MIT Collaboration: Blacks in American Medicine
This project was the result of a collaboration with students in the CMS.S62/S98 (Digital Humanities II) class at MIT, which was offered jointly with our class Active Archives, Engaging Interfaces at FH Potsdam.
The collaboration was supported by a trip to Cambridge, Massachusetts in May 2018.
The overarching goal of our work on this archive was to tell the untold stories of the black community in American Medicine. In particular, we want to ensure that the contributions of the individuals in the archive to medicine, science, and civic life were properly documented, recognized, and preserved, while leaving space for new members of the community to be added by users. We wanted to provide an user interface that allowed visitors to the archive to visualize relationships between individuals in the community, issues the community faced and how they approached addressing them, as well as mapping personal and professional journeys across time and geography.
Our guiding questions were therefore:
How can we create an archive that allows for contribution from the community, acknowledges gaps in data, and can be continuously built upon in the future?
How do we digitize, represent, and aggregate the source documents in a way that is reflective of their materiality and emphasizes the stories they tell?
Our intended audiences for this archive are people who have personal connections to medical practitioners in the archive, people without direct personal connections who seek to learn more about the history of the black community in America through the lens of healthcare, and people interested in American history who might not otherwise encounter the contributions of black people in advancing medicine and building communities.
In the early 1980s, Professor Ken Manning (Thomas Meloy Professor of Rhetoric in MIT’s Program in Writing and Humanistic Studies, and Science, Technology, and Society) became interested in studying the history of African Americans in medicine. Growing up in Dillon, South Carolina in the mid-1950s, Manning first became interested in the history of medicine when visiting the office of local African American physician, Dr. Gordon, who said his office was “American,” not “Southern” as the waiting rooms were not segregated. Noticing only African Americans in the waiting room, the then six-year-old Manning asked Gordon where the white patients were, to which Gordon replied that they came after-hours to receive treatment for venereal diseases and afflictions that they didn’t want their regular doctors knowing that they had. Manning said he suddenly then became aware that the racial divide in America was transcended when it came to seeking certain kinds of medical treatment.
Manning said he realized that practicing medicine was not purely scientific, and that it was a facet of society that ran along U.S. social and political lines. Studying the work of black physicians “gives a lens through which you can see the conditions under which a whole group of people lived,” said Manning. “It is because almost everybody experiences the healthcare system at some point in their lives, whether it’s at birth, whether it’s at death, and many times in between.” By looking at the health and the healthcare options of a particular group of people at a given time, one can figure out its place in society, said Manning.
When Manning began collecting documents related to black healthcare practitioners, he did not limit his scope to the work of physicians only - he also looked at nurses, pharmacists, physicians assistants, as well as the lives of their patients, and the institutions at which they worked and studied. Manning began by examining medical school graduating classes and medical licenses granted every year from 1860 to the early 1980s. Through this, he assembled a database containing biographical information for nearly 23,000 members of the black medical community including data such as name, gender, date of birth, place of birth, university attended, graduation date, medical speciality, positions held, and institutions at which positions were held. When available, Manning also included data on religious affiliations, membership in clubs and societies, certifications, awards, military services, and non-medical positions held.
The biographical database compiled by Ken Manning.
In addition to the database - now saved as a .csv file - Manning also sought out and collected photocopies or original hard copies of manuscripts and collections of black medical practitioners’ work and correspondence. “Manuscript and archival collections provide an essential body of information for the project,” wrote Manning in an initial update report on the project. “These collections identify the major characters, the important issues, and the scenarios in which characters confront and attempt to resolve issues.” Beyond seeking information on the public record related to which physicians practiced and where, Manning sought to find documents that revealed the physicians analysis, introspection, and discourse with other members of the black community on issues that affected them, both professionally and personally.
