COVID-19 vaccination hesitancy, misinformation and conspiracy

theories on social media: A content analysis of Twitter data

Tasmiah Nuzhath1,2, Samia Tasnim1, Rahul Kumar Sanjowal3, Nusrat Fahmida Trisha1, Mariya

Rahman1 , Farabi Mahmud4, Arif Arman4, Susmita Chakraborty1, Md Mahbub Hossain1

1
School of Public Health, Texas A&M University, College Station, TX, USA
2
Hagler Institute for Advanced Study at Texas A&M University, College Station, TX, USA
3
College of Engineering, Department of Computer Engineering, Texas A&M University

College Station, TX, USA

4
Department of Population Sciences, University of Dhaka, Dhaka, Bangladesh

Corresponding author:
Tasmiah Nuzhath
School of Public Health
Texas A&M University
College Station, TX 77840
Email: nuz_t@tamu.edu
Abstract

Background: The coronavirus disease (COVID-19) pandemic has caused a significant burden
of mortality and morbidity. A vaccine will be the most effective global preventive strategy to
end the pandemic. Studies have maintained that exposure to negative sentiments related to
vaccination on social media increase vaccine hesitancy and refusal. Despite the influence
social media has on vaccination behavior, there is a lack of studies exploring the public's
exposure to misinformation, conspiracy theories, and concerns on Twitter regarding a potential
COVID-19 vaccination.

Objective: The study aims to identify the major thematic areas about a potential COVID-19
vaccination based on the contents of Twitter data.

Method: We retrieved 1,286,659 publicly available tweets posted within the timeline of July
19, 2020, to August 19, 2020, leveraging the Twint package. Following the extraction, we used
Latent Dirichlet Allocation for topic modelling and identified 20 topics discussed in the tweets.
We selected 4,868 tweets with the highest probability of belonging in the specific cluster and
manually labeled as positive, negative, neutral, or irrelevant. The negative tweets were further
assigned to a theme and subtheme based on the content

Result: The negative tweets were further categorized into 7 major themes: "safety and
effectiveness,” "misinformation,” "conspiracy theories,” "mistrust of scientists and
governments,” "lack of intent to get a COVID-19 vaccine,” "freedom of choice," and "religious
beliefs. Negative tweets predominantly consisted of misleading statements (n=424) that
immunization against coronavirus is unnecessary as the survival rate is high. The second most
prevalent theme to emerge was tweets constituting safety and effectiveness related concerns
(n=276) regarding the side effects of a potential vaccine developed at an unprecedented speed.

Conclusion: Our findings suggest a need to formulate a large-scale vaccine communication

plan that will address the safety concerns and debunk the misinformation and conspiracy
theories spreading across social media platforms, increasing the public's acceptance of a
COVID-19 vaccination.
Introduction

COVID-19, an acute respiratory illness caused by the SARS-CoV-2 virus, was declared a

pandemic on March 11, 2020 [1]. The pandemic has caused a significant burden with over 69

million infections and an estimated 1.5 million deaths worldwide as of 10 December 2020 [2].

Globally, countries responded to the crisis by implementing risk mitigation strategies to contain

the spread of the virus, which included social distancing and mobility restrictions [3]. However,

these measures have led to catastrophic consequences to the global economy and healthcare

systems and have also adversely affected the population’s physical, social, and psychological

health. Therefore, a vaccine will be the most effective global preventive strategy for limiting

the spread of the virus [4,5]. COVID-19 vaccine coverage of ~67% is estimated to be sufficient

to achieve herd immunity based on the assumption that the basic reproductive number (R0) of

the virus is three [6,7].

Evidence suggests that vaccines have been effective in preventing infectious diseases [8]. The

effectiveness of vaccines depends on people’s immunization uptake; however, the availability

of vaccines does not always translate into utilization [9]. This is because there has been a

decline in vaccine confidence that has resulted in vaccination hesitancy and refusal, which has

further contributed to the global decline in vaccination uptake and has caused infectious disease

outbreaks worldwide [9,10,11]. Therefore, vaccine hesitancy or refusal poses serious global

health risks. Consequently, the World Health Organization (WHO) has included vaccine

hesitancy in its top 10 global health threats for 2019 [12].

A key factor in low vaccine confidence is exposure to misinformation and falsehoods [13].

Misinformation and negative sentiments are highly contagious and can potentially reduce

vaccine uptake rates [14]. An example is the spread of mistrust, and unsubstantiated theories
on vaccine safety against the 2009 swine flu outbreak had a significant negative impact on

people’s willingness to be vaccinated [15].

Multiple potential COVID-19 vaccine candidates show promising phase 3 clinical trial results,

and several COVID-19 vaccines have received emergency authorization for use in particular

countries, including the United States [16,17]. However, people’s willingness to take the

COVID-19 vaccine to stop the community spread of COVID-19 is currently insufficient in

several countries, including the US, several countries from Africa, Europe, and Asia [18,19].

Globally, vaccine confidence is low, with 52% of adults stating their willingness to get

vaccinated within less than three months after the COVID-19 vaccine becomes available to all

[20]. The low confidence will then act as a barrier to achieving herd immunity, which is

essential to end the pandemic.

Therefore, there is a critical need to gain a deeper understanding of people’s perceptions of a

potential COVID-19 vaccine. Furthermore, to enhance vaccine confidence among the

population, it is essential to develop and implement strategies to mitigate the concerns and

combat the misinformation and rumors related to the safety and effectiveness of a COVID-19

vaccine.

Traditional methods such as surveys, which are extremely expensive and time-consuming,

cannot capture real-time perceptions and are subject to a social desirability bias [21]. A prompt

understanding of the people’s concerns in real-time can be obtained from social media [22].

Social media provides real-time content on people’s knowledge and opinions on several issues,

including population-level health and behavior [23]. Twitter, a popular social media platform,

can be a reliable source for assessing the public’s knowledge, personal experiences and

identifying health information needs [24]. Analyzing Twitter data thereby provides a snapshot

of health knowledge and informs targeted health information communication [24]. Its
interactive nature can also be used as an effective educational tool for health-related

interventions [24].

