Data Science in Politics of Yellow Fever: Medical Research before “Data”

By Tannaz Motevalli, Sarah Eilers, Laura Hartman, and Erika Mills

In the exhibition Politics of Yellow Fever in Alexander Hamilton’s America includes a variety of items, selected by the curator from the NLM historical collections, which reflect how scientists and the public hypothesized the cause and spread of yellow fever throughout U.S. cities during the 18th and 19th centuries. During the 1793 outbreak in Philadelphia, science and medicine were largely based on experience, observation, and hypothesis, in contrast to contemporary data-driven research. The National Library of Medicine’s Strategic Plan 2017–2027 charts a pathway for NLM to embrace and pursue data science as a means to address existing and emerging challenges in biomedical research and public health through data science. And with that, a corollary: how can we use the language, concepts, and framework of data science to comprehend how scientists of the past used data before the concept of data science emerged? Politics of Yellow Fever in Alexander Hamilton’s America will serve as our case study.

The overarching concepts of data science are not new—examples abound in history of researchers sharing, analyzing, and processing knowledge, information, and data. The ways these processes are undertaken have shifted, as have the priorities informing those actions, and the language used to describe them. One helpful distinction is the difference between data and information. Data means the facts and figures gathered in research, which alone and without context are relatively useless. But when data is processed and analyzed it becomes actionable information. Items in this exhibition, therefore, are not merely examples of data or data sets; each item represents a piece of information developed from data by early American physicians, epidemiologists, authors, and others.

The Growing Currency of Data Through Yellow Fever Epidemics

Around the end of the 18th century, a variety of formal and informal sources—scientific publications, medical organization meeting notes, newspaper articles, and personal correspondence—largely captured information about yellow fever epidemics and related research findings. These records combined facts and figures about the disease with observations and opinions. By the time the U.S. Army Yellow Fever Commission confirmed that yellow fever was transmitted by mosquitoes in 1901 (a theory first proposed by Dr. Carlos J. Finlay, a Cuban physician, in 1881 at the International Sanitary Conference), public health research and record-keeping around the disease had undergone a change, becoming more rigorous and data-driven, and reflecting an expanded understanding of yellow fever. The items showcased in the exhibition—public health records and research findings from the 18th and 19th century—illustrate this transition.

Benjamin Rush, a signer of the Declaration of Independence, a physician, and a prominent player in the debate about the origins of and best treatment for yellow fever during the 1793 Philadelphia epidemic, wrote a letter to his sister detailing the devastation the disease had wrought on the city. He describes deserted streets, 190 fresh graves in the Catholic Church yard, and 100 people having succumbed to the disease in the previous week. Rush would go on to write books and newspaper articles about the epidemic that featured personal accounts and his medical views, as well as some figures on infection and death. Looking back on his works now, his scholarship is particularly important as it provides public health information about the 1793 epidemic at a time when health record-keeping wasn’t rigorous.

Lexicographer Noah Webster, famous for his work on the American Dictionary, conducted an epidemiological study of the 1793 yellow fever outbreak in Philadelphia and published his findings in a book that included infection and death statistics. In an article in The Spectator newspaper, he starts with a few statistics from his research, then zeroes in on his conclusions about the disease and a discussion of epidemiological methods. Here, public health data is presented alongside Webster’s observations and inferences.

By the 19th century, some scientists believed the climate had an impact on the development and spread of yellow fever. This chart appeared in a report published by the Sanitary Commission of New Orleans, which had been tasked with finding causal factors and prevention methods in a city that had been hit by yellow fever multiple times throughout the mid-1850s. Printed alongside discussions on public health and climatic data in tables, the chart represents what we now call data visualization.

Coming to the Wrong Conclusions

Can research that is data-driven still bring us to the wrong conclusions? How do we ensure accurate data analysis?

“Yellow Fever: a Map of Old Slip and infected Vicinity” shows the incidence of yellow fever cases in New York in 1819. The numbers on each house refer to a detailed case report contained in the accompanying work A Statement of the Occurrences During a Malignant Yellow Fever in the City of New-York. The author, physician Felix Pascalis Ouviere, analyzed this map and concluded that the randomness of the cases indicated that the air was to blame. Like many physicians of his time, it never occurred to him that mosquitos flying through the air might be carrying the contagion.

The idea of bad or impure air causing disease is known as the miasma theory. Miasma is a Greek word meaning polluted. The miasma theory of disease says that rotting organic matter gives off fumes or pollutes the air, creating unhealthy conditions and illness.

Even though public health officials in the 19th century drew incorrect conclusions from the disease incidence maps, they galvanized to clean up urban areas. This 1878 the Sanitary Commission of New Orleans drew this map to identify “public nuisances” in the city. “Public nuisance” is a legal term than means something that threatens the good of society. Thus, public health officials considered anything believed to contribute to miasma as a public nuisance including: boarding houses, cemeteries, factories, slaughterhouses, stables, and even construction sites, because they were disturbing the ground where rotting organic material might lie hidden. The 19th century saw widespread public health initiatives to improve city infrastructure for sewage removal, and extensive legislation for disposing of dead bodies and managing refuse from factories, markets, and slaughterhouses, which improved the well-being of residents but didn’t mitigate outbreaks of yellow fever.

Today, increasingly large and complex data sets motivate researchers in all subjects to gain the skills that data science can provide to ensure accurate and ethical, and effective use of data. Critical examination of a variety of historical records reveals how the influx of available and evolving data created obstacles for researchers of the time. Through this perspective, we see how the past echoes our current challenges in data-driven medical research. In the second part of this series, we discuss the use and influence on data after the discovery of the true cause of yellow fever.

Stay tuned for part two—Data Science in Politics of Yellow Fever: Discovering the Cause of Yellow Fever

Explore the online exhibition Politics of Yellow Fever in Alexander Hamilton’s America.

Tannaz Motevalli is a researcher and exhibition coordinator for the Exhibition Program in the History of Medicine Division at the National Library of Medicine.

Sarah Eilers is the Manager of Historical Audiovisuals in the History of Medicine Division at the National Library of Medicine.

Laura Hartman is Rare Book Cataloger in the History of Medicine Division at the National Library of Medicine.

Erika Mills is outreach coordinator for the Exhibition Program in the History of Medicine Division at the National Library of Medicine.