A look into social media engagement with racial health inequality during the COVID-19 era
Updated: Apr 7
The world is currently facing an unprecedented global pandemic. The COVID-19 pandemic has been devastating socially, economically, and mentally in addition to the virus’ physical toll. However, the damage caused by COVID-19 is not equal due to a variety of social factors that have historically led to inequitable health outcomes for marginalized individuals, especially people of color.
Activism Always is an interdisciplinary team with a shared passion from social impact and data (Miki O’Reggio, Jin Pu, Hoa Nguyen, Chelsie Lui). As a student-founded initiative interested in how accessible data strategy can transform impact in the mission sector, Activism Always has partnered with COVIDxNOW, a global consortium of leaders interested in sourcing solutions around COVID-19. Our project aims to assess engagement with this very real issue, and to survey what (digital) voices have been leading this conversation about COVID-19 and health inequality.
The Activism Always Team conducted a two-part analysis to learn more about the relationship between discussions around coronavirus and racial health inequalities online and offline. Our team cleaned and visualized the data to glean insights around social media engagement with the issue of racial health inequality (especially during the era of COVID-19) in comparison to the real-life racial breakdown of COVID-19 cases. The two platforms we analyzed for social media engagement were Twitter and Google Trends. The dataset we analyzed for COVID-19 cases was from the COVID Racial Data Tracker, available through the COVID Tracking Project and the Center for Antiracist Research.
As of 2020, it is accepted by the American Medical Association (AMA) that racism is a threat to public health; systemic racism impacts health inequalities, both in conscious and unconscious biases in health care and research.
It is also acknowledged by the AMA in the same article that race is a social construct rather than a biological one, and the origins of racial essentialism (or that biology and genetics defines a racial category) are ones rooted in oppressive desire to excuse and explain inequalities. The AMA rather uses race/racial category as a proxy for ancestry.
The history of racial health inequality in the United States has a long, traumatic past. To learn more about racial health inequality across the dimensions of why it exists, how it exists, and its impact, here are some selected resources:
Healthcare disparities in the US, Khan Academy https://youtu.be/7qAld9bGwlA
Does Racism Play a Role in Health Inequality, Williams https://youtu.be/Mudlev9tTIY
How racism makes us sick Ted Talk, Williams https://youtu.be/VzyjDR_AWzE
The US medical system is still haunted by slavery, Vox https://youtu.be/IfYRzxeMdGs
Nothing Protects Black Women From Dying in Pregnancy and Childbirth, Martin and Montagne https://www.propublica.org/article/nothing-protects-black-women-from-dying-in-pregnancy-and-childbirth
As the novel coronavirus has dramatically changed our lives in the past year, much like many of the ignored systemic inequalities and shortcomings of our current world, these health inequalities are clearer than ever. The most marginalized communities are being hit with the virus the hardest: undocumented, poor, incarcerated, and homeless. There is an overrepresentation of people of color being affected by COVID-19, disproportionate to the population. More difficult, many of these populations are not tested often or at all, with factors like work, geography, and documentation status on the line for these individuals.
These gaps and inequalities are not a new phenomenon, even as they have been more visible in the past year. In the United States, there is an established history of medical violence towards marginalized communities. For Black and Latinx individuals (among many marginalized groups in the United States) this distrust runs deeply, and it introduces another dimension of complication to the current pandemic. In addition to improving care within the medical institutions to serve these communities, there also needs to be a dimension of actually getting people to trust the institutions to not repeat their history of racist practices.
Methodology and Analysis
In better understanding the relationship between racial health inequalities and COVID-19, our data team (led by Hoa Nguyen and Jin Pu), conducted a two-part analysis. They analyzed online data through social media listening of publicly accessible Twitter posts and Google Trends data; they then analyzed offline data through analyzing the publicly accessible dataset published by The COVID Racial Data Tracker.
Section One — Online Data: Social Media Listening
The data collected for section one of our analysis was sourced from social media locations: Google Trends and Twitter. For Google Trends assessed 3 related keywords. For Twitter, we assessed 7 related hashtags about health equality and racism in the context of COVID-19 (#blackcovid, #latinxcovid, #covidracism, #covidinequality, #RacismIsAVirus, #IAmNotCovid19, #healthequity). This data was abstracted from March 1st, 2020 to December 31st, 2020.
