The Topography of Disease

In 1831, cholera hit the United Kingdom and began killing tens of thousands during successive waves of terrifying outbreaks. The disease was fatal for as many as a quarter of its victims, often within just days of exposure. Many scientists of the day embraced the theory that epidemics were caused by the foul air, or miasma (from open cesspools, raw sewage and rotting rubbish), that hung over large swaths of most major cities. That is, until a young English doctor named John Snow documented an outbreak of cholera in his London neighborhood in 1854.

As the story goes, Snow carefully mapped the locations of the victims’ homes and demonstrated that the deaths were clustered around a public water pump on Broad Street in the Soho district. By interviewing the victims’ families, he was able to trace nearly every case back to that water pump, bolstering his theory that cholera is a waterborne disease and convincing the local authorities to remove the pump’s handle. The story has become legendary, and Snow’s map is often portrayed as a breakthrough moment in both cartography and epidemiology.

The truth, however, is much messier. Snow did indeed do excellent work that helped advance the science, and his map still rightfully stands as a shining example of medical cartography. But it wasn’t until long after Snow’s death in 1858 that his theory was proved correct and his work was hailed as a turning point. “His map has become an icon and Snow himself an almost mythic figure,” writes medical geographer Tom Koch in Cartographies of Disease. “Few focus, therefore, on Snow’s failure to convince his contemporaries of his argument, the limits of his thesis in the context of his time.”

And Snow wasn’t the only one mapping cholera in England in 1854. The physician Henry Wentworth Acland tracked an outbreak in Oxford that affected 290 people that year. His work, which resulted in the 170-page Memoir on the Cholera at Oxford and an accompanying map, was “perhaps the most comprehensive study of an urban disease of its day,” according to Koch.

Henry Wentworth Acland’s map of the cholera outbreak in Oxford, made in the same year as Snow’s. Credit: Princeton University Library

Some experts found Acland’s work more convincing than Snow’s, partly because of its breadth and thoroughness. But Acland’s research had another big advantage: Its conclusions supported the prevailing miasmatic theory of disease, which had been developed over centuries. Snow, on the other hand, was bucking the mainstream with his waterborne-disease theory.

While Snow’s analysis was focused on one possible explanation for the outbreak, and his argument rested on the visual clarity of his map, Acland took a more statistical approach that considered many potential disease factors. In addition to mapping victims, Acland included sites that had previously been deemed unhealthy (brown dots), those that had subsequently been cleaned up (brown circles), streams that were unpolluted, and those that had been contaminated (dashed lines), including point sources of the contamination such as outflows of raw sewage (see close-up below). Areas with poor drainage were shaded green.

Detail from Acland’s map. Credit: Princeton University Library

Snow was content to stop mapping the cholera deaths that occurred after he thought his case had already been made. By contrast, Acland mapped the entire list of victims in 1854, as well as those of two previous outbreaks. He used different symbols for the locations of victims’ homes from 1832 (blue dots), 1849 (blue bars) and 1854 (black squares and bars). And, most important for his argument, Acland mapped the physical topography of the town with five-foot (1.5 meter) contour lines. His map, together with his statistical analysis, showed a clear correlation between elevation and the disease. In each of the three outbreaks, people in low-lying areas suffered a much higher rate of infection and death. Even the higher spots that had unhealthy brown dots fared better than the lowlands.

Acland’s map neatly backed up the miasmatic theory, suggesting that the toxic air would collect and remain in low areas with less wind. “His statistics are showing that he had an excellent argument and evidence,” says Este Geraghty, chief medical officer for the mapping software company Esri. But Acland failed to see the whole picture, making his a cautionary tale, she says. “You have to determine what things mean, not just the outcome and the correlation.”

Acland saw polluted water as a potential contributor to pestilent air, not as a medium for the spread of an invisible agent of disease. Consequently, he hadn’t paid attention to sources of drinking water. Instead, like many of his contemporaries, he looked to the weather for clues to how elevation could be influencing the disease.

“That there is a connection between the state of the Atmosphere, or of the imponderable agents of the globe, and the existence of the Epidemic, is scarcely doubted by those who have carefully attended to its history,” he wrote.

Acland meticulously charted the time line of the 1854 outbreak against a host of local climate variables, including temperature, barometric pressure, wind, rain, humidity, cloud cover and ozone levels (below). But he was unable to find anything that waxed and waned precisely with the number of new cholera cases. What Acland was able to demonstrate, with copious data, is that 1854 was an abnormal year for weather in Oxford. Comparing it with the 25 years prior, he found that rain was abnormally low, as was wind speed. The list of things that were abnormally high included temperature range, pressure, thunder and lightning, days with hail, and appearances of the northern lights.

Acland’s chart showing weather variables he speculated might have contributed to the Oxford outbreatk. Credit: Princeton University Library

He couldn’t quite put any of these variables together with elevation to form a reasonable explanation, especially considering that the previous outbreaks didn’t follow the same pattern. But Acland was still confident that if another epidemic were to occur, “the rapidly advancing science of Meteorology” would be able to use data to clarify which of the abnormalities played a role.

Acland’s study was more comprehensive, and, at the time, more convincing than Snow’s, but it had one glaring flaw: His conclusion was definitively wrong. “Like much of the science of every era, it missed an intervening vector,” Koch writes. What Acland failed to see was that at higher elevations, water typically came from wells or streams, while lower areas mostly relied on rivers that were often polluted with sewage. But, Koch notes, “the mapping—and here, the mapmaking—were clear, consistent, and, if ultimately incorrect, still rigorous.”

Snow’s theory was largely dismissed during his lifetime, but he ultimately triumphed long after his death. Conversely, Acland enjoyed recognition from his peers for his work on cholera but lived to see the theory it supported disproved. Still, his work is worth more than a footnote, argues data visualization journalist Alberto Cairo.

“The myth of the hero who singlehandedly wrecked miasmatic theory obscures the fact that those who held onto it were also thoughtful fellows,” Cairo writes. “It is unfortunate that we don’t study them further, as we humans learn much more from our mistakes—both individual and collective—than from our successes.”

Published by: https://blogs.scientificamerican.com/observations/the-topography-of-disease/