What John Snow knew

In the aftermath of the covid-19 pandemic, while the effects and the disruption caused by it are still fresh in our minds, it can be interesting to look at our response to the pandemic and compare it against earlier such events to find out things we did better and things that stayed the same.

Six Pandemics ago

A little over 200 years ago, the world dealt with a different pandemic. The cholera pandemic is believed to have started in India in the year 1817. In 1832 the virus reached London and the death toll began rising. The virus disproportionately affected the poorest areas of London where sanitation was relatively bad. At the time, we were still a good 60 years away from discovering viruses, and in the absence of solid science-backed solutions, most of the precautionary measures were close to guesswork.

At the time, there were two broad categories of beliefs around the disease.

• Theory #1 - The vast majority of Londoners strongly believed that cholera was caused due to God's vengeance against mankind.
• Theory #2 - A small section of the educated elite on the other hand believed in the then scientifically accepted Miasma theory. The Miasma theory stated that disease was spread through the inhalation of air from a poisonous and cloud-like miasma.

The proponents of Theory #2 were also the ones with the most influence and decision-making abilities. Acting on the Miasma theory they proposed distancing themselves from the poorer areas of London to avoid miasma as a solution to avoid Cholera. But since cholera wasn't an airborne virus, the solution didn't work and cholera outbreaks continued to cause a large number of deaths in the next 4 decades.

Enter John Snow

John Snow was a young physician born into humble beginnings in York in the year 1813. As a 14-year-old, Snow was a surgeon's apprentice and he eventually moved to London to gain a formal education in medicine. Snow, however, was unconvinced by the miasma theory. His logic was simple, if cholera was indeed caused by noxious air, then why were the lungs of those who were infected, completely unaffected. Why would it cause dehydration and diarrhea instead? Thinking along those lines Snow started working on a revolutionary theory. His research resulted in a pamphlet he published in 1849 called 'On the Mode of Communication of Cholera'.

John Snow's Theory

At the time, London received its drinking water supply from the river Thames, which was also where the city sewers dumped the cities waste. So according to Snow, Londoners were in essence consuming their waste, and Cholera was spreading through the contaminated water. Snow’s theory, as rational as it might seem in hindsight was just a theory. He didn’t have enough evidence to support it and despite his efforts, he couldn’t even get the parliament to discuss it because in addition to the missing empirical proofs, what his theory implied was disgusting. This added to the fact that the influential and educated elites of the time were still strong proponents of the miasma theory meant one thing, John Snow needed proof.

So John Snow started walking door to door, mapping out the cholera outbreak manually. And although as a medical professional he had access to more medical information about the outbreaks than a regular citizen would’ve had, he lacked local knowledge about the London neighborhoods he was trying to map. Given Cholera's nasty symptoms and the paranoia surrounding the disease, not everyone was willing to talk about it either.

Henry Whitehead

Do you know who 19th-century Londoners loved though? religious figures, enter Henry Whitehead, Henry was a vicar at St Lukes church in SoHo and a strong proponent of theory #1, that cholera was God’s vengeance. Over the last decade, Whitehead too had seen his fair share of Cholera induced deaths, and as a religious man, he found it hard to justify those deaths.

Whitehead believed that if Cholera was indeed the wrath of God, there had to be a way to alleviate God's anger. He had heard of John Snow’s theory, but without enough evidence he didn't accept it either, on the contrary, Whitehead believed that he could prove Snow wrong. Interestingly, in an attempt to prove Snow wrong, Whitehead and John Snow started working together to investigate the deaths. As they kept going over more and more cases, Whitehead soon realized that Snow’s insane theory made a lot of sense.

With Whitehead’s help, Snow was able to get more detailed information about the infections and his map was finally at a level where they could see the entire picture. Whitehead was also able to help Snow in finding patient zero in the case of a particular outbreak that happened in an area called Broad Street. Broad Street became the proving ground for Snow’s theory. Patient zero was a 5-month-old girl who contracted Cholera and died on the 2nd of September 1854 at the beginning of the epidemic. Whitehead found out that the girl’s mother washed her soiled clothes and emptied the water in a cesspool in front of her house from where the dirty water was believed to have leaked into the broad street pump that caused the outbreak.

As the investigations went on, using his map and Whitehead's familiarity with the area and the residents, Snow was able to find links between most nearby outbreaks and the Broad Street pump.

A case that I found particularly interesting was the Lion Brewery. The brewery was located in an infection hotspot, yet the 70 odd workers of the brewery never got sick. After talking to the workers, Snow discovered that the employees were given a daily allowance of beer in addition to their wages and they ended up not drinking any water at all.

The end of the pandemic

Even with all his data, Snow wasn’t able to convince the elites that the miasma theory was incorrect, however, with whitehead’s help, he presented his case to St James’s Vestry and was able to persuade them to disable the Broad Street pump by removing its handle. 200 years later, the pump ( minus the handle ) is still there and is a tourist attraction for history nerds.

John Snow passed away in June 1858 and sadly never saw his work vindicated. It took a lot of effort on Whitehead’s part, more outbreaks, and an incident since known as "The great stink" for people to finally recognize the validity of Snow’s theory that led to Bazalgette’s sewer systems in 1871 that finally stopped the cholera outbreaks.

”You and I may not live to see the day, and my name may be forgotten when it comes, but the time will arrive when great outbreaks of cholera will be things of the past; and it is the knowledge of the way in which the disease is propagated which will cause them to disappear.” - John Snow, in a letter to Henry Whitehead

Two centuries later

When you remove the historical context and the cool life-saving beer stories from the mix, John Snow's problem was the inability to access sensitive data to prove his theory to the parliament, if Henry Whitehead had not stepped into the picture, the pandemic could've claimed a much higher number of lives. When you think about it, there could've been many John Snows in history who had theories that could have had a significant impact on the rest of the world, but maybe they never found their Henry Whiteheads.

Over the last two centuries, we came up with multiple ways of using sensitive information about people to get answers to important questions without giving up the privacy of the people involved. While some of these technologies were already being used by tech giants such as Google (Federated Learning) and Apple (Differential Privacy) to analyze private user information in a privacy-preserving manner, Covid-19 contact tracing efforts caused many of these technologies to gain popularity. MIT's private kit & Safe Paths, in particular, is one of the most recent privacy-by-design based applications that used user's location data to perform contact tracing while ensuring their privacy.

In essence, we could enable John Snow to make the map and prove his point without needing a Henry Whitehead to talk to everyone who lived in Broad Street. During my last two years as a founding engineer at s20.ai, a lot of our work was centered on incorporating privacy-by-design principles into ML products. I plan on using this story as an anchor point to talk about some of the privacy-preserving data sharing techniques I'm familiar with, including:

• K-Anonymity
• Differential Privacy
• Private Join and Compute
• Federated Learning
• SGX based hardware enclaves

Follow me on twitter for updates on the future parts.