A selection of the documents collected by Ken Manning
Over nearly four decades, Manning compiled tens of thousands of documents - published papers, newspaper articles, magazine clippings, class notes, meeting and symposia agendas, letters and correspondence, reports, poems, art, research notes, and more - related many of the 23,000 individuals in the database. These papers and photocopies were held in folders and stored in dozens of cardboard boxes, roughly grouped based on to which individual they corresponded. These boxes of documents were housed physically at MIT, making up an archive of sorts, but were not publicly available, and were not part of the MIT official archives - interested parties would need to seek out Manning directly to get access to the boxes.
A previous group, Andy Stuhl and Evan Higgins from HyperStudio, had created a prototype of a tool with which users could search through the data represented in the .csv database via a faceted browser of the biographical information related to 23,000 individuals. The browser was integrated with HyperStudio's Chronos timeline which allowed events to be filtered based on the user’s input, and displayed different event categories in relation to each other. While this tool offered a useful connection to the database information, it was limited in its ability to let the user engage actively with it. There was not an option to draw new connections between doctors that was not already available in the filters, there was no visual vocabulary connected to the information, and there was not the wealth of primary documents from Manning’s physical archive. Above all, the tool did not foreground the heart of the archive, as Manning had said - the personal stories in the documents themselves.
The user interface of the previous tool constructed to represent the BAM data
Therefore, the challenge for our team was to start with an enormous body of physical documents, some data points, some oral histories recorded by Manning, and create the beginnings of a project that will continue in the hands of future students and researchers: an active, sustainable archive that enables storytelling, discovery, collaboration, and discourse within its audience of users.
Guiding Literature and Archival Theory
One the most foundational works in shaping our approach to creating this archive was new media scholar Wolfgang Ernst’s Digital Memory and the Archive. Ernst discusses the evolution of the archive as a concept, and how the infrastructure of digital spaces has transformed the way archives are created, maintained, and used with respect to notions of an analog archive. Most notably, he brings up the idea of computer memory and computer storage with respect to information management, and how the time it takes to retrieve information has been so shortened by this technology, archives are now almost always in motion. Digitalization has collapsed the binary between stored information and information in motion/transmission. The web ecosystem has allowed for a flattening of the temporal dimension of information and for its spatial dimension - it can be anywhere at once, and accessed/modified at any time. Ernst points out that one of the consequences of such immediate archival power is that it transforms the process of archiving from a cultural record to “a performative form of memory as communication”.
Since Ken Manning’s work sought to preserve the cultural memory and continue the active construction of new cultural memories for a specific group of people, we wanted our archive to be more than an online repository of documents. Ernst’s work was concerned with studying the physicality of digital archives - the tools, mechanisms, hardware, software, computations - and how these mechanisms affected the creation of cultural memory. We, too, examined our approach and the tools we selected to create our digital version of Manning’s archive. Honoring his original vision for the project, we wanted to construct our archive with mechanisms that both preserved the connections of the data and documents to their spatial-temporal origins, but also enable new connections to be drawn across space and time through storytelling and curation. This combination of storytelling, curation, and added narration was described as “storymaking” in “Introducing the Active Archives Initiative: Making Stories, within the Archive” by Evan Higgins on the HyperStudio blog page. Even before this project was fully realized there was an emphasis placed on constructing narratives based on raw materials and being aware of the subjectivity surrounding those narratives, whether the user chose to focus on the document’s purpose, the time period it emerged from, or the geographical information it yields.
Another scholarly work that informed our process in constructing the BAM archive was “Graphesis: Visual Forms of Knowledge Production,” the work of visual theorist Johanna Drucker. In the text, Drucker discusses how our growing relationships to screens - in our laptops, our smartphones - have put an increased emphasis on the use of visuals to convey information. More than just reproducing existing non-visual information, Drucker argues, graphic representations actively construct and control our relationship to knowledge.
Drucker also argues that observation of the phenomena is never the fully representative of the phenomena itself. Information collected is dependent on biases and objectives and the way it is displayed adds another layer of biases and objectives on top of that. Therefore, the knowledge derived from graphic representations is dependent on interpretation, not on the actual phenomenon. But the rendering of statistical information into graphical form gives it a simplicity and legibility that hides every aspect of the original interpretative framework on which the statistical data were constructed. Acknowledging how data is constructed - the biases in its collection and display - allow the viewer to more accurately interpret the data, to accept the ambiguity and better relate it to the actual phenomenon.