First and foremost, finding the right search terms that encapsulate discussion and interest in racial health inequality about the coronavirus pandemic was difficult. The term we found — “Covid inequity” — is a comparably rare term compared to other search terms related to the pandemic (such as “Coronavirus update”, “COVID-19 Vaccine”); only a few states have record on this search term.
From the line chart, people are paying less and less attention to the daily updates of COVID-19, a possible sign that people are “getting used to the new normal”. Since November, the COVID-19 vaccine has become a heated topic. However while different terms were gaining and losing momentum, “Covid inequity” stayed low-key and did not vary throughout the time.
Nevertheless, this is only about the web search index; people who care about inequality may not search in Google. Another platform we analyzed for online engagement with racial health inequality and coronavirus was Twitter.
To start with, finding the right hashtag for racial health inequity in COVID-19 is not as easy as finding hashtags that represented a broadly established movement, such as “BLM”. From our team’s analysis, there were no trending hashtags that broadly represented a movement, campaign, or initiative addressing racial health inequality and COVID-19. There were some hashtags (such as #racismisavirus) that could be considered as a broadly representative trending hashtag, but in comparison to other hashtags during this period, it did not get nearly as much attention.
Overall, #racismisavirus was the most popular hashtag that addressed this intersection between race and COVID-19; #healthequity was the second-most popular hashtag following #racismisavirus. Other related hashtags were #blacklivesmatter and #sdoh.
Compared to March 2020, the number of mentions with #healthinequity almost doubled after the breakout of Covid19. (However, to note, the volume of mentions was still not high.)
#racismisavirus was a trending hashtag from March 2020 to June 2020, and was the top 1 hashtag in May 2020 from our collected tweets. From July 2020, while #racismisavirus stayed in the range of the top 3 hashtags for several months, the number of mentions decreased drastically.
Organizations that were called out/mentioned the most were Robert Wood Johnson Foundation (@RWJF), American Medical Association (@AmerMedicalAssn), and the American Public Health Association (@ PublicHealth)
Individuals that were mentioned the most were Abdel _Defiant ! (@Abdnys), uché blackstock (@uche_blackstock), Dr. Aletha Maybank (@DrAlethaMaybank), and Sir Michael Marmot (@MichaelMarmot)
Section Two — Offline Data: The COVID Racial Data Tracker
The data collected for section two of our analysis were sources from the publicly accessible dataset as hosted by The COVID Racial Data Tracker (COVID Tracking Project and the Center for Antiracist Research). The dataset was downloaded from the official website (https://covidtracking.com/race). The background of the dataset was written in PowerBI, and our team abstracted the columns Case by Population and Death by Cases. We then normalized all the numbers in the dataset by dividing the number over the average level across different established racial categories.
This dataset does not include all the confirmed cases and deaths; our limited dataset may reflect inaccurate and/or biased conclusions.
Cases by Population (Normalized)
Overall, Black, American Indian and Alaskan Native (AIAN), and Latinx populations have higher infection rates compared to white and Asian populations. This is indicated by darker coloration on the maps in PowerBI.
For Black folks, Maine (ME), Vermont (VT), and Oregon (OR) were reported to have the highest normalized infection rate; in other terms, the Black community were found to be most widely affected in those states compared to other populations. For AIAN populations, some states did not report data about this race. This may be due to a variety of reasons, including difficulty gathering data from Native populations, difficulty distributing tests in rural regions). The American Indian population had a higher normalized affection rate than the Black population in many states. While there was no outstanding pattern, it brings to attention the difficulty many Native communities are facing in the wake of COVID-19.
For the Latinx population, many states were denoted with white coloration due to no reporting data. But for those states which reported data, Latinx population normalized number was always higher than 1. This indicated they were more widely affected than other races.
Death by Cases (Normalized)
Since not many cases were reported in Latinx, multiracial, Native Hawaiian and Pacific Islander (NHPI) populations, many parts of the map have a gray coloration, which denoted that their were either no cases recorded for that population in the state (that no deaths were recorded at that time). We focused our analysis on the AIAN, Asian, Black and white populations.
All races suffered in terms of death by cases (normalized). AIAN, Black people, and Asians had a wide range of death rates across the map (some states had better numbers, while others showed worse cases). AIAN had a notably high normalized death rate in VT, Utah (UT), and Mississippi (MS). Black populations especially suffered in Michigan (MI) and California (CA). Asian suffered the most cases in the west region, in states such as CA and UT. The death rate of White people was less varied across the country; they had high normalized death rates in New Mexico (NM), CA, and Colorado (CO).
Detailed Statistics Per State
Detailed statistics are provided on the state basis. If interested in certain states, the dashboard can list all the information. Take CA as an example: NHPI, Latinx, AIAN had the highest “cases by population” rate, while Asian, Black and white had the highest “death by cases” rate.
Discussion and Limitations
The biggest takeaway from our analysis is that racial health inequalities are a real issue ― both seen in prior established research but also in the offline analysis we conducted using the dataset from the COVID Racial Data Tracker. However, in terms of public discussion about this topic, it has not reached the level of mainstream popularity as many other social movements and concerns have during the same period. It was difficult to find trending hashtags related to racial health inequality and even more difficult to find ones that addressed the intersection between racial health inequality and coronavirus.
The offline data showed a gap between the cases in infection and death for people of color (more specifically Latinx, Black, AINA, and Asian populations) compared to the white population. For many of the populations, especially smaller populations such as NHPI and mixed raced, we were unable to conduct any analysis. There was a clear sign that certain populations’ cases were not represented as clearly in certain states compared to others.
We expected to see variations in cases, as aligned with our background research. And we did find there were distinct variations. However, our work is non-conclusive and non-representative of the real state of coronavirus cases. To start, our work only looks at a limited range of data around this topic (both in our online and offline data). Only analyzing from this limited set of data, we cannot generalize our findings outside of the time range and platforms we collected from. It should be noted however, that while the issue of racial health inequality is not discussed at length in these digital spaces, it does not mean that they are not being discussed. Rather, many of these discussions are localized on the interpersonal and community level; for people who experience these injustices most closely, this is not a new discovery or phenomenon.
Additionally, while race is a major (and can be noted as the most significant) factor in determining health outcomes, in following research, it would be interesting to consider the intersections of other factors in relation to race and COVID-19: documentation status, wealth, gender to name a few. Additionally, it would be interesting to adapt this methodology with several different datasets collecting information about race and coronavirus cases; we could compare the different frequencies found between the datasets to find a clearer image of what locations and groups may experience significant numbers of cases compared to the national average. On that note, it is important to be clear our project only looked at cases within the United States, and the reported findings of this work cannot be generalized outside of the United States as well.
Overall, from this project, our team has used our social media listening methodology to better understand public engagement with the topics of racial health inequality and COVID-19.
The results of our analysis align with existing research on this area of focus. Although it is a positive that our research matches up with the existing work, the reality of these health gaps is a failure in itself. As the medical community is aware and acknowledges the fact that racial health inequalities in the United States are a byproduct of a history of systemic racism, there is a lack of public interest and action in this arena (according to our social listening analysis).
These gaps cannot be closed and this injustice cannot be solved fully through individual actions. However, considering the ability of digital communities to generate awareness and organize movement, greater activation of social media and mainstream public attention towards this topic could assist collective action towards making societal level changes.
Our greatest hope in conducting this work is sparking interest in the relationship between race and health inequality, contextualizing the gap through the setting of the ongoing COVID-19 pandemic. As we move through the first month of 2021, a year since the first case of COVID-19 was detected in the United States, there have been great strides to the contain spread of the virus: people coming together with their communities, rapid adoption of new socially distanced policies, and the accelerated development of COVID-19 vaccines. As we move forward towards vaccine distribution, and towards a safer and healthier future, we cannot return to a previous “normal” where health care and health outcomes are so clearly inequitable across populations.
From COVIDxNOW: COVIDxNOW Global Economic Leaders Consortium is aimed at unlocking all possible solutions to the impacts of COVID-19 and in the process creating hope, opportunity and job creation across the globe. Learn more about their work at https://www.covidxnow.org/.
Activism Always’ hybrid service combines an internal AI platform with strategy analytics, maximizing organizations' data capabilities and impact with an accessible price and format. If you have any questions or interested in a FREE informational consultation, contact us at firstname.lastname@example.org.
This essay was originally published on Medium on January 28, 2021. View the original article here.