Social media is frequently used to disseminate vaccine-related information; however, it has

also provided easy access to incorrect information and misinformed opinions along with denial

of scientific studies, which have left concerned families susceptible to strongly voiced

misleading opinions in the media [25]. Misinformation and rumors about the current pandemic

have also been rapidly propagating on social media platforms, including Twitter, and

negatively influenced COVID-19 related health behavior [26].

Additionally, there are many anti-vaccine advocates in social media, particularly on Twitter,

where anti-vaccine beliefs comprise a substantial amount of vaccination-related tweets [27].

Several studies have maintained that exposure to negative sentiments, misinformation, and

rumors about vaccination on Twitter has increased vaccine hesitancy and refusal and a decline

in vaccination uptake [27,28,29].

Given the influence and wide-reaching capabilities of Twitter, we will be conducting a content

analysis of tweets to gain a fuller understanding of the perception people have of a COVID-19

vaccination. The aim of the study is to determine the public’s sentiment towards the

immunization and to identify the major thematic areas of safety concerns, misinformation, and

rumors related to the vaccine on Twitter. To the best of our knowledge, this is the first study to

use Twitter data to understand people’s opinion of a potential vaccination against coronavirus.

The new knowledge gained from this study is essential to effectively combat the spread of

misinformation, to designing policy strategies to gain public trust, and increasing the rate of

COVID-19 immunization uptake to control the pandemic globally, once one or more COVID-

19 vaccines become available to the public.

Method:

Search Strategy and Data Collection

We retrieved data using the Twint package, a publicly available, advanced Twitter scraping

tool written in python [30]. We retrieved 1,286,659 tweets posted within the timeline of July

19, 2020, to August 19, 2020, and contained the keywords: “vaccine” or “#vaccine” in it. This

period overlapped with announcements of 3 potential COVID-19 vaccines progressing into

phase 3 trial and the approval of a Russian vaccine, Sputnik V [16]. This study was approved

by the Texas A&M University Institutional Review Board on December 2020 (IRB 2020-

1417).

As the tweets can contain emojis and newlines, we converted tweets to the lemmatized format

to process them using the Latent Dirichlet Allocation algorithm [31]. First, we removed all the

URLs from the tweet texts. We then converted the tweet text to lowercase, removing leading

and trailing whitespaces, tokenizing and removing the accents from the tweet texts.

After the texts were cleaned by the simple preprocessor, we lemmatized the texts. The

lemmatization process converts the word to its basic form based on the context. For example -

if we lemmatize “vaccines” it would return the form “vaccine.” When the lemmatization

process was completed, we removed the stop words such as, “the, such, as, etc.,” from the

remaining tweet texts.

To create the model, we needed to develop the corpora by taking all the unique lemmatized

words from the text and forming the bag of words. By removing the usernames, we de-

identified the tweets. Our dataset included tweets in English Language only.

Topic Modelling Technique

Following data extraction, we employed Topic Modelling to screen tweets for inclusion and to

identify the common topics, excluding any posts that are not relevant. Topic modeling is an
unsupervised machine learning technique for identifying clusters based on the co-occurrence

of words and has been previously used in studies related to vaccination and social media

[31,32]. We used the Latent Dirichlet Allocation (LDA), a topic modeling algorithm, to

determine the relevant tweets compared to the irrelevant tweets containing COVID-19

vaccination keywords [31]. To determine this number of clusters, we ran a sensitivity test with

multiple values for a number of clusters but got a saturated coherence score near 20 topics.

After the LDA returns the clusters, we assigned each tweet to a corresponding topic. From each

of the 20 topics, we selected the first 20 tweets with the highest probability of belonging in the

specific topic for the initial coding process.

Coding Strategy

Codes were adapted from analyses of Twitter contents and themes previously identified in the

literature [21,32,33]. We refined these codes through an inductive approach that involved

manually evaluating the top 20 tweets from each of the 20 topics to supplement the codebook

we had created based on the previous literature. A final codebook presented the definition of

the sentiment categories, the themes and sub-themes for the negative contents.

Then we selected 5000 tweets from our original dataset to conduct content analysis. Previous

research has shown that qualitative content analysis can be done using less than a thousand

tweets [34]. However, we used a larger set of tweets to ensure optimal representation of the

public’s perspectives on social media. After extracting these tweets, duplicate entries were

discarded, and 4,868 tweets were labeled by 3 coders (RKS, MR and NFT). The coders first

manually assigned sentiments of the content to identify if the tweet was positive, negative,

neutral, or irrelevant. The negative tweets were then assigned to a theme and subtheme based

on the content. 10% of tweets were randomly selected and split between two additional
researchers (TN,ST) to check the accuracy of this annotation,. Coder agreement was confirmed

by determining inter-rater reliability (Cohen's kappa score was 0.80).

Figure 2 provides a summary of the above steps undertaken to collect the dataset for content

analysis.

Figure 1: Flowchart for Data Processing

Tweets retrieved posted within the

timeline of July 19, 2020 to August 19,
2020 using Twint package
(n = 1.1 million)

Used Topic Modelling to screen tweets for

inclusion and to identify the common topics,
number of topics = 20

Eligible Tweets with the highest probability Excluded retweets and duplicates
from each topic for content analysis (n=132)
(n =5000 tweets)

Final sample of tweets selected for

content analysis (n=4868)
Definition of terms used to code tweet sentiment

We determined categories for the tweets based on previous research that used Twitter to assess

the sentiment of public health topics [35]. A codebook was developed that included relevant

definitions of categories. Tweets were classified based on their sentiment toward vaccines:

positive, negative, neutral or irrelevant categories.

1) Positive: A tweet was considered positive if it mentioned research discoveries related

to the COVID-19 vaccine trial, if a person is seeking information on how to volunteer

for the clinical trial, or if the tweet reflected a positive attitude toward the COVID-19

vaccination.

"Don't worry, we will return to normal with vaccines worldwide.

