the water was dirty, and people often died of various contagions. An outbreak of cholera in Soho in 1854 was just one flare-up in a persistent world-wide cholera epidemic that lasted for decades. But Snow did something remarkable.

Cholera is a real winner, disease-wise. An infected person is gripped with sudden, intense diarrhea and vomiting leading to dehydration. Death follows in a few hours. Part of the terror cholera held in the 1800s was uncertainty about how the disease was transmitted. Some thought it was a punishment for sin. Others blamed rats or other nasty things. The prevalent theory was that bad air or miasma caused it, and there were a few who held a new theory that invisible germs were to blame. The germ theorists didn’t carry much weight until a few years later when Louis Pasteur proved that germs are real.

Snow did not cure cholera or bring the disease to an end. His great achievement was more limited, but still impressive from the perspective of science and epistemology. Snow identified one particular source of disease and prescribed an effective solution to it. It was a small victory for public health, but a huge achievement for the art of discovery and truth-finding. It was the start of the whole field of epidemiology.

Snow didn’t pull a Sherlock Holmes-like stunt of detached reasoning based on superhuman deduction. Rather, he carefully collected and analyzed a mountain of data about where people got their water and how they stored and handled it. It amounted, in Snow’s careful mind, to a massive experiment with everybody in London taking part:

The experiment was on the grandest scale. No fewer than three hundred thousand people of both sexes, of every age and occupation, and of every rank and station, from gentlefolks down to the very poor, were divided into two groups without their choice, and, in most cases, without their knowledge; one group being supplied water containing the sewage of London, and amongst it, whatever might have come from the cholera patients, the other group having water quite free from such impurity.

In early September there was a sudden spike in cholera deaths. People were dying all over London. But Snow connected many of the victims to the Soho neighborhood. This was not a simple or obvious observation. Many of the sick had sought a change of air or gone to stay with relatives when they got sick. Others fled the neighborhood as soon as they heard there was an outbreak, taking the infection with them.

But Snow sifted his data, asked follow-up questions, and determined that the victims were mainly from Soho. More especially, they were all people who had drank water from a particular well on Broad Street. The victims were rich and poor, old and young, virtuous and dissolute. The only thing they had in common was their water source. Snow also noticed that workers at the Broad Street brewery seldom got sick. They breathed the same air and walked the same streets as the victims. But they drank their own beer, and never from the pump.

Snow concluded that the Broad Street water pump was infecting the people who drank from it. Later analyses suggested the well may have been polluted when the diapers of a diseased baby were washed and rinsed at the well. In any case, Snow realized that if people stopped using the well, they would stop getting sick from it. He reported this to the town officers, and they disabled the pump.

Popular reports on this incident say Snow removed the handle from the Broad Street water pump and the epidemic immediately halted. That version actually diminishes Snow’s accomplishment, because cholera deaths (caused by other sources) continued all over London while Snow was theorizing and even after the pump handle was removed. The number of deaths in the immediate neighborhood fell before the handle was removed, because people scattered. It took some sophisticated focus and rigorous effort on Snow’s part to show how the Broad Street pump related to the much larger scope of disease and death.

Abraham Wald and the Bomber Planes

This story is the best known of the three so I’ll relate it quickly.

During World War II, America’s Air Corps was losing too many planes while bombing targets in Germany. A crashed plane was a costly material loss, but also an unbearable human loss. The Air Corps demanded action. An order was given to strengthen the bombers with heavy armor plating. But a plane that was armor-plated all over would be too heavy to fly. The order was given to put the armor plating only on the parts of the plane that needed to be protected.

Inspection of planes damaged in recent bombing runs showed that the heaviest damage was on the wings, fuselage and tail sections. So the order was given to put the extra plating on the wings, fuselage and tail sections. Bombing resumed, and the casualty rates were worse than before. Reinforcing the heavily damaged parts of the planes made the planes harder to fly and didn’t protect them from enemy fire.

This was when Abraham Wald stepped in. He was a member of the science team assigned to fix the planes. Wald noted that their evidence was from planes that survived and landed safely back in England. The planners had no information about the really critical damage to planes, because all the planes they needed to study had crashed over Europe or in the North Sea. The data they were gathering from the surviving planes was not showing what was critical, but what was not critical. So Wald recommended something counter-intuitive. He said the Air Corps should reinforce the parts of the planes where damage was never observed. They did that, and losses immediately went down.

Wald is described as a math genius. And perhaps he was. But it is a mistake to ascribe this success to math. Math was leading to the wrong conclusion. Wald corrected the mistake by simple observation and intuition. He recognized that all the evidence they could observe by looking at planes that survived could not help them understand the planes that crashed.

Wald is credited with discovering, or at least dramatizing the effect of, survivorship bias. That principle is important in a lot of research. I used to work at a university and was often asked to report on circumstances that helped students graduate on time or succeed in other ways. I knew the results would be very different if I looked only at graduates (100% of whom had succeeded) than if I included the students who dropped out along the way. The challenge of coping with survivorship bias is an important consideration for all researchers.

