1 Introduction

Vaccines remain the cornerstones in fighting most vaccine-preventable diseases; thus, future pandemics could spread, prolong, and worsen if vaccine campaigns do not reach their targets. The World Health Organization (WHO) and other health organizations recommend vaccine coverage should reach at least 70% of the population to achieve the herd immunity goal [1]. To achieve this goal, vaccines must be accepted and trusted by at least 55% of the population globally; this was the first concern of the involved parties during COVID-19 pandemic [2,3,4]. Vaccines are usually produced using either attenuated or killed microorganisms, their outer surface proteins, or ingested or inhaled toxins to boost the immune system and create antibodies that identify and counteract infecting microorganisms [5, 6]. Thus, effective and safe vaccines are crucial to controlling the COVID-19 pandemic [7, 8]. Progress in conceiving the Severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV) vaccines has been noted by many studies [9,10,11,12], which saved time for developing the currently marketed SARS-CoV-2 vaccines and provided platforms for the future across-the-board application of SARS-CoV-2 vaccines.

However, the absence of reliable and accessible information about COVID-19 and its vaccines led to a rise in information production containing both correct and fabricated news about the virus and vaccinations, thus, creating another threat defined as an "infodemic" [13, 14]. An "Infodemic" is an overload of data containing false and reliable information; that imposes a challenge for the public to distinguish accurate from inaccurate information [15, 16]. The COVID-19 infodemic waves started by spreading conspiracy theories about the virus's origin. One such theory claimed that the virus was a laboratory-made bioweapon that was developed in Wuhan, China [2]. (Mis)information about COVID-19 outbreak and vaccination was typically spread via virtual channels (e.g., news channels and social media platforms) rather than non-digital platforms [17]. The rapid and global nature of virtual communication channels is particularly conducive to the spread of misinformation (e.g., "fake news") [2].

A limited number of studies have explored the nature of misinformation and disinformation on online social platforms and their impact on people's hesitation, acceptance, and willingness to receive the COVID-19 vaccine in Arab countries. However, investigations into the infodemic, involving disinformation on social media and online platforms, have included examinations of fake news in Arabic content on Twitter during the pandemic [18]. Other studies have explored conspiracy theories surrounding the virus [19] and the COVID-19 vaccine on Twitter [20] and YouTube [18]. A study investigated Arabic content by examining anti-vaccine hashtags on Twitter during the pandemic. Findings concluded that the reasons behind vaccine hesitancy and opposition among Arabs included conspiracy theories about vaccines, suspicions regarding vaccine efficacy and associated risks, violations of civil liberties and freedom, and a lack of trust in official organizations and government [19]. Further studies are recommended in the Middle East and North Africa (MENA) Region to study and investigate the public confidence in the presented vaccine and determine which factors influence vaccine hesitancy and refusal [20].

Vaccine hesitancy is a serious public health challenge and it is defined as a delay or refusal of vaccination despite its availability by people for themselves and their family members [21]. The public vaccine hesitancy, delay and opposition of COVID-19 vaccination in Yemen remains unknown for several reasons, including the country's limited testing and monitoring capacity in the northern region [4, 22]. A principal reason is that in areas controlled by the de-facto authority (DFA) in the north of Yemen (also known as the Houthis), authorities deny the existence of COVID-19 virus, and no testing for the virus is taking place. Over 80% of the population—around 30 million people live in the northern governorates [22]. Only the southern governorates benefited from the vaccine rollout in Yemen, whereas on Dec 2, 2021 (4,774,000) vaccine doses were allocated under the COVID-19 Vaccines Global Access (COVAX) initiative [4, 22]. Until January 2022, only 2.6% of Yemenis have received their first COVID-19 vaccine dose until [22]; this meager percentage is due to the highly prominent Infodemic in Yemen, where political leaders share misinformation and conspiracy theories around the vaccine [23,24,25]. The Infodemic, on top of a strained healthcare system resulting from the ongoing civil war, humanitarian crisis [26], and simultaneous infectious disease outbreaks (e.g., cholera, diphtheria, and dengue [27]), increased fear, panic, and stress in the Yemeni society, particularly among healthcare workers (HCWs). Furthermore, Yemenis who are non-university educated, with low income, and living in rural areas are highly susceptible to misinformation about COVID-19 and correspondently showed hesitancy to take the COVID-19 vaccine [25].