In constructing our archive we tried to build in features and representations that allowed for the data and documents to be viewed graphically and textually in several different arrangements, so as not to provide the viewer with the semblance of one “definitive” way to experience the data. We also sought to add features that allowed the user to see the source documents in several different contexts, alongside information that provided more details on how it was collected. Giving the user more options and more transparency seemed to be the best course of action so as to not give a false semblance of one “right” way to explore the BAM data and documents.
Other texts that shaped our approach included Noah Lenstra’s “social inequalities in shaping of cultural heritage infrastructure,” which discussed the necessity of providing and supporting community capacity for marginalized communities working with institutions to create archives of the community’s experience and history. If marginalized communities are to feel autonomy in the space, and have more control over the production of digital heritage artifacts, proper planning and collaboration must be built into the institution’s plan for constructing the archive itself. Since the BAM archive is dedicated to preserving and furthering the digital storytelling capacity of a marginalized group, it was critical that we engage with the community while designing the archive. We also strove to create opportunities for user co-creation, access, and contribution so that the archive could not only be active, but sustainable and empowering.
A final article that informed our work on the BAM archive was Rodighiero Dario, Kaplan Frédéric, and Beaude Boris’ “Mapping Affinities in Academic Organizations,” in which the authors discuss the possibilities of visually representing “affinities” in different organizations, focusing in their case study on the different labs at l’École polytechnique fédérale de Lausanne. The authors define affinities as relative to the closeness between peers, when people they have something in common, an interest, a research discipline, a spoken language, a career, among others. They also differentiate between potential affinities (commonalities that have not resulted in a formal collaboration) and actual affinities (those that have). Further, they distinguish between affinities that have been recorded as some kind of data, and therefore exist as a digital trace, and affinities that leave unreadable, invisible, or difficult to find traces.
In our data for the BAM project, there are a number of individuals who have commonalities in terms of their areas of study, practice, their geography, their birthplaces, their membership in organization, the conferences they attended or organized, and other potential/actual affinities. What is especially challenging our body of data is the lack of consistent digital traces available to represent and map these commonalities. We were therefore mindful of presenting both the individual, the collective, and the combination thereof in our visualizations, especially since our prototype contained documents containing a disproportionate amount of data about a small number of individuals.
The visual grammar of Rodighiero et al’s work in mapping their laboratories of interest is compact and efficient, as with each zoom in the user can move down through levels of governance and structure to see the contributions of the individuals (or even just their membership) in every lab. This structure also allows for the easy inclusion of members who do not have any available digital traces or data. In our case, we thought perhaps this approach could be useful in representing the biographical data of those doctors and individuals in medicine whose data was incomplete or whose associated documents we had not yet had time to process.
Existing Archives as Case Studies
Our approach to designing our archive was also informed by examining several of existing archives, particularly those that dealt with similar issues we expected to encounter in our own work. The first such archive was “Geography of the Post: U.S. Post Offices in the Nineteenth-Century West,” as it deals with representing incomplete data sets with fidelity and transparency. The site discloses that the project team was unable to locate and map every post office mentioned in Helbock’s data set, and has placed a percentage bar on the map’s legend to indicate what percentage of the locations during a given period are unmapped. The site also acknowledges that is does not capture when an office changed names. A name change shows up as if it were a brand-new post office, with the old office “closing” and the new office “opening,” even if it was the same post office in continuous operation.
The missing data for the name changes is somewhat circumvented by the design of the map itself, which overlays the post offices at the same location to create a darker color dot, somewhat like a heat map, indicating the constant presence of a post office, regardless of name. The unmapped percentages are somewhat less effective. They remind the user that the data is incomplete, but the percentage metric is pretty ineffective in terms of visualizing the missing pieces of data, especially in a design that privileges the location metric. This project tackles several issues related to using incomplete data sets in creating interactive archival representations. It acknowledges within the archive itself (unmapped vs. mapped percentage feature) that a portion of its data is missing – yet fails to fully elaborate or clarify which data are missing. The percentage metric is only useful in trying to visualize the magnitude of missing data. It does little to elaborate on the nature of this missing data. The name change missing data, however, is handled more skillfully by using the “heat map” overlay feature of post offices in the same location.