#NoNewNormal #ScienceWillWin"

2) Negative: A tweet was considered as negative if it mentioned vaccines as unnecessary,

provided misinformation, conspiracy theories, or raised safety and effectiveness

concerns about vaccine against coronavirus.

"So, the plan is still to wait for an untested vaccine that was produced in less

than a year when it normally takes 10-years?”

3) Neutral: Tweets were considered neutral if the tweet’s overall sentiment was neutral, if

tweets provided information that did not pertain to emotional words or mentioned a

covid-19 vaccine clinical trial in general.

"Phase 3 clinical trial of investigational #vaccine for #COVID19 begins

#Moderna"

4) Irrelevant: Tweets were considered irrelevant if the tweets were not related to COVID-

vaccine or if the context is not understood. Posts with pictures or URLs only without

providing a context in the tweet were also considered as irrelevant.

"Fear is the greatest pandemic and love is the greatest vaccine"

Results

Content analysis: classifying tweets into categories

1306 out of 4868 tweets were categorized as negative, out of which the majority (32.47%)

tweets were classified as misinformation, followed by tweets expressing concern regarding

vaccine safety and effectiveness (21.13%), tweets related to conspiracy theories on potential

COVID-19 vaccines (17.99%), and mistrust of scientists and governments (12.86%).

Additionally, 10.64 % of tweets expressed a lack of willingness to get the COVID-19 vaccine

when it became available, and a future mandatory COVID-19 vaccine program was perceived

as a limitation to freedom of choice in 3.60 % of tweets. Lastly, 1.30% of tweets expressed

concerns or restrictions regarding a potential COVID-19 immunization due to religious beliefs.

Figure 2: Percentage of tweets by Themes

The following seven themes and related sub-themes emerged from our content analysis and are
summarized in Table 2.

Table 2: Themes and Sub-themes of tweets expressing negative sentiments

Theme Sub-Theme Definition

Safety and Vaccine will have side effects Vaccines against coronavirus is harmful or will cause
effectiveness (n = 140, 50.72%) injuries.
(n =276) Fast paced Vaccine Concerns that the first COVID19 vaccines would be
Development released far ahead of the traditional timeframe.
(n =34,12.32%)
Vaccine will be Ineffective A potential COVID-19 vaccine will create only
(n =102,36.96% ) temporary/ineffective immunity
Conspiracy Profit from developing a Production of a potential covid-19 vaccination is
theories COVID-19 vaccine motivated only by a quest for profit by big pharma
(n =235) (n =121, 48.79%) companies

Unusual theories Unique theories about the vaccine against coronavirus.

(n =127, 51.21%)
Misinformation Falsehoods Unsupported statements made regarding a future
and falsehoods (n =81,19.10%) COVID-19 vaccination
(n =424)
Vaccines are untested Potential COVID-19 vaccines will not be adequately
(n =44, 10.38% ) tested for safety.

Vaccines cause illnesses of Vaccine will cause autism or other illnesses.

unknown origin
(n =21, 4.95%)
Vaccine is Unnecessary Vaccination is not necessary to protect against
(n =278, 65.57%) coronavirus.
Mistrust (n Mistrust of Scientists and Showing mistrust of scientists and vaccine advocates
=168) vaccine advocates involved in financing, developing, or distributing a
(n =105, 62.50%) potential COVID-19 vaccination
Mistrust in the government Mistrust in the governmental institutions concerning a
(n =63, 37.50%) potential COVID-19 vaccination.
Lack of intent Individual unwillingness to get Tweets expressing a user’s intention to get a vaccine or
to get a COVID- a COVID-19 vaccine discouraging others from getting vaccinated against
19 vaccine ( (n =114, 82.01%) coronavirus.
n=139)
Discouraging others to get
vaccinated against coronavirus
(n =25, 17.99%)
Freedom of Violation of individual rights Civil liberties will be violated by taking away an
choice ( n=47) (n =24, 51.06%) individual’s right to choose in getting a potential COVID-
19 vaccination
 Totalitarianism Accusations of totalitarianism were made, which
(n =23, 48.94%) included warnings that citizens would be forced to
comply with a mandatory vaccination program.
Religious Religious concerns Tweets expressing religious concerns about the
beliefs (n =17) (n =17, 100%) ingredients used in developing the vaccine or about the
vaccines in general.

Misinformation:

The predominant theme to emerge was “Misinformation” (32.47%, n =424). This theme

encompassed tweets making misleading statements that were not substantiated by evidence.

Some of the contents of the tweets under this theme are as follows: herd immunity was safer

than getting a vaccination to protect against the coronavirus, vaccines are medical experiments,

vaccines are not part of real science, COVID-19 vaccine will not work on people with obesity,

the vaccine will be only for the elderly population, vaccines don’t exist for any viruses, and

lastly, vaccines are not adequately tested. Some tweets have also expressed concern about the

process of testing vaccine safety in countries that have low coronavirus infection rates such as,

“'How could a country reporting so few cases actually be able to test a vaccine for safety and

efficacy?'".

The majority of the tweets belonging to this theme were classified under the sub-theme of

“Vaccine is unnecessary” (65.57%, n =278). In this sub-theme, tweets expressed various

unsupported information and claimed that the vaccine for COVID-19 is not necessary for

protection from coronavirus infection. For example,

"Why do we need a vaccine for a disease that has a 99.5% survival rate?"

"I already have a vaccine called a human immune system."

"Herd immunity is safer than a vaccine."

The next sub-theme was classified as “falsehood,” with 19.10% (n =81) of the tweets making

false claims about vaccination. For example,

"VACCINES ARE AN ILLUSION ... There will never be a functional vaccine\, and
 there never was a functional vaccine."

"The vaccines don't work as well on obese people... A lot of you are obese."

"Vaccines are not double blind tested with inert placebo and there aren’t any

longitudinal safety studies."

Out of the total tweets under the theme misinformation, 10.38% (n=44) of the tweets were

categorized as “vaccines are untested”. This group of tweets questioned the credibility of the

vaccine developed within such a short time and claimed that untested vaccines would be made

available for the general public.