In an attempt to contribute to the growing body of literature and scientific studies that aims at understanding the vaccine hesitancy related to infodemic in Yemen; our study aims to answer How does misinformation, driven by unaddressed information voids and lack of risk communication, impact vaccine acceptance and public health actions in conflict settings like Yemen during the COVID-19 pandemic?? Additionally, how did Yemenis' social behavior towards COVID-19 vaccine uptake and their hesitancy towards vaccination, both online and offline, contribute to the infodemic? Furthermore, what were the primary reasons hindering Yemenis from accepting the vaccine in areas where it was available?”.

In this study, in an effort to assess the influence of the COVID-19 Infodemic in vaccine hesitancy, the current paper explores different infodemic themes on social media, which performed as barriers to vaccine acceptancec. In addition, this study aims to highlight how an infodemic disrupts governmental/ international efforts in conflict settings, and therefore, there is a critical need for establishing action plan for an infodemic management in response to future outbreaks, significantly in conflict settings.

2 Methodology

2.1 Study design

This study used a mixed-method approach. The qualitative data were collected from social media platforms, including Twitter and Facebook, using the Early AI-supported Response with Social Listening (EARS). EARS is a WHO-developed platform that allows social listening to real-time conversations and misinformation on the vaccine during the pandemic [28]. Our results were then complemented with a quantitative cross-sectional community survey with 400 participants through cluster sampling using the SMART household survey methodology adopted by UNICEF in Aden city.

2.2 Data collection

2.2.1 Online social listening EARS

Using the WHO Early AI-supported Response with Social Listening (EARS), EARS is a tool that has an access to public social media content on Twitter, Facebook and Instagram in English, French,, Spanish, Portuguese and Arabic languages to provide health decisions makers and researchers with real-time analysis of public online conversations in multiple languages [29] The AI-powered tools allow for researchers to adjust filters to choose which language, platform and narratives related to COVID-19 they want to explore and analyse. Using EARS, we accessed to people’s narratives, questions trending in public conversation through customizing EARS’ features choosing Yemen from the country list that involve 40 country and adjusted the time line to the selected period of the study and selected Twitter and Facebook one in each time [29, 30].

We collected 1869 total (1.774 tweets) from Twitter and (95 posts) from Facebook mentioning COVID-19 vaccines on from all over Yemen. The data gathered from January to April 2022, when the search for the COVID-19 vaccine was at its highest, according to Google Trends. The tweets and posts retrieved using EARS were in Arabic and then underwent translation to English by Arabic-English native researchers. EARS shows the trending keywords that have been repeated in public conversations in Yemen, and among the Arabic words was " تطعيم كوفيد"، "تطعيم" "كوفيد-١٩" ،" تطعيم فيرس كورونا" ،

The keywords included “COVID vaccine*” OR “COVID-19 vaccine*” OR “COVID19 vaccine*” OR “coronavirus vaccine*” OR “vaccine* for coronavirus” OR “vaccine* for COVID.” Each tweet/post record comprised of the account name, contents, and post time. Since our study aimed to assess public attitudes toward COVID-19 vaccination, we only included tweets or Facebook posts from individual accounts and excluded those from news and organizational accounts. Therefore, duplicate tweets, tweets with identical text but different tweet identifications, retweets, quotes without comments, and irrelevant tweets were removed.

2.2.2 The survey

On the other hand, the survey was conducted in the city of Aden, where we randomly sampled 400 participants for a face-to-face cross-sectional survey through cluster sampling using the SMART household survey methodology adapted by UNICEF in this city. The sample size was set to achieve a confidence interval (CI) of 95%, precision of 10%, and 0.5 proportion and was calculated to be 384 potential participants; thus, we aimed at a sample size of 400 participants. Participants were of all ages and genders and from different regions of Aden (AlBreiqa, Crater, Mansoura) a mean age of 41 (SD 15) years and a range of 18–86 years. We executed data quality control by training interviewers, pre-testing the tools, scanning the questionnaire, assembling the data on a tablet, and assigning a supervisor to monitor data collection in real-time day-to-day. COVID precaution measures have been considered to protect both interviewers and participants. In addition, participants were informed about the study's purpose, risks, benefits, the right to voluntary participation, the right to withdraw, privacy policy, and policies for sharing data. Questionnaires took approximately 10–15 min to be administered to each participant.

2.3 Data analysis

2.3.1 Online social listening EARS

After the EARS platform uses quantitative analysis to recognize the most frequent themes reflecting potential information voids and areas of concern/questions in Arabic and English; This platform relies on the volume of online items that mentioned COVID-19 vaccine together with a keyword related to each tracked topic [28].

These tweets and posts were then investigated in more detail via qualitative analysis to comprehend the context and identify when and why the public was hesitant to take the COVID-19 vaccine. Our investigation prioritized the widespread perplexity or repeatedly asked questions, the dynamic of infodemics that were particularly relevant to vaccine hesitancy during our study period.

The 1.855 collected tweets and the 14 posts were first reviewed and translated into English. Then, four data analysis steps were followed to gain insight into the social media users' concerns and questions about the COVID-19 vaccine:

  • Interpreting the data by reading each tweet and post and underlining statements.

  • All underlined statements were coded.

  • All codes were grouped into six themes.

All statements in all themes were read to reflect the overarching participant's concerns and questions about the COVID-19 vaccine.

The translation step involved reviewing the data before translation to exclude content that did not meet the research criteria. Subsequently, the translation was conducted using Google Translate and underwent further review by native Arabic and English speakers to ensure the preservation of the essence of the Arabic content.

The qualitative analysis phase involved two coders who independently reviewed the gathered and separately translated data. Each researcher then developed their main codes and themes, which were reviewed by a third researcher to ensure coding reliability.

The analysis of content extracted from Facebook and Twitter was conducted separately. Content from Facebook included only a few posts from public pages and accounts due to privacy and ethical considerations. Similar content from Facebook and Twitter was merged into similar codes and themes.

2.3.2 The survey

This survey has been a part of big study conducted in Yemen. The survey contained 33 questions. The questionnaire collected information on (1) socio-demographic characteristics; (2) knowledge of COVID-19 transmission and prevention measures; (3) attitudes toward COVID-19 prevention measures; and (4) practices regarding COVID-19 prevention. Participants answered most knowledge questions with “Yes/No,” or “I do not know.” Other knowledge questions were open-ended which included asking participants to list forms of COVID-19 transmission and prevention methods they utilized. For both types of questions (closed and open-ended) those who gave an incorrect or an incomplete answer was scored “0,” and the correct answer was scored “1.” Similarly, in the section on attitudes and practices, a score of “1” was given to answers that reflected positive attitudes and/or good prevention practices toward COVID-19 and a score of “0” was given to answers that reflected negative attitudes and/or poor COVID-19 prevention practices. Practices were deemed good/poor through comparison with guidelines and literature on COVID-19 prevention practices. Two researchers, both with graduate-level quantitative training, agreed on categorizing answers as good/poor via consensus. Thereafter, an average score was calculated for each section and compared with a predetermined scale: 0–40% = poor knowledge/negative attitudes/ poor practices; 41–70% = intermediate knowledge/fair attitudes/fair practices and 71–100% = good knowledge/positive attitudes/good practices. The items 11, 12, and 30–33 have been utilized in this study as they were specific questions related to the vaccine-infodemic, assessing knowledge, attitudes, and practices towards the COVID-19 vaccine. A descriptive statistic was presented according to the impact of misinformation on the attitudes towards COVID-19 and preventive measures. Descriptive analysis using mean values and standard deviations for continuous variables and the count and percentages for the dichotomous or categorical variables were used in describing the data using Stata® 13 software.

3 Results

3.1 Qualitative themes from social media

Investigating the Facebook and Twitter content using the AI-EARS dashboard resulted in a total of 1869 document (1.855 tweets) from Twitter and (14 posts) from Facebook that mentioned the COVID-19 vaccine in the defined period from January to April 2022. There are 41 categories of COVID-19-related narratives, including vaccine distribution, misinformation, disinformation, travel measures, and other vaccine-related categories.

Table 1 showcases six COVID-19 vaccine themes, five of which supported vaccine hesitancy in Yemen, which included (1) Concerns over the vaccine's side effects, (2), Concerns about the vaccine’s safety and effectiveness, (3) Conspiracy theories, (4) mistrust in the health authorities and (5) underestimation of the COVID-19 health risks. However, the sixth theme, travel, highlights an additional identified theme that may have motivated COVID-19 vaccine uptake.