In our archive, we strove to acknowledge and make visible the gaps in our data, as oftentimes the biographical information in the database was incomplete, missing details like gender, university, place of practice, and others. We also sought to make clear that the archive was a work in progress and build in features for users to contribute to the archive - avoiding any potential misinterpretation that the data set was complete and finished.
Other archives from which we drew design inspiration were the spatio-temporal network visualizations from the “Museum of the World” and the “NYC Foodiverse” archives, both of which plot data over space in addition to another dimension. In the “Museum of the World” archive, human artifacts are plotted across continents and time, and sorted based on themes, and once a user clicks on an artifact, network lines connect it to other artifacts based on their relationship to a theme, technique, or time period. The Foodiverse represents data on restaurants - reviews, ratings, health code violations - in New York City through a network cloud, a map, and a graph, giving the user multiple modes through which they can visualize the data.
While our initial data set, drawing from the biographical database, was predominantly spatial/geographic in nature, as each individual practitioner was associated with a place of birth (usually within the U.S.), a place of education, and a place of practice. It made sense, therefore, to honor the geographic quality of this database by mapping individuals across a map of the U.S., which is one way users can plot the data in our archive. However, we did not want to limit or reduce the other dimensions of the data set, especially when also including the documents after they were scanned, so we were sure to build in other ways for users to view and interact with the data set. Therefore, in addition the map visualizations, we also included features to allow the user to search and sort within the physicians represented in the archive, and a story-driven homepage to guide the user into the data and documents by presenting them in a way that created a narrative linking their content. For further work, we would propose creating a network of affinities visualization to represent relationships between physicians, independent of a geographic structure like a map.
We also drew on the work of the Civil Rights and Restorative Justice Project from Northeastern University’s Law School, which has investigated over 400 cases of racial homicides and lynchings, and has collected over 20,000 items related to these events. The CRRJ’s mission is to offer holistic portrayals of the individuals, families, and communities who suffered at the hands of domestic terrorists, and to give voice to this community. Unlike other repositories and lynching databases that center on the horrific images of dismembered black bodies, the CRRJ archive has created a reading room that presents photography, documents, and summary “case studies” related to each victim, seeking justice and remembrance for the victims without centering on images on violence. Since so many documents related to racial homicides were lost, damaged, or not properly filled out by a variety of institutions, the archive’s team is still in the process of collecting documents. The CRRJ team has coordinated three “history harvests,” June 2017 in Tuskegee, Alabama, October 2017 in Birmingham, and March 2018 in Selma to convene descendants and witnesses to share their story and contribute materials for the archive.
The CRRJ is working to create an active archive to bring about restorative justice, and acknowledge and explore the effects of systemic violence on a marginalized community. Our archive, though not focusing on incidents of violence and murder, is also working to shed light on the issues faced by the African American community throughout history, through the lens of medicine and healthcare. The CRRJ’s approach is therefore an excellent case study for us to understand how to design an archive that honors and preserve the experience of the individual alongside that of the collective community. They are also a model care for how we would want future groups to approach engagement with the community - perhaps by organizing similar “history harvests” or reaching out directly to family members of the people in the BAM archive.
On the German side, the team members from FH Potsdam began the project by analyzing the biographical data of over 23,000 doctors that were provided. Since the data was in an excel file extracted from an old database that used decades-old format that made it hard to readily use, the data was parsed into json files that are easier to use for web development. We then added geolocation data for the doctors using Google Maps, and used the data to create various visualizations based on a world map. After obtaining PDF and text of the documents, we finished the prototype by adding ability to search the doctors and the documents as well as pages to explore the stories of the archive.