"Is any vaccine safe to use when it was only tested for like 2 months?"

"Can't see how they would have a thoroughly tested vaccine by then? (And that really

works). Promises, promises!"

"I will not use the new vaccine\, I will wait until it has been really tested."

The last sub-theme was “Vaccine causes autism and other unknown illnesses” (4.95%, n =21).

These tweets held vaccines responsible for autism in children and claimed that vaccines cause

diseases.

"What about PTSD? Could it also be due to vaccines?"

"Vaccines will cause autism by 2032. 50% of all children will have autism."

"Vaccination causes autism and other disabilities in children.”

Safety and effectiveness:

The second common theme was “Safety and effectiveness,” comprising 21.13% (n =276) of

all tweets that expressed negative sentiment. This category of tweets expressed perceived

concerns regarding the safety and effectiveness of a vaccine developed at a rushed speed, with

some users posting tweets where they stated that they would prefer being exposed to the

coronavirus than get the vaccine. One such tweet was, "I'll take my chances with covid-19. I'm
not going to take some rushed vaccine."

Among the tweets discussing the vaccine’s safety and effectiveness, tweets expressing concern

about the potential adverse side effects of the potential covid-19 vaccine were found to be

widespread (50.72%, n =140). For example,

"A COVID vaccine may very well be the next REAL pandemic."

"Covid vaccine will kill 50 million Americans”.

"What happens when the death rate from the Vaccine is worse than the death rate from

the Virus? I fully expect that to be the case."

Next emerging sub theme was regarding the “vaccine will be ineffective”. It comprised 36.96%

(n=102) tweets among the 276 tweets under the theme “safety and effectiveness”. Majority of

the tweets in this category claimed that the vaccine will not be potent enough to protect against

covid-19 infections.

"The body does not sustain antibodies to COVID. How can there be a vaccine?"

"How can vaccines work if the virus is changing day by day! The vaccine may not work!

I believe prevention is better than vaccines and Chlorite is better than that!"

"There is no proven immunity to this virus. Medical Professionals have found that

antibodies disappeared after infection and stated that there is no guarantee that an

effective vaccine will be found."

Tweets within the sub-theme of “Fast-paced vaccine development” (12.32%, n =34)

highlighted the rapid production of the vaccine and expressed concern regarding the safety and

efficacy of a rushed vaccine.

"I think most people would trust a vaccine but a rushed though vaccine is another

matter."

"Because the vaccine development is so accelerated, we will not have all the safety data

to risk exposure of vulnerable populations to vaccines.”

 "Also, who wants a warp speed vaccine? Not me!"

Conspiracy theories:

The third theme that emerged was "Conspiracy theories," (19%, n =248) and covered a wide

range of unconventional and unfounded theories regarding the potential covid-19 vaccine.

Nearly half of the tweets (48.9%, n =121) were categorized under the sub-theme "Profit for big

corporations," and consisted of tweets that stated how the big corporations or pharmacological

companies and health care services would make enormous revenue from the production and

distribution of the vaccine. Some of the tweets also alleged that Bill Gates and Dr. Anthony

Fauci will be making a financial profit from potential COVID-19 vaccines

"The same doctors are telling you it works. Fauci and Gates don't want anything that

jeopardizes the money to be made from vaccines. You people must wake up."

"No money to be made from HCQ. Millions to be made from the ineffective vaccine."

"Big Pharma wants the vaccine, it will make some company a lot of money."

The remaining half of the tweets (51.21%, n =127) have stated various unusual theories and

are summarized in Table 3.

Table 3: Unusual Theories related to COVID-19 vaccination

Theories Tweets

Vaccine is being developed to limit or "How can someone who believes that we are overpopulated be
control population size the one to create the vaccines to keep us alive? Think smart
folks."

"Don't trust Gates. He is a Globalist who wants to kill Americans

with his vaccine."

Vaccine will contain Microchip or tracking "This one will have no problem accepting the mandatory vaccine
device when it comes out and then absolutely no problem accepting the
mandatory microchip when that comes out!!
#covididiot #sheep”
 "It's been rumored that there are nanobots in the vaccine for
contact tracing"

Vaccine will cause infertility "What percentage of people will be injured by the vaccine? What
if the vaccine rendered a percentage of people infertile?"

"Is this a backup in case the sterilization the COVID vaccine

does fails?"

5G/3G technology related to COVID-19 "How can a vaccine protect against 5G toxicity?"
infection and vaccine
"Just tell the stupid people that the vaccine only has 3G."

Vaccine is made of materials from "Most vaccines are made from murdered babies"
murdering babies or contains unusual
ingredients "Please do your research. It's not just aborted fetal tissue. There
are monkey cells, cocker spaniel cells, formaldehyde, aluminum,
and glucosphate. But please go and do your own research. Ask
your doctors for vaccine inserts."

Vaccine makers created COVID-19 "Almost makes you wonder if these vaccines were made before
the pandemic."

"Why didn't they also develop a vaccine while inventing a killing

virus?"

COVID-19 vaccine will control the mind "I already know that Covid vaccine got some mind control shit in
it"

COVID-19 vaccine is a hoax "Fake vaccine fake COVID-19."

"Fake Vaccine to Cure the Fake COVID Hoax!"

Vaccine may be laced with marijuana "They were working on a weed-based COVID-19 vaccine in that
factory."

Vaccine may turn people into zombie "You all will finally see some zombies once this vaccine is
released."

"What if the vaccine causes a zombie apocalypse?"

Vaccine will alter DNA "The COVID vaccine will be the first of its kind. An RNA vaccine
that will alter your genetic makeup and permanently change your
DNA. Nanobots will be smaller than red blood cells & injected
straight into your veins, which allows for WiFi connectivity to
merge humans and tech."

"Bill Gates openly states his vaccine has DNA altering properties
in it ( NOT a Vaccine but ILLEGAL AND UNLAWFUL roll out of
Gene Therapy), also Bill Gates vaccines have a history of
undisclosed Sterility chemicals in them, research the MILLIONS
in Africa and India he's destroyed"

Vaccine contains bleach "Me too. Trump will have a vaccine with bleach in it. He is trying
to kill us!"
 "Wouldn't trust your vaccine - it probably has bleach in it."