Table 1 COVID-19 vaccine themes and codes from social media
Full size table
  1. 1.

    Concerns about the severity of the vaccine's side effects.

    Twenty-nine tweets in the study time frame contained the term "vaccine", including apprehensions and fears surrounding the potential vaccine's side effects. Further, questions about the types of adverse effects and their severity were echoed in many tweets. Similarly, several queries were posed about ways to protect oneself from the post-vaccine side effects. Twitter users expressed particular concerns about receiving the COVID-19 vaccine and the booster dose if they have active cold symptoms. Some users emphasized in their Tweets that having mild side effects like shivering after taking the vaccines is a sign that the vaccine works, while others stated that they experienced mild effects.

    Example 1

    "How to prepare for the day of the corona vaccination and the side effects?" Jan 7, 2022.

    Example 3

    "The symptoms after taking the vaccine were a slight fever with joint pain. I administrated paracetamol and put compresses over the vaccine injection side, and thank God I am better now".

    Example 4

    "Is it okay to receive the booster dose while I still have the cold and cough symptoms to go away?". Jan 6, 2022, male.

  2. 2.

    The vaccine’s safety and effectiveness.

    1. a.

      Concerns about severe complications, including strokes and abortion for pregnant women.

      Vaccine safety and effectiveness were detected in (17 posts) from Facebook and (63 tweets) from Twitter. Multiple concerns about vaccine safety were evident, particularly fear surrounding severe complications, including, strokes and abortion for pregnant women. Other concerns included worries about the vaccine's safety for pregnant women and children in the long term.

      Example 1

      "Pregnant women who received the Pfizer vaccine for Covid-19 during their pregnancy faced abortion and lost their pregnancy". Jan 27, 2022.

      Example 2

      "Is the vaccination against corona the cause of the spread of hepatitis among children?". Apr 30, 2022.

    2. b.

      Second and booster doses raised doubts about the vaccine's effectiveness.

      Additionally, people raised questions about the effectiveness of the COVID-19 vaccine in protecting against COVID-19 infection, especially after releasing recommendations to receive two doses and a booster to reach a high level of COVID-19 immunity.

      Example 1

      "There is no effective vaccine against Corona, and it did not stop the spread of the disease, nor did it provide protection for vaccinated people from the infection. Are we convinced that a fourth dose will benefit?" Mar 25, 2022.

  3. 3.

    Conspiracy theories over the vaccine development.

    1. a.

      The vaccine changes the DNA and causes sudden deaths.

      Findings from 28 posts on Facebook and 83 tweets from some residents reflected some of the conspiracy theories circulating in Yemen about the intention behind vaccine development. For example, some described the aim of the vaccine as altering the human DNA and causing sudden death.

      Example 1

      "Vaccine is an international conspiracy to kill people and to reduce the population number." Jan 4, 2022.

      Example 2

      "I hope the government examines all coming vaccines to assure that it doesn't pose any threats to our children in the long-term causing autism and other diseases, especially if it comes from countries that hate Muslims". Jan 1, 2022.

      Example 3

      "The corona vaccine contains lipid nanoparticles that change the shape of the protein of the body's cells to an abnormal shape called "prion." Jan 18, 2022.

    2. b.

      COVID-19 vaccine is a western ruse to destroy muslims.

      Other claims stated that the vaccine is an obvious tactic from the Western countries to destroy the Arab nations and Muslims and that pharma companies did make a significant profit from COVID-19 vaccine deployment. 109 tweets mentioned vaccine opposition by people living in wealthy countries while vaccines were sent to Arab countries as an experiment. Other Tweets included rumors and invalid information about the reported deaths of children after taking the COVID-19 vaccine.

      Example 1

      "Pfizer revenues are about $ 5 trillion from the COVID-19 vaccine, considering about a billion people take it worldwide. The U.S. government takes 50% of this revenue in the form of an income tax through which the U.S. government can cover the public debt in ten years".

      Example 2

      "Children who die after vaccination against COVID-19 are 52 times more likely than children who were not vaccinated with COVID-19 vaccine. A parliamentary hearing revealed that the Pentagon concealed the side effects and deaths of large doses of the COVID vaccine.”

  4. 4.

    Mistrust in the health authorities.

    1. a.

      Criticism of the government’s weak performance during the pandemic.

      Many Tweets contained harsh criticism of the government’s weak performance during the pandemic, which was identified in 66 tweets. Yemenis’ opinions varied between a lack of confidence and significant mistrust in the authorities’ efforts.

      Example 1

      “We don’t have tests, vaccines, or statistics in Yemen (northern part), and everything is normal as before Corona!”. Male, Jan 4, 2022.

      Example 2

      “Only 54% of the country has been vaccinated, and the vaccination rate is low” Apr 5,2022.

      Example 3

      “Corona vaccine is ineffective, and the national television is a killing organization that plans to reduce the population number” Jan 10, 2022.

    2. b.

      The unfair distribution of vaccines in the country.

      Many complaints came from the northern part of the country, such as the Sana’a governorate, regarding the rapid increase in COVID-19 cases and local authorities limiting testing and COVID-19 vaccine rollout. There was a discrepancy in the tweets (96) from social media users in the southern region describing that all measures were taken during the fourth stage of the vaccine distribution. News and tweets mentioned that “Johnson& Johnson and Sinovac” vaccine deployment was limited to the liberated governorates. Seven tweets announced the beginning of the COVID-19 vaccination rolling out in Marib and one about Aden. Further, there was news about releasing a list of the locations for receiving the Johnson and Johnson vaccine in Ta’izz Governorate.

      Example 1

      “The Houthis and the World Health Organization misled people regarding the distribution of the Corona vaccine in the areas of the Houthi authorities” Jan 11, 2022.

      Example 2

      “Marib Health Office launches a fourth campaign for immunization against Corona.” Jan 9, 2022.

      Example 3

      “We need vaccines against corona, especially in the northern regions,” Jan 1, 2022.

  5. 5.

    Low perception of COVID-19 health risks.

    1. a.

      COVID-19 is a mild infection like the common cold or flu.

      Low perception of COVID-19 health risks was vastly discussed on Facebook and Twitter. Yemenis expressed their opinions regarding the severity of the disease on their health in 70 documents (23 posts, 54 tweets). Many quotes mention COVID-19 is a mild infection like the common cold or flu symptoms. Others suggested that the governments should focus on other vital issues, such as traffic, which kill more people than corona.

      Example 1

      “It is an ordinary flu that only comes in the winter; during these days of the seasonal flu outbreak, some people get Corona without symptoms. This topic has been overestimated, and based on statistics, traffic accidents kill more.” Male, Jan 1, 2022.

      Example 2

      “There is nothing called coronavirus; it is just a normal cold except that the psychological effect plays a role in believing such news.” Male, Jan 2, 2022.

  6. b.

    The influx of news that they perceived to over-report COVID-19 infection risks.

    Additional findings entailed some recommendations for take-home remedies and herbal treatments for those who become infected with COVID-19 instead of receiving the COVID-19 vaccine. Many social media users in Yemen justified their rejection of the COVID-19 vaccine due to the influx of news they perceived to over-report COVID-19 infection risks and this was identified in 64 tweets.

    Example 1

    “Best treatment for corona is “Support” means if you feel the symptoms; forget it and avoid stress and anxiety and consider the symptoms like cold; take usual medication and eat garlic, lime, and onion and avoid ice cream” Jan 5, 2023.

  7. 6.

    Travel restrictions: traveling abroad was the only motivator for vaccine acceptance.

    Although the abovementioned themes may have discouraged COVID-19 vaccine uptake, findings showed that traveling abroad was the only motivator for vaccine acceptance. Tweets (230) and 27 posts contained several questions and concerns about travel restrictions and requirements, especially in January 2022. The inquiries asked about countries requesting the vaccine certificate upon arrival and the requirements for traveling to Yemen from elsewhere. Repeated questions were about the travel restrictions to neighbouring countries, including Saudi Arabia and Kuwait. Similarly, Facebook users complained in the posts about the shortage of the Covid-19 vaccine as travelers received all the doses first before the rest of the people as it was compulsory to travel outside.

    Example 1

    “A Yemeni residing in Qatar had two doses of vaccination more than ten months ago, and after the two doses, he got Corona and recovered. Is he allowed to enter with an Umrah visa from Qatar, or does he have to take the third dose?". Feb 27, 2022.

    Example 2

    "Is it required for those who want to enter the Saudi Arabia Kingdom for a visit or Umrah for the Corona vaccine certificate?".

