Understanding COVID‐19 misinformation and vaccine hesitancy in context: Findings from a qualitative study involving citizens in Bradford, UK

Abstract Background COVID‐19 vaccines can offer a route out of the pandemic, yet initial research suggests that many are unwilling to be vaccinated. A rise in the spread of misinformation is thought to have played a significant role in vaccine hesitancy. To maximize uptake, it is important to understand why misinformation has been able to take hold at this time and why it may pose a more significant problem within certain contexts. Objective To understand people's COVID‐19 beliefs, their interactions with (mis)information during COVID‐19 and attitudes towards a COVID‐19 vaccine. Design and Participants Bradford, UK, was chosen as the study site to provide evidence to local decision makers. In‐depth phone interviews were carried out with 20 people from different ethnic groups and areas of Bradford during Autumn 2020. Reflexive thematic analysis was conducted. Results Participants discussed a wide range of COVID‐19 misinformation they had encountered, resulting in confusion, distress and mistrust. Vaccine hesitancy could be attributed to three prominent factors: safety concerns, negative stories and personal knowledge. The more confused, distressed and mistrusting participants felt about their social worlds during the pandemic, the less positive they were about a vaccine. Conclusions COVID‐19 vaccine hesitancy needs to be understood in the context of the relationship between the spread of misinformation and associated emotional reactions. Vaccine programmes should provide a focused, localized and empathetic response to counter misinformation. Patient or Public Contribution A rapid community and stakeholder engagement process was undertaken to identify COVID‐19 