Vaccine will contain hormones "They’ve been adding hormones to the vaccines"

Mistrust:

The next theme was labeled as “Mistrust,” which comprised 13% ( n=168) of all negative

tweets. The majority of these tweets expressed mistrust of science, scientists, and vaccine

advocates, including Bill Gates (62.5%, n=105).

"Does anyone really believe we'll be told the truth about what's really in the vaccine

once it is released to the public? I am sure they won't."

"If it is based on science why are there 150 different vaccines?"

"Any vaccine that has anything to do with Gates and Fauci is a big NO."

The remaining tweets (37.5%, n=63) conveyed mistrust toward governments regarding their

motive towards pushing a potential Covid-19 vaccine.

"Putin has a vaccine he's going to give to Trump & he's going to try to kill us all with

it!"

"Boris can go to hell with his vaccine."

"NO WAY. I can make my own bleach vaccine at home. As if it's not bad enough that

he won't do anything to help. Why does he have to continuously get out there and spread

misinformation? #LyingTrump”

Intent to get a COVID-19 vaccine:

Tweets on the theme "Lack of intent to get a COVID-19 vaccine" (10.7%, n=140) expressed

individual or collective unwillingness to vaccinate against coronavirus in the future when a

vaccine becomes available.

The majority of the tweets (82.1%, n=115) were classified under the sub-theme of "Individual
unwillingness to vaccinate." These tweets demonstrated the public's opinion or future

compliance about receiving COVID-19 vaccine dosage in the future.

"We don't know enough about covid 19 to get the correct vaccine. and 150 possible

vaccines being made so how does anyone know it will help. Until it can be proven I'm

not putting myself at risk"

"Anyone who tries to give me a vaccine will be put down"

"WILL YOU GET A COVID-19 VACCINE AS SOON AS IT IS RELEASED? I voted NO

on WFSB Hartford. What's your vote? I will when I know its really really safe to do. I

would never do one right as its available"

The remaining tweets in this theme (17.8%, n =25) were categorized as “Discouraging others

from getting vaccinated against coronavirus.” These tweets either concluded their community

would not vaccinate or expressed negative remarks regarding the vaccine and demotivated

others to take it.

"Run as far as you can from that covid-19 vaccine."

"WAKE UP AMERICA DO NOT TAKE THE VACCINE!"

"Say no to vaccines and no to being forced to vaccinate."

Freedom of choice:

The next theme emerging was labeled "Freedom of choice" and consisted of 3.6% (n =47) of

the total tweets classified as expressing negative sentiments. In this theme, half the tweets

(51%, n=24) were classified under the sub-theme- "Violation of individual rights," which

expressed concerns about the violation of civil rights due to any future mandatory covid-19

vaccine programs.

"Who cares. As Americans, we should be able to choose our own medical treatments.
 My body is my choice. But they're pushing for vaccine that most of us won't take"

"@dougducey @realDrBirx @realDonaldTrump Please do not make whatever vaccine

they come up with mandatory! Let "we the people" have our freedom of choice to choose

if we want it or not. Thank-you"

"And more importantly, governments had better NOT mandate taking a vaccine for this

COVID-19 thing that hasn't proven to be The Plague like everyone wants to think it is."

The remaining tweets (48.9%, n =23) were classified under the sub-theme of “Totalitarianism.”

It consisted of tweets that considered mandatory vaccines to be a case of subservience. For

example,

"If you don't take the vaccine you won't be able to travel, go to stores, buy gas. Even

when there isn't a statewide mandate on masks, stores are demanding that you wear

the mask. Corporations are policing us now. Others are policing us now."

Religious beliefs

Religious concerns about a potential vaccine consisted of 1.3% (n =17) of the total tweets

categorized as negative. Some examples are given below,

"That vaccine will be the mark of the Beast!"

"That's so that we can be the guinea pigs for their demonic vaccine."

"All vaccines are made by the kuffar and contain pork enzymes, including the one used

for hajj (for Asians)."

1 Discussion

2 The findings of the current research highlight the plethora of misinformation and conspiracy

3 theories surrounding a potential COVID-19 vaccination spread through popular social media

4 platform -Twitter. Our study undertook a qualitative content analysis of Twitter posts to better

5 understand the public's perceptions of and exposure to information on a potential immunization

6 against coronavirus. We classified tweets into different sentiment categories and analyzed posts

7 expressing negative sentiments towards a potential coronavirus vaccine. Through the analysis,

8 we identified seven major themes: "safety and effectiveness," "misinformation," "conspiracy

9 theories," "mistrust of scientists and governments," "lack of intent to get a COVID-19 vaccine,"

10 "freedom of choice," and "religious beliefs." Negative tweets predominantly consisted of

11 misleading statements that immunization against coronavirus is not necessary as the likelihood

12 that a COVID-19 infection will result in death is low. The second most common tweets

13 expressed concern regarding the side effects of a potential vaccine being developed at an

14 unprecedented speed. Our results also show that a significant number of tweets spread

15 conspiracy theories regarding the ingredients used in developing vaccines, while others alleged

16 that large pharmaceutical companies had primarily financial rather than health-related motives

17 for developing vaccines. We have also found tweets that showed mistrust towards science and

18 scientists, particularly Dr. Anthony Fauci along with vaccine advocates such as Bill Gates, who

19 is allegedly involved in developing COVID-19 vaccines with a microchip to promote the

20 establishment of a global surveillance network. Our findings are consistent with recent studies

21 that have focused on understanding the public's willingness to get immunized against

22 coronavirus once a vaccine becomes available [18,19,20,36]. This suggests a need to formulate

23 a large-scale vaccine communication plan which will address the safety concerns and debunk

24 the misinformation and conspiracy theories spreading across social media platforms to increase

25 the public's acceptance of a COVID-19 vaccination.

26 Vaccine-related trust is a global health concern. In recent years, many countries reported having

27 new cases of vaccine-preventable diseases such as measles, mumps, and rubella, which can be

28 attributable to many psychosocial factors, including a lack of trust in vaccines in many contexts

29 [37]. Often a low number of such cases for several decades might have resulted in a public

30 perception that those vaccines are no longer useful. In this way, the success of vaccination

31 programs might have affected the credibility of those programs in a paradoxical manner, which

32 must be addressed for preventing infectious diseases and associated population health burden.

33 Despite a lack of credible evidence, some people may suggest that vaccines may cause certain

34 illnesses or adverse biological consequences, leading to anti-vaccine movements that have

35 gained momentum in different populations [38]. Such misinformation, often coupled with a

36 general distrust of science, has been accelerated during the COVID-19 pandemic. In this way,

37 a critical and persistent public health issue such as vaccine distrust is likely to affect the ongoing

38 efforts by the research community, vaccine manufacturers, and regulatory agencies who are

39 working together to ensure the safety and efficacy of COVID-19 vaccines. The ongoing

40 scientific developments should be widely communicated to restore the public's trust in vaccines

41 and vaccination programs and engage the general public, which may prevent such

42 misconceptions [39].

43 Vaccine related conspiracy theories may gain traction when different opinion leaders spread

44 them. For instance, in 1990, a conspiracy theory was spread that a tetanus vaccine was used in

45 Cameroon to sterilize girls and women [40]. As a result, vaccination rates against tetanus

46 declined to as low as 13% in the following five years. Another conspiracy theory related to

47 vaccination in Nigeria during 2003 suggested an oral polio vaccine (OPV) was an American

48 plot to sterilize Muslim girls and spread the human immunodeficiency virus [41].

49 Such conspiracy theories and misleading statements resulted in the suspension of OPV use in

50 five northern Nigerian states. The impacts of such conspiracy theories were enormous.
51 Between 2002 and 2006, wild poliovirus cases in the country increased fivefold. In both

52 cases, the conspiracy theories were endorsed by local opinion leaders that appeared to have

53 some credibility to the general public, who might not have had access to accurate

54 information.

55 Similar studies from Pakistan suggest that those parents who had refused OPV during annual

56 OPV campaigns attributed their refusal commonly to the perception that the vaccine was

57 associated with fertility and disapproval of the religious leaders in their communities [42]. In

58 low-income contexts where institutions are not optimally functioning, and local opinion

59 leaders shape the public perceptions and practices, vaccine hesitancy may have unique

60 psychosocial correlates that should be examined in the context of the COVID-19 pandemic.

61 Vaccine hesitancy can also be associated with a lack of trust in conventional medicine and

62 health systems. A study among 5200 Spanish participants found that distrust in conventional

63 medicine was associated with vaccine hesitancy, markedly higher in people who used

64 complementary and alternative medicine [43]. This may offer valuable insights into people's

65 health behavior who do not trust conventional medications or vaccines but have some trust in

66 alternative therapies. In countries such as India and China, where alternative medicine is

67 widely used, [44] public health authorities may partner with those providers to empower the

68 populations with scientifically accurate information and deliver vaccines, thereby establishing

69 trust in public health measures. Local healthcare providers, both formal and informal, should

70 be engaged in community-based participatory planning for the communities' health goals and

71 how they can collaborate and serve respective populations addressing the local concerns and

72 challenges. Future research is needed to better understand such local issues that are likely to

73 be unique in different contexts and people. Perhaps, such knowledge would complement the

74 findings of the current study and enable the researchers to understand how different
75 psychosocial factors interact across populations and result in a varying range of concerns

76 related to vaccines and vaccination programs.

77 Effective health communication would be necessary for addressing complex issues such as

78 vaccine hesitancy during and after this pandemic. A robust public health framework

79 involving regulatory authorities and well-established activities may promote health during

80 public health emergencies. It is necessary to strengthen such measures globally and allow

81 institutions to exchange their expertise to encourage scientific rigor of authorizing vaccines.

82

83 Limitations

84 Firstly, we limited our sample size due to the labor-intensive nature of manual coding therefore

85 our finding may not represent all the topics discussed on Twitter relevant to COVID-19

86 vaccination. Secondly, due to the rapid changes in information related to COVID-19 vaccine,

87 the categorization of the tweets may be subject to bias. We removed URLs and excluded videos

88 and images of tweets without considering the information it contained. Finally, our search

89 timeframe was limited and therefore might not capture the recent topics discussed on social

90 media. Thus, there is a need for future research to continue monitoring the public's opinion as

91 the COVID-19 vaccines become available in order to monitor the misinformation being shared

92 on the social media platforms and enhance our understanding of vaccination hesitancy

93 surrounding a COVID-19 immunization.

94
95 Conclusion:

96 Anti-vaccine movements have always been vocal online [25], and the COVID-19 “infodemic”

97 has significantly compromised the prevention efforts [26]. If the current misinformation surge

98 across the web is not nipped in the bud, the vaccination movement will likely suffer,

99 endangering public safety. Public health organizations should recognize this imminent threat

100 and launch health education programs using traditional and social media platforms to immunize

101 the public against misinformation. Health care professionals must learn to adapt to the evolving

102 scientific evidence and relevant misinformation and safety concerns associated with the

103 vaccine to allay the fears and aversions regarding this rapidly developed immunization.

104 A global collaboration may enable the public health authorities to address common concerns

105 raised by communities and population groups globally. As Twitter and other social media

106 continue to evolve, the efforts to communicate effective health information should be based on

107 scientific statements that have global consensus among the researchers, regulatory authorities,

108 and other key stakeholders. Media partnerships should be strengthened to control inaccurate

109 and hazardous media contents and to disseminate credible information targeted to promote

110 uptake of a potential covid-19 vaccine to end the pandemic.

111

112 Conflict of interest

113 The authors have no potential conflicts of interest to declare.

114 Acknowledgment

115 None.

116 Funding

117 No financial support was received at any stage of conducting this study.
118 References

119 1. World Health Organization. WHO Director-General's opening remarks at the media

120 briefing on COVID-19-11 March 2020.

