A Doctor’s Influence during an Influenza Epidemic

In December 1889, as the Russian influenza appeared poised to reach the United States, Dr. Roberts Bartholow published an editorial in Medical News, a journal published in Philadelphia, in which he defined influenza, predicted the likely pattern of transmission, and offered recommendations for treatment. The content of this editorial, and the ways in which it was reprinted, summarized, and even contested in newspapers across the United States was the focus on a paper presented at the American Association for the History of Medicine in April 2015 and subsequently published in the journal, Medical History. This article used traditional humanities tools – close reading of texts – in combination with search capacities made possible by digitized newspaper collections to interpret the ways in which this doctor’s editorial shaped reporting on the influenza across the United States.

Asking about Bartholow’s influence during an influenza epidemic allows the historian to pursue a relevent epidemiological question: During a disease outbreak, what factors shape the opinions of medical experts and the general public regarding the scope and severity of a disease? Including digitized sources and digital humanities tools allows historians to pursue this question in new ways that suggest innovative perspectives and original insights—as well as the need for continued research and interpretation.

Map of Locations Reprinting or Referencing Bartholow’s editorial, created by the author, using google analytical tools. Online version: https://drive.google.com/a/vt.edu/open?id=10E7qipZXhzkOqe__VQfEpusnF68&usp=sharing

Map of Locations Reprinting or Referencing Bartholow’s editorial, created by the author, using google analytical tools.
Online version

Illustration 2 is a map created using a spreadsheet and the google mapping tool that identifies the locations of newspapers that reported on Bartholow’s editorial in Medical Age. The map shows the broad geographic range in which this editorial was noted by a local newspaper. In the online map, clicking on each point shows the title of the newspaper, the date, and a link to the original article, if it is available in a public database.

Illustration 3 provides a different approach to visualizing information in order to focus more specifically on the possible causal effects of one doctor’s intervention. One of the specific recommendations made by Bartholow was the use of antipyrine to treat headaches that accompanied influenza. Figure 3 traces the number of pages in the digital newspaper collection, Chronicling America, in which antipyrine appeared in 1889-1890. As this chart illustrates, the term appeared significantly more frequently in January 1890, the month after Bartholow’s article appeared and the worst stage of the influenza epidemic in the United States. In fact, 10% of the 272 pages on which this term appeared occurred in the two weeks starting December 28, 1889, the day that so many newspapers published all or part of Barthlow’s editorial specifically endorsing the use of this treatment. As a point of comparison, a more generic term for medical treatment, such as tonic, appeared on more than 30,000 pages in these same two years.

Chart of newspaper pages with term antipyrine, created from data available from Chronicling America, Library of Congress: http://chroniclingamerica.loc.gov/

Chart of newspaper pages with term antipyrine, created from data available from
Chronicling America, Library of Congress

Yet the question remains, did Bartholow’s intervention have a causal relationship on the use of the term antipyrine? A review of the content of selected newspaper articles suggests that the increased incidence and visibility of influenza, but not specifically Bartholow’s article, had a causal effect on the frequency of the use of this term. Newspapers reported on widespread use of antipyrine in cities with hundreds of cases of influenza, on the shortage of this treatment that was blamed on the monopoly in production exercised by a German physician, and on individuals who became ill from taking too much antipyrine (see for example, Tombstone epitaph., December 28, 1889, Evening star., January 08, 1890, and The evening world., February 22, 1890.)These articles generally did not reference Bartholow, thus suggesting that multiple factors, and not just one doctor’s editorial commentary, despite how widely it was disseminated, contributed to the more frequent appearance of this term in newspapers.

The relationship between antipyrine and the influenza epidemic can be visualized on an even larger scale using the Voyant tools for text analysis. Figure 4 uses the full text version of the journal, Medical News, for two years, 1889 and 1890, with nearly 3 million words in the four volumes. The term antipyrine appears just 125 times in this entire corpus. Using the collocation tool in Voyant, it is possible to graph the 50 words which appear most frequently in proximity to this term, as shown in

Chart of collocated terms with “antipyrine,” created using google fusion charts and voyant text visualization tools, using complete text of Medical News, 1889-1890, available from
Medical Heritage Library

Figure 4. This network of collocated terms clearly indicates that antipyrine appeared most frequently in close proximity to terms related to administering this treatment, such as grain, dose, or patient. The term also appears in close proximity to other treatments including, quite frequently, cocaine. Yet the word influenza does not appear as one of the more frequent collocations with antipyrine, despite the fact that Bartholow’s editorial appeared in this journal and then was widely reprinted in newspapers across the United States.

Compared with the humanist’s more traditional methods of close reading and contextualized interpretation, data visualization tools do not provide the depth or complexity desired by medical historians. Yet these tools can provide insights because of their capacity to look across scales (millions of words in a single journal or hundreds of thousands of newspaper pages) and suggest potentially important associations. The challenge for medical historians seeking to understand causal relationships, therefore, is to balance the additional contributions of new tools with the recognized value of more traditional methods.

The following two tabs change content below.

Tom Ewing

Tom Ewing is professor of history at Virginia Tech. He directed the history of tuberculosis project funded by 4VA in summer 2016.

Latest posts by Tom Ewing (see all)

About Tom Ewing

Tom Ewing is professor of history at Virginia Tech. He directed the history of tuberculosis project funded by 4VA in summer 2016.
Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *