Advocacy, Hesitancy, and Equity: Exploring U.S. Race-Related Discussions of the COVID-19 Vaccine on Twitter

Background: Our study aimed to describe themes of tweets related to COVID-19 vaccines, race, and ethnicity to explore the context of the intersection of these topics on Twitter. Methods: We utilized Twitter’s Streaming Application Programming Interface (API) to collect a random 1% sample of publicly available tweets from October 2020 to January 2021. The study team conducted a qualitative content analysis from the full data set of 1110 tweets. Results: The tweets revealed vaccine support through vaccine affirmation, advocacy through reproach, a need for a vaccine, COVID-19 and racism, vaccine development and efficacy, racist vaccine humor, and news updates. Vaccine opposition was demonstrated through direct opposition, vaccine hesitancy, and adverse reactions. Conspiracy and misinformation included scientific misinformation, political misinformation, beliefs about immunity and protective behaviors, and race extermination conspiracy. Equity and access focused on overcoming history of medical racism, pointing out health disparities, and facilitators to vaccine access. Representation touted pride in development and role models, and politics discussed the role of politics in vaccines and international politics. Conclusion: Our analysis demonstrates that Twitter can provide nuances about multiple viewpoints on the vaccine related to race and ethnicity and can be beneficial in contributing to insights for public health messaging.


Introduction
From September to December 2020, a nationally represented survey in the United States (U.S.) found that respondents' intentions to take the vaccine rose from 39.4% to 49.1%, and non-intent decreased from 38.1% to 32.1% [1]. Racial and ethnic numerical minorities reported having greater vaccine hesitancy than White people. Specifically, studies have consistently found that Black and Hispanic people report higher levels of vaccine hesitancy than White people in the U.S. [2][3][4]. In addition, Asian people and people reporting more than one race or other have higher levels of vaccine hesitancy than White people [4]. As of April 2021, 65% of White people, 11% of Hispanic people, 9% of Black people, 1% of Asian people, 1% of American Indian or Alaska Native, and <1% of Native Hawaiian reported receiving at least one dose of a COVID-19 vaccine among the 55% of data that had race and ethnicity reported [5]. It appears that vaccine hesitancy may be one factor impacting vaccine uptake.
A driving factor for vaccine hesitancy among specific racial and ethnic groups is the history of and current racism in the U.S. healthcare system [6,7], from Tuskegee to the present day [8]. In a nationally representative study of U.S. adults ages 50+, health care discrimination was associated with higher odds of elevated HbA1c and C-reactive protein, a marker of inflammation and a predictor of coronary heart disease [9]. Black Americans report significantly more experiences of racial discrimination in health care settings, leading to apprehension towards individual providers, as well as mistrust of the overall healthcare system [6]. Discrimination in the healthcare system could impact how people perceive trust in a vaccine. For instance, the Black community reports less vaccine trust and higher levels of perceived potential harm from the vaccine compared to other races in the U.S. [10][11][12]. Americans who get immunized for the flu yearly describe having a trust in public health and government authorities, as well as having trust in vaccine production and health care providers administering it [7]. Yet, vaccine hesitancy among African Americans is connected to an increasing concern about the government's motives [7,12].
The politicization of the COVID-19 vaccine increased in 2020 [11], which could have exacerbated concerns. Analyzing the context of the time period in which the pandemic has unfolded is critical to understanding COVID-19 vaccine hesitancy. Since the beginning of the COVID-19 pandemic, the U.S. has also undergone a historic presidential election and a significant rise in racial tensions and social justice movements. At the beginning of the pandemic, some elected officials continuously used offensive terms linking the coronavirus to the Asian community believed to fuel violence against the Asian American community and sparking a public outcry against the use of the racist terms [13,14]. During the summer social justice protests of 2020, members of the federal government utilized divisive rhetoric, leading to a further disconnect between elected officials and members of the Black community [15]. The use of discriminatory language established a level of distrust in the government among communities of color, which has resulted in concerns and hesitancy about the safety and efficacy of the COVID-19 vaccine. Consequently, communities of color have cited the federal government as the least reliable source of information regarding COVID-19, and doctors and other public health officials as being the most reliable [15]. It is paramount to understand the power of social media to influence behaviors such as COVID-19 vaccine uptake during this pandemic.
As news progressed about the vaccine production, social media outlets were used as a space to voice opinions about efficacy, equity, ethics, representation, and conspiracies surrounding this process [16,17]. Even before the pandemic, the anti-vaccination community has used social media to influence health decisions with their opinions about vaccines, and COVID-19 vaccines have not been an exception [18][19][20]. Misinformation on social media has played a major role within the COVID-19 pandemic, as there are no standards of regulation regarding quality, accuracy, and availability of the information [21]. Vaccine hesitancy is highly influenced by misinformation propagated on social media [22]. In addition, Twitter can be used to spread racism. Emboldened anonymity along with echo chambers can fuel racist tweets [23]. In particular, there has been a rise in negative tweets referencing Asians with the emergence of COVID-19 pandemic with the apex in negative Asian sentiment occurring during the week of March 16, when President Trump used the phrase, "Chinese virus" [14,24]. The intersection of racism and the pandemic could have a similar impact on vaccine perception.
With the availability of Pfizer, Moderna, and Johnson & Johnson vaccines in the U.S., uptake is of the utmost concern for curbing the pandemic. When more individuals become vaccinated in a community, the virus is less likely to be transmitted due to herd immunity [25]. Dr. Fauci stated that 70% to 85% of the American population needs to have the COVID-19 vaccine to achieve herd immunity [26]. Twitter provides a natural laboratory for researchers to study perceptions of the vaccine in order to understand public sentiment within this complex social and medical milieu. Our study aims to describe themes of tweets related to COVID-19 vaccines, race, and ethnicity to explore the context of the intersection of these topics on Twitter using qualitative content analysis. This type of analysis has the potential to contribute to message development to support vaccine uptake.

Methods
A random 1% sample of publicly available tweets in the United States was collected from October 2020 to January 2021, using Twitter's Streaming Application Programming Interface (API) that included key words from all the categories of race/ethnicity from a keyword list of 518 race-related terms, COVID-19 from 75 coronavirus-related keywords [14], and vaccine from 28 vaccine-related keywords (Appendix A). Terms for vaccines included variations of vaccine, vaccination, immune, immunization, vax, and anti-vax. The first COVID-19 vaccine was given on 14 December 2020 in the U.S., so the study timeline captures the anticipation of having a vaccine, as well as early dissemination. The study team conducted a qualitative content analysis from the full data set of 1110 tweets. This study was determined exempt by the University of California, San Francisco Institutional Review Board.
The study team developed the codebook based on a literature review and reviewing tweets from the sample. Specifically, the team had an extensive multi-session discussion about the first 335 tweets from the sample to solidify the codes and definitions in the codebook. The final codes were vaccine support, vaccine opposition, conspiracy and misinformation, equity and access, representation, and politics. Using this coding scheme, all the tweets were double coded by two study members independently. Eight members of the study team discussed all discrepancies in the coding and came to a consensus on the final code for each tweet to achieve complete inter-rater reliability. Utilizing thematic analysis, the team analyzed tweets within each code to identify themes [27], which are explained in the results section. With this extensive consensus building process, we sought to maintain data trustworthiness through utilizing multiple data analysts with different racial backgrounds and life experiences.

Results
From the 1110 tweets, themes emerged from the overarching categories of vaccine support, vaccine opposition, conspiracy and misinformation, equity and access, representation, and politics. Table 1 provides the detailed themes of each category along with illustrative tweets. In general, many of the race terms in our sample were used as a descriptor, and some of were in a derogatory manner. Of note, many of the terms associated with Asian people utilized anti-Asian rhetoric related to the coronavirus like "Chinese virus."

Vaccine Support
This category represented 21% of the sample. There were six themes in this section. Vaccine affirmation focused on tweets reporting the excitement that Black and Latinx people were getting the vaccine, as well as some Twitter users viewing it as an act of "dismantling systemic racism." Advocacy through reproach were tweets that expressed agitation (many times through expletives or insults) that were "calling out" perceived despicable behavior as a way to assert that people should get vaccinated. The metaphor of a need for a vaccine for COVID-19 and racism recognizes the interplay between the pandemic and racial tension and the deep desire to remedy both of the issues. Vaccine support was also shown through positive tweets about vaccine development and efficacy. There was a trend of racist vaccine humor through stating what people already did (e.g., like eating certain types of food), and then saying "don't worry what's in the covid vaccine." This pattern was used many times without race terms as well. News updates supported vaccine support by providing accurate, and potentially inspiring updates about the vaccine within communities of color.

Vaccine Opposition
This category represented 12% of the sample. There were three themes in this section. Direct opposition were tweets that strongly asserted that they would never take the vaccine. Tweets displayed vaccine hesitancy through expressing reservations about the vaccine and saying that they wanted to see other groups get the vaccine first or gain understanding of connections to negative consequences. Some tweets focused solely on potential or rare adverse reactions.

Conspiracy and Misinformation
This category represented 15% of the sample. There were four themes in this section. Scientific misinformation were tweets that directly provided information counter to research about the vaccine. Political misinformation focused on perceived government intervention that was not true, such as the government using vaccines to insert microchips and governmental vaccine use for population control. Some tweets focused on beliefs about immunity and protective behaviors, which focused on the strength of the immune system over and above the protection of the vaccine. The race extermination conspiracy tweets posited that the vaccine was created to "kill off [people of color] POC."

Equity and Access
This category represented 18% of the sample with three themes emerging. Overcoming history of medical racism references tweets that acknowledge the past, but urges Black people to still get vaccinated. Tweets also focused on pointing out health disparities related to COVID-19 based on pre-existing conditions, hospitalization, and work conditions. Twitter users shared information to promote facilitators to vaccine access through prominent speakers and organizations that are from and represent communities of color, as well as petitions to demand access.

Representation
This category represented 13% of the sample with two themes. Tweets highlighted pride in development through celebrating that a Black women developed the vaccine. Twitter users also emphasized the importance of role models from Black and Latinx communities by publicly taking the vaccine or promoting vaccine uptake. Role models included organizational leaders, civil rights leaders, and celebrities.

Politics
This category represented 20% of the sample. Twitter users discussed the role of politics in vaccines through mentioning politicians and political parties concerning their stance on mask mandates, access to the vaccine, speed of the vaccine development, and the presidential transition. Tweets also mentioned international politics, and they were commenting on the COVID-19 impact and vaccine roll-out in other countries.

Discussion
This is the first study, to our knowledge, to conduct a qualitative content analysis on the intersection of COVID-19 vaccines, race, and ethnicity. Our analysis demonstrates that Twitter can provide nuances about multiple viewpoints on the vaccine related to race and ethnicity. The tweets revealed vaccine support through vaccine affirmation, advocacy through reproach, a need for a vaccine for COVID-19 and racism, vaccine development and efficacy, racist vaccine humor, and news updates. Vaccine opposition was demonstrated through direct opposition, vaccine hesitancy, and adverse reactions. Conspiracy and misinformation included scientific misinformation, political misinformation, beliefs about immunity and protective behaviors, and race extermination conspiracy. Equity and access focused on overcoming history of medical racism, pointing out health disparities, and facilitators to vaccine access. Representation touted pride in development and role models, and politics discussed the role of politics in vaccines and international politics. The information gleaned from this qualitative content analysis can be beneficial in contributing to insights for public health messaging.
Discussion of COVID-19 on Twitter can display sentiments of anticipation, anger, and fear [28]. Some people may oppose vaccinations because of mistrust of healthcare providers and big pharma, denialism ("rhetoric employed in order to give the appearance of legitimate debate where there is none, with the goal of rejecting an argument for which there is a consensus of expert opinion") and social media's role in providing confirmation bias [29], (p. 4484). Our research found that COVID-19 vaccines are susceptible to the some of those factors. To address it, tactics could include engaging people with opposition and/or hesitancy through discussion with healthcare professionals and having open forums on social media sites [29,30]. Furthermore, COVID-19 vaccine opponents were more likely to share unreliable information (34.5%) compared to vaccine proponents (11.3%) [17]. Our data also shows that the public's perception of the COVID-19 virus and vaccination may not align with accurate information and can even promote racism, like the use of the term "Chinese virus." Discussions around perception of the COVID-19 vaccine and access to the COVID-19 vaccine are important to analyze so adjustments can be made in the delivery and approach to deploying the vaccine [31]. Our study found tweets that were explicitly related to equity and access. Research demonstrates that the COVID-19 pandemic has disproportionately impacted communities of color due to socioeconomic status and lack of access [32]. Moreover, older Americans living in low-income households who are Black or of American Indian race have a higher associated risk of illness from COVID-19 [31,33]. Since equitable COVID-19 distribution is linked to removing the associated barriers [34], Twitter sentiment under the theme of equity and access communicates a relevance for vaccine equity and a priority of vaccine coverage to go to individuals at a higher risk of infection [31,33]. Our study findings support the perception that equitable access to the COVID-19 vaccine rollout should be addressed amidst barriers by race and socioeconomic status.
Representation is an area that could be utilized for positive messaging about the vaccine. In our study, tweets referenced Dr. Kizzmekia Corbett's work in the Moderna vaccine development process. In recent years, Twitter has been used by leaders and officials from all over the world to communicate information about policies and current events. This uptake by world leaders contributes to the growing influence of social media on behaviors of citizens, especially during times of crisis [35]. Willingness to receive the COVID-19 vaccine was associated with recommendations made by former Vice President Biden, the Centers for Disease Control and Prevention, or the World Health Organization [12]. There is a significant correlation between vaccine trust and an individual's trust of news sources who promote trust in vaccines [11]. There is an opportunity to partner with leaders identified on Twitter to share information about the benefits of vaccine development. For instance, there could be a Twitter campaign that features Dr. Corbett, as well as other significant leaders in this area. Additionally, our study included exemplary messaging expressing vaccine support that could potentially be utilized, along with different approaches taken, including affirmations and attempted humor. In fact, research shows that humor helps exhibit positivity and a sense of cohesion during COVID-19 [36]. Yet, it is important that the humor does not include racist tones like the tweets in our study.
Many of the political tweets expressed conflict and disapproval. Previous research has suggested that the age of social media as a primary form of news distribution has created a space in which individuals can selectively expose themselves to information, furthering the divide between political parties and crafting one-sided narratives based on a predetermined set of political views [37]. Research on the varying sources of COVID-19 misinformation finds that misinformation issued by politicians garners more attention and a more significant reaction than misinformation from many other sources due to their high-profile nature [38]. Spread of this misinformation from politicians themselves as well as their supporters may be incentivized by political motives [38]. Overall, the current climate of political polarization makes social media a breeding ground for the spread of misinformation, political incentivizing, and selective distribution of facts regarding COVID-19 vaccines. A media approach could attempt to disentangle the science from the political process.
The approach to this content analysis study has some limitations. The process by which the tweets were acquired does not provide information about the users' identity; therefore, the race or ethnic groups with which they identify were unclear. However, it is clear which race and ethnic groups were referred to in the tweets, so it provides information about how people are discussing these groups in public discourse. The API system does not collect the tweets within the context of a conversation or an algorithm for the user, so some meaning may have been unclear and individual exposure to tweets is undetermined. Yet, the number of tweets collected provided an overall sentiment of the main ideas. In addition, a team of eight people discussed all tweets with discrepancies in order to grapple with the interpretation. While the study used some previously published vaccine-related terms [39], it is possible that the keywords used for this data collection may not completely reflect the terminology used to discuss the topic, therefore not capturing all relevant tweets.

Conclusions
Due to the health disparities present in the United States related to race, it is vital to distribute the COVID-19 vaccine in communities of color. As seen in the results of this study, a segment of Twitter users expressed their hesitancy toward the vaccine. This was attributed to several causes as discussed, but we recommend the practice of utilizing qualitative analysis of Twitter data to potentially contribute to the development of messaging for public health campaigns. For instance, the analysis can help with message development, and then the messages would need to be tested with the intended audiences before dissemination. Our study showed how this type of analysis can offer nuanced insight about public sentiment, as well as areas to target in a media outreach, from who should be the messenger to the specific topic. Minimally, public health practitioners can encourage and amplify positive sentiments and affirmations of the vaccine on social media. A coordinated effort could potentially result in contributing to attitude change about vaccine uptake.  The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We are grateful to the California Center for Population Research at UCLA (CCPR) for general support. CCPR receives population research infrastructure funding (P2C-HD041022). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Institutional Review Board Statement:
This study was determined exempt by the University of California, San Francisco Institutional Review Board .

Informed Consent Statement: Not applicable.
Data Availability Statement: Twitter data were collected using Twitter's Application Programming Interface (API). Twitter's API is free and open to the public.