121 https://www.who.int/director-general/speeches/detail/who-director-general-s-opening-

122 remarks-at-the-media-briefing-on-covid-19---11-march-2020. Published March 11,

123 2020. Accessed April 01 2020.

124 2. Johns Hopkins University & Medicine. Covid-19 Map - Johns Hopkins Coronavirus

125 Resource Center, https://coronavirus.jhu.edu/map.html; 2020 [Accessed December 10

126 2020]

127 3. de Bruin YB, Lequarre AS, McCourt J, Clevestig P, Pigazzani F, Jeddi MZ, Goulart

128 M. Initial impacts of global risk mitigation measures taken during the combatting of

129 the COVID-19 pandemic. Safety Science. 2020;104773.

130 https://doi.org/10.1016/j.ssci.2020.104773

131 4. Dinleyici EC, Borrow R, Safadi AP, van Damme P, Munoz FM. Vaccines and routine

132 immunization strategies during the COVID-19 pandemic. Human vaccines &

133 immunotherapeutics. 2020;1-8. DOI: 10.1080/21645515.2020.1804776

134 5. Wang J, Peng Y, Xu H, Cui Z, Williams RO. The COVID-19 vaccine race: challenges

135 and opportunities in vaccine formulation. AAPS PharmSciTech. 2020;21(6), 1-12.

136 doi: 10.1208/s12249-020-01744-7

137 6. Randolph HE, Barreiro LB. Herd Immunity: Understanding COVID-19. Immunity.

138 2020;52(5), 737-741. DOI: 10.1016/j.immuni.2020.04.012

139 7. Kwok KO, Lai F, Wei WI, Wong SY, Tang JW. Herd immunity–estimating the level

140 required to halt the COVID-19 epidemics in affected countries. Journal of Infection.

141 2020;80(6):e32-3. DOI: 10.1016/j.jinf.2020.03.027

142 8. Andre FE, Booy R, Bock HL, Clemens J, Datta SK, John TJ, Lee BW, Lolekha S,

143 Peltola H, Ruff TA, Santosham M. Vaccination greatly reduces disease, disability,

144 death and inequity worldwide. Bulletin of the World health organization.

145 2008;86:140-6. https://doi.org/10.2471/BLT.07.040089

146 9. Phadke VK, Bednarczyk RA, Salmon DA, Omer SB. Association between vaccine

147 refusal and vaccine-preventable diseases in the United States: a review of measles and

148 pertussis. Jama. 2016;315(11):1149-58. DOI: 10.1001/jama.2016.1353

149 10. Larson HJ, Jarrett C, Eckersberger E, Smith DM, Paterson P. Understanding vaccine

150 hesitancy around vaccines and vaccination from a global perspective: a systematic

151 review of published literature, 2007–2012. Vaccine. 2014;32(19):2150-9.

152 https://doi.org/10.1016/j.vaccine.2014.01.081

153 11. Hotez PJ, Nuzhath T, Colwell B. Combating vaccine hesitancy and other 21st century

154 social determinants in the global fight against measles. Current Opinion in Virology.

155 2020;41:1-7. https://doi.org/10.1016/j.coviro.2020.01.001

156 12. World Health Organization. Ten threats to global health in 2019.

157 https://www.who.int/news-room/spotlight/ten-threats-to-global-health-in-2019.

158 Published March 21, 2019 Accessed November 10, 2020.

159 13. Badur S, Ota M, Öztürk S, Adegbola R, Dutta A. Vaccine confidence: the keys to

160 restoring trust. Human vaccines & immunotherapeutics. 2020;16(5):1007-17.

161 https://doi.org/10.1080/21645515.2020.1740559

162 14. Salathé M, Vu DQ, Khandelwal S, Hunter DR. The dynamics of health behavior

163 sentiments on a large online social network. EPJ Data Science. 2013;2(1):1-2.

164 https://doi.org/10.1140/epjds16

165 15. Cohen J, Enserink M. As swine flu circles globe, scientists grapple with basic

166 questions. Science. 2009; 572-573. DOI: 10.1126/science.324_572

167 16. Zimmer C, Corum J, Wee S. Coronavirus Vaccine Tracker The New York Times.

168 https://www.nytimes.com/interactive/2020/science/coronavirus-vaccine-tracker.html.

169 Updated December 5, 2020. Accessed December 7, 2020.

170 17. Ledford H, Cyranoski D, Noorden RV. The UK has approved a COVID vaccine —

171 here’s what scientists now want to know. Nature.

172 https://www.nature.com/articles/d41586-020-03441-8. Published December 3, 2020.

173 Accessed December 7, 2020.

174 18. Lazarus JV, Ratzan SC, Palayew A, Gostin LO, Larson HJ, Rabin K, Kimball S, El-

175 Mohandes A. A global survey of potential acceptance of a COVID-19 vaccine. Nature

176 medicine. 2020;1-4. https://doi.org/10.1038/s41591-020-1124-9

177 19. Neumann-Böhme S, Varghese NE, Sabat I, Barros PP, Brouwer W, van Exel J,

178 Schreyögg J, Stargardt T. Once we have it, will we use it? A European survey on

179 willingness to be vaccinated against COVID-19. 2020;1-6.

180 https://doi.org/10.1007/s10198-020-01208-6

181 20. Boyon N. COVID-19 vaccination intent is decreasing globally. Ipsos.

182 https://www.ipsos.com/en/global-attitudes-covid-19-vaccine-october-2020. Published

183 November 5, 2020. Accessed December 7, 2020.

184 21. Chew C, Eysenbach G. Pandemics in the age of Twitter: content analysis of Tweets

185 during the 2009 H1N1 outbreak. PLoS One. 2010;5(11):e14118.

186 https://doi.org/10.1371/journal.pone.0014118

187 22. Dredze M, Broniatowski D, Smith M, Hilyard K. Understanding vaccine refusal: why

188 we need social media now. Am J Prev Med. 2016;50(4):550–552.