On the American side numerous difficulties were encountered in the project development, and a lot of the difficulties were in just starting the project. None of the documents were digitized, and the digitization process proved to take a lot longer than expected. Most of the documents were photocopies of old papers, which meant that there were portions of the documents that were hard to read, and a lot more sections of the documents where FineReader was not able to create a reasonable transcript of the documents. This meant that proofreading process became more tedious as some documents had to have large sections transcribed manually.
As German members of the team we did not have easy access to the documents but only to the the biographical data at the start. This meant that for most of the project timeline we did not have great opportunities to integrate the documents into the design. A lot of our efforts were focused on understanding and representing the biographical data, while designing and building the framework that would allow users to explore the data, understand its incompleteness and engage in the archive, including the documents. The time that we were then able to actually spend on integrating the documents into the user experience was disproportionate to the importance of this part of Ken Manning’s archive.
While this prototype gives an overview of our basic vision for the archive, there are many features that we were not able to fully develop or to implement. Our vision for this archive is for it to be a constantly updating and shifting body of documents, responsive to new information, while preserving the materiality and original content of the documents. There are several features that we believe should be developed further, or added to the archive to achieve this goal.
Currently the BAM archive is largely comprised of front-end designs with a limited connection to a functioning back-end. For example, the search function and the map display both display information collected in the archive, but they don’t have all the digitized data integrated at this time. The archive pages shown above are part of a click-through version, so while the stories about the prominent members of the archive are available, at the moment it is not possible to click through those stories outside of the set path. Ultimately the click-through version will be replaced with a completed archive as seen in the pictures.
An essential element of an active archive is the community that engages with it, so there are several features designed for user engagement and input into the archive.
One feature that is essential to the ability of the archive to incorporate new information is the “contribute” feature. While our prototype include mock-ups of this feature, they are not yet fully functional. Without these features there is no way for new information or stories to be added, so these features are essential to BAM’s status as a living archive.
The major obstacle in implementing these features is that they require a person to filter through submissions. Users who are contributing information alone may only be submitting what they remember, either form stories or their own experience. While this input is valuable, it can be hard to corroborate, which could lead to potentially false or misleading information in the archive. A moderator will need to check the information submitted to the archive and verify that it is relevant, true, and submitted to the correct category. The ability for the contribute forms to autocomplete known and verified information could reduce the amount of information a moderator would have to sift through, as well as streamline the submission process for the user. Similarly, a moderator will need to edit submitted stories for length, clarity and content, and insert links to supporting documents and physicians. Machine reading could aid in some of these duties by recognizing key phrases in the new material and suggesting related documents and stories. Filtering through these submissions will be a significant amount of work, and will require someone dedicated to moderating the archive.
Additionally, individual submissions may not preserve the materiality of documents in a way that is consistent with the original documents archived. To digitize the original documents, we used scanners to produce high quality pdfs. Many people don’t have access to such scanners, and may not know how to take a usable photograph of the documents. This presents a problem, as preserving the materiality is essential to protecting the history of the documents and verifying the information the user is submitting. A standardized document image submission process will need to be developed to allow as many people as possible to contribute, while preserving the materiality and legibility of the documents.
While this prototype was designed with the ease of use of several different types of users in mind, these uses have not yet been tested by outside people. In order to ensure that no errors in design were made, and that the interface is intuitive to users unfamiliar with the content, outside user testing is needed.
We would like to set up sessions for users of each audience type to explore the prototype. The different audiences (personally related, interested in physician, interested in topic) will have different knowledge levels, and different objectives. Therefore, it is important to test the prototype from each of their perspectives before finalizing the design of the archive. These tests will ensure intuitive use for all audiences.
The BAM digital archive design prototype has fulfilled many of our goals, presenting information in a responsive way that preserves the materiality and allows users to explore and contribute their own narratives. As the functionality is added to the prototype we will allow users to add in their stories and explore the role of a moderator of the archive. There will also be more material added from the physical archive, letters and newspaper articles adding more richness to the stories and filling in gaps in the archive. Once our archive has been fully implemented we will conduct user tests and expand vision based on the needs of the users.