Thank you, Abraham Wald!

Andrew Degbe and the BeePee

In the mid-1980s, Andrew Degbe was one of six young African men trained to work with a US-funded rural development initiative, the Liberian Rural Communication Network. I was a Peace Corps Volunteer assigned to support them.

Our work involved visiting small villages in the West African rain forest and presenting information to the people living there about primary health, education, foreign aid, and other topics. We had to travel to those villages from our base in Zwedru, and we got around riding motorbikes.

To ensure gasoline would always be available for these travels the team bought volumes of gasoline and stored it in 55-gallon drums. Arthur Gboryu was tasked with dispensing the fuel and keeping it secure. He kept the drums in a locked shed. He dispensed the fuel gallon by gallon with a green plastic jug. He recorded in a notebook exactly how much fuel each team member used and how much remained in the drum.

In tribal cultures like in Liberia, there is always pressure to help a family member — and everyone of your tribe is family. It isn’t seen as stealing, but as doing the right thing. So the manager (Mr. Juasemai) was putting heavy responsibility on Arthur. He chose Arthur because he was not a local guy and he had the fewest family or tribal connections in the area. Arthur was conscientious with the responsibility.

And yet a day came when the drums were empty though the notebook showed that several team members still had fuel owing to them. Arthur fell under suspicion: either he miscalculated the usage or he was stealing fuel on the side.

Mr. Juasemai asked me to figure out what had happened. He gave me the job because I was an outsider with no tribal loyalties, and because I hadn’t joined in on the gas-sharing scheme. (I liked talking with the gasoline vendors around town and didn’t mind giving money to the Muslim who had a corner on the fuel supply as much as my colleagues did.)

Arthur was adamant that he’d done nothing wrong. I wanted to believe him. And yet the accounts testified against him. The book said there should be 30-odd gallons still in the empty drums.

Enter Andrew Wlekiomun Degbe (pronounced DEB-weh). He was the youngest and, to be honest, not the brightest member of our team. His communication work was less incisive than the others’ and he made more mistakes in his translations. But he was clever enough to notice the key detail that solved this mystery.

One of the variations that occurs from culture to culture is the choice whether or not to identify individual things into groups or categories. Eskimos are said to have 20 or more words for varieties of frozen precipitation but no general word for snow. The Liberians in the 80s would refer to the orange carbonated beverage as “Fanta” and to its purple sister product as “Grapfanta.” There was no categorical word for pop or soda.

And so it was no surprise when I overheard Degbe say, during a discussion about the missing fuel, that Arthur’s dispensing jug was not a gallon but a “beepee.” A gallon was yellow plastic and had an embossed shell on it. Arthur’s jug was green: an altogether different thing.

And Degbe was right.

West Africa at that time was a mixing pot of foreign influences. There was a massive Firestone rubber plantation run by Americans, a big Spanish logging concern, and foreign aid programs run by many different countries. In the two years we lived in Zwedru, we encountered a wry Englishman, several surly Germans and Danes, a couple of lewd Italians, several cheerful Japanese and Lebanese, two engaging doctors from India and a demure Iranian.

With that diversity of influence, the systems were mixed. The English system and the metric system bounced up against each other frequently. The yellow “gallon” was American. It had a capacity of one gallon, which is 3.78541 liters. The green “beepee” had originally dispensed four liters of British Petroleum motor oil. Four liters is 1.05669 gallons. And there was the answer to the mystery.

Each time Arthur had dispensed fuel to a team mate, he had innocently added an extra 5.669% to the gallon. The drums were filled in the capital city of Monrovia under US-AID supervision, and measured in proper gallons. But they were dispensed up-country with the green four-liter jug that came easily to hand.

The discrepancy wasn’t enough for anyone to notice, and none of the guys on the team had thought the difference mattered. But over time it created the shortfall. I did the math and found that the slight error multiplied by the number of units Arthur had dispensed over several months time amounted to almost exactly the deficit we were trying to explain. The remaining disparity was probably from evaporation and spillage.

I honestly don’t know if Degbe’s remark was meant only to distinguish a yellow container from a green one, or if he comprehended the metric-to-English discrepancy. I just know that as soon as I heard him say “That’s not a gallon. That is a beepee” I knew the solution. I want him to get the credit. I would be delighted if A. Wlekeomun Degbe were added to the list of the world’s great thinkers. As I said, I don’t think he was especially bright, but there are people sitting in endowed faculty chairs and high elected offices who are no smarter than him.

Epistemology is the search for justified true belief. Each of these three stories relates a time when a correct conclusion was drawn from available evidence and correct action followed. A little bit of math came in useful. But the main ingredient in all three cases was careful thought built upon clear definitions and consistent measurement. Those three elements — and not computers and software — are the essentials to knowing what’s true.