189 doi: 10.1016/j.amepre.2015.10.002

190 23. Odlum M, Yoon S. Health information needs and health seeking behavior during the

191 2014-2016 Ebola outbreak: A Twitter content analysis. PLoS currents. 2018;10.
192 doi: 10.1371/currents.outbreaks.fa814fb2bec36e29b718ab6af66124fa.

193 24. Sinnenberg L, Buttenheim AM, Padrez K, Mancheno C, Ungar L, Merchant RM.

194 Twitter as a tool for health research: a systematic review. American journal of public

195 health. 2017;107(1):e1-8. https://doi.org/10.2105/AJPH.2016.303512

196 25. Dubé E, Vivion M, MacDonald NE. Vaccine hesitancy, vaccine refusal and the anti-

197 vaccine movement: influence, impact and implications. Expert Rev Vaccines.

198 2015;14(1):99–117. https://doi.org/10.1586/14760584.2015.964212

199 26. Tasnim S, Hossain MM, Mazumder H. Impact of rumors and misinformation on

200 COVID-19 in social media. Journal of preventive medicine and public health.

201 2020;53(3):171-4. DOI: 10.3961/jpmph.20.094

202 27. Tomeny TS, Vargo CJ, El-Toukhy S. Geographic and demographic correlates of

203 autism-related anti-vaccine beliefs on Twitter, 2009–15. Soc Sci Med. 2017;191:168–

204 175. https://doi.org/10.1016/j.socscimed.2017.08.041

205 28. Dyda A, Shah Z, Surian D, Martin P, Coiera E, Dey A, Leask J, Dunn AG. HPV

206 vaccine coverage in Australia and associations with HPV vaccine information

207 exposure among Australian Twitter users. Human vaccines & immunotherapeutics.

208 2019;15(7-8):1488-95. https://doi.org/10.1080/21645515.2019.1596712

209 29. Dunn AG, Surian D, Leask J, Dey A, Mandl KD, Coiera E. Mapping information

210 exposure on social media to explain differences in HPV vaccine coverage in the

211 United States. Vaccine. 2017;35(23):3033-40.

212 https://doi.org/10.1016/j.vaccine.2017.04.060

213 30. Zacharias C. TWINT - Twitter Intelligence Tool. https://pypi.org/project/twint/.

214 Published April 29, 2020. Accessed 22 Nov. 2020.

215 31. Blei DM, Ng AY, Jordan MI. Latent dirichlet allocation. Journal of machine Learning

216 research. 2003;993-1022.

217 32. Surian D, Nguyen DQ, Kennedy G, Johnson M, Coiera E, Dunn AG. Characterizing

218 Twitter discussions about HPV vaccines using topic modeling and community

219 detection. Journal of medical Internet research. 2016;18(8):e232.

220 DOI: 10.2196/jmir.6045

221 33. Wolfe RM, Sharp LK, Lipsky MS. Content and design attributes of antivaccination

222 web sites. Jama. 2002;287(24):3245-8. DOI: 10.1001/jama.287.24.3245

223 34. Lee JL, DeCamp M, Dredze M, Chisolm MS, Berger ZD. What are health-related

224 users tweeting? A qualitative content analysis of health-related users and their

225 messages on twitter. J Med Internet Res. 2014;16(10):e237. Published 2014 Oct 15.

226 doi:10.2196/jmir.3765. DOI: 10.2196/jmir.3765

227 35. Mamidi R, Miller M, Banerjee T, Romine W, Sheth A. Identifying key topics bearing

228 negative sentiment on Twitter: insights concerning the 2015-2016 Zika epidemic.

229 JMIR Public Health and Surveillance. 2019;5(2):e11036. DOI: 10.2196/11036

230 36. Hotez, P. J. (2020). COVID19 meets the Antivaccine Movement. Microbes and

231 Infection. DOI: 10.1016/j.micinf.2020.05.010

232 37. Dermody, T. S., DiMaio, D., & Enquist, L. W. (2020). Vaccine Safety, Efficacy, and

233 Trust Take Time. Annual Review of Virology, 8. DOI: 10.1146/annurev-vi-08-

234 102220-100001

235 38. Smith, T. C. (2017, July). Vaccine rejection and hesitancy: a review and call to action.

236 In Open forum infectious diseases (Vol. 4, No. 3). Oxford University Press.

237 DOI: 10.1093/ofid/ofx146

238 39. National Vaccine Advisory Committee. (2013). Protecting the public's health: critical

239 functions of the Section 317 Immunization Program—a report of the National

240 Vaccine Advisory Committee. DOI: 10.1177/003335491312800203

241 40. Feldman-Savelsberg P, Ndonko F, Schmidt-Ehry B. Sterilizing vaccines or the

242 politics of the womb: retrospective study of a rumor in Cameroon. Med Anthropol

243 Q. 2000;14:159–79. DOI: 10.1525/maq.2000.14.2.159

244 41. Ghinai I, Willott C, Dadari I, Larson H. Listening to the rumours: what the northern

245 Nigeria polio vaccine boycott can tell us ten years on. Glob Public

246 Health. 2013;8:1138–50. DOI: 10.1080/17441692.2013.859720

247 42. Murakami H, Kobayashi M, Hachiya M, Khan Z, Hassan S, Sakurada S. Refusal of

248 oral polio vaccine in northwestern Pakistan: a qualitative and quantitative study.

249 Vaccine. 2014;32:1382–87. DOI: 10.1016/j.vaccine.2014.01.018

250 43. Hornsey, Matthew J., Josep Lobera, and Celia Díaz-Catalán. "Vaccine hesitancy is

251 strongly associated with distrust of conventional medicine, and only weakly

252 associated with trust in alternative medicine." Social Science & Medicine (2020):

253 113019. DOI: 10.1016/j.socscimed.2020.113019

254 44. Patwardhan B, Warude D, Pushpangadan P, Bhatt N. Ayurveda and traditional

255 Chinese medicine: a comparative overview. Evid Based Complement Alternat Med.

256 2005;2(4):465-473. doi:10.1093/ecam/neh140. DOI: 10.1093/ecam/neh140