    Example 3

    "The Ministry of Health decides to allow travelers who have the complete course of Corona vaccine to enter the country without a "PCR" examination" male, Feb 15, 2022.

3.2 Survey's results

Regarding the number and gender of participants; (male = 205, female = 195) took part in in the survey from regions of Aden (AlBreiqa, Crater, Mansoura). Whereas more than half of our survey respondents (N = 242, 60.50%) think vaccine misinformation has no negative impact on their health (Table 2), both physical and mental, a third of our sample size succeeded in acknowledging that vaccine misinformation affects their physical and psychological health (Table 2). Unfortunately, the remaining respondents (N = 24, 6%) did not answer this question.

Table 2 People's perception of health misinformation's impact on health
Full size table

Misinformation’s Impact on health decisions: Although more than half of our respondents (N = 220, 55%) did not acknowledge that misinformation has an actual impact on their health (Table 3), more than one-third (N = 152, 38%) admitted that the COVID-19 vaccine's misinformation has affected their health decision (Table 3). However, the remaining respondents (N = 28, 7%) did not give answers in this part.

Table 3 People's perception of health misinformation's impact on health decisions
Full size table

Reasons triggering vaccine hesitancy: Even though more than half of our respondents have previously stated that the COVID-19 vaccine misinformation didn't affect their health nor did it influence their health behaviours (Tables 2 and 3), in Table 4, nearly half of the participants (N = 243, 61%) exhibited hesitancy behaviors toward accepting the COVID-19 vaccine. This hesitancy was attributed to various reasons, such as suspicion regarding vaccine efficacy (N = 30, 7.5%), as well as belief in vaccine conspiracy theories (N = 32, 8%). Additionally, rumors and claims that the vaccine causes further diseases were reported by respondents, with (N = 55, 14%) and (N = 53, 13%), respectively. In contrast, unwillingness to try new vaccines was reported by (N = 48, 12%), while roughly (N = 20, 5%) mentioned other reasons for their hesitancy attitudes towards the COVID-19 vaccination.

Table 4 People's attitudes and perceptions toward COVID-19 vaccine uptake
Full size table

4 Discussion

Looking into our results, we consider information gaps being one of the main drivers of vaccine hesitancy alongside the different types of mis/disinformation. From the beginning of the vaccine rollout in Yemen, information gaps around the vaccine's side effects caused many concerns and doubts, captured in our social listening from Facebook and Twitter. A similar study investigated internet search queries about the COVID vaccine's side effects [31]; this study objectively evidenced that the increased unaddressed concern on social media regarding the vaccine's side effects caused a significant hesitancy in vaccine uptake [31]. Although all vaccines for COVID-19 cause common side effects such as pain and swelling at the injection site, fever, chills, fatigue, joint pain, nausea, muscle soreness, and headache, the findings from various clinical trials of COVID-19 vaccines concluded that the vaccines were well-tolerated and had a favorable safety profile [32]. Nonetheless, feeble efforts were made by any Yemeni stakeholders to address people's questions and concerns on social media regarding the vaccine side effects, which gave a chance for misinformation to generate on social media, some of which were true, but the majority were false claims; a similar study gathered vaccine-related content on social media from 52 countries; from the content compiled in that study, 91% were rumors, and 9% were conspiracy theories [33]. Hence, the absence of risk communication on social media and failure to fill information voids by not addressing public concerns were the initial cause for the well-spread misinformation related to the vaccine’s side effects in Yemen, a similar study in the literature from 2020 has emphasized that one conceivable way to assure proper risk communication is using social media channels, and providing an ongoing, timely and consistent media presence [34], which unfortunately was massively absence in Yemen.

With the introduction of the second dose and the booster dose globally, the majority in Yemen have not yet received their first dose, and therefore they started suspecting the effectiveness of the vaccine against new variants and if it's still worth it. Furthermore, a recent study stated that the primary immunization with two BNT162b2 or ChAdOx1 doses delivered no or narrow protection against symptomatic disease with the Omicron variant [35], which supports the peoples' claims and concerns. However, the previous study stated that boosting with BNT162b2 following either primary course extremely increased protection [35], nonetheless, Yemen has only introduced the Oxford AstraZeneca vaccine. Apart from the challenges and the effort to get the vaccine only from the southern governorates. Therefore, some of the peoples' concerns and doubts in Yemen were rooted in reality, but anti-vaxxers used these concerns to their advantage to promote misinformation on social media and boosted conspiracy theories about the vaccine's effectiveness.

Fear of vaccines, especially in LMIC, is not a new issue. A recent study from 2022 stated that this fear is a problem with historical roots that go back to when vaccines came with significant risks of inducing debilitating adverse effects or death, weak understanding of the measures taken by medical researchers to ensure the safety of the COVID-19 vaccines is poorly understood by the public provoking active anti-vaccination groups on social media that generate fear of vaccines [36]. The same issue was present in Yemen, where anti-vaxxers spread misinformation on social media that the COVID vaccine changes the human DNA and causes sudden death. Another recent study in 2020 from five countries UK, Ireland, USA, Spain, and Mexico, has yielded that raised susceptibility to misinformation negatively influences people's self-reported compliance with the COVID-19 preventive measures, as well as people's willingness to get the vaccine or to recommend it to vulnerable family members or friends [37]. Another study conducted in Yemen has reported that misinformation and information gaps (low access to trustworthy sources and answers to the public's questions and concerns) have driven the population not to acknowledge COVID-19 existence and even practice social pressure on those who complied with the WHO guidelines [38].

The issue of mistrust in the Yemeni Government was predominant in our social listening results as many Yemeni social media users have harshly criticized the Government’s management of COVID-19, as the De-facto authorities in the north refused to receive the vaccines because they didn’t acknowledge COVID-19. In addition, a recent study about the impact of the infodemic in Yemen has also reported that the De-facto authorities in the north had denied COVID-19 existence [38]. Likewise, the same study indicated that no cases were reported in the north, and only the southern governorates, where the vaccine is available, reported only a few COVID-19 cases and deaths due to low testing capacity [38]. Although the people's mistrust of the Yemeni government during COVID-19 is highly justified, we still think it can be rebuilt if infodemic management approach and robust risk communication techniques are to be used. The WHO guideline for emergency risk communication policy and practice also emphasizes that establishing and maintaining trust is arguably the first and most crucial step in effective emergency risk communication [39].

The population in Yemen showed a very low willingness to take the vaccine due to a low perception of COVID-19 health risks and its vaccine benefits in protection, whereas social media was full of misinformation, considering COVID-19 the normal flu. A similar study from Yemen showed that the community called COVID-19 with different names to make it sounds normal and not a serious disease that requires prevention or vaccination [38]. Low perception of the impact of misinformation was also evidenced in our survey, in which around 60% of our respondents claimed that misinformation doesn’t affect their health and doesn’t affect their health decision to take the vaccine. However, when they were asked why they hadn’t taken the vaccine yet, their answers showed that their fear of the vaccine was due to rumors and conspiracy theories and none of them mentioned real challenges like the accessibility or availability of the vaccine which indicates that the main challenge for the COVID-19 vaccine acceptance in Yemen is misinformation.

5 Conclusion

Our research highlights the infodemics' impact on vaccine acceptance, where unaddressed information voids about the vaccine had evolved into misinformation, forcing people to mistrust all public health actions, including the vaccine. Our research also shows that in conflict settings like Yemen, the absence of Risk Communication and Community Engagement (RCCE) response amid COVID from the health authorities is a primary driver for mistrust in the government during health emergencies, intensifying the impact of infodemic on the population's health decisions and practices.

5.1 Limitation

Although the AI-EARS platform is global and country-focused, it does not include state, regional, or community-level information. As with other social listening dashboards, Facebook content due to limited content there are ethical considerations, especially regarding transparency and privacy. And, because dashboards condense complex data into a simplified format, they risk skewing the accuracy of interpretation. The AI-EARS Platform has greater access to all tweets as opposed to Facebook, where it has only access to public group pages and not to individual accounts. During the research period, we encountered only content from tweets and posts by community members. Few tweets released by health authorities focused solely on publishing the location of vaccines and vaccine centers, with no tweets or posts addressing public concerns and questions on vaccine. AI-EARS tool was in the pilot phase and it did not identify any content from official governmental accounts. Another limitation of our study was the utilization of a binary answer scale (yes or no) for the items participants responded to in the survey, which may have restricted the depth and nuance of their responses. While this approach allowed for straightforward analysis, it limited the ability to capture more nuanced attitudes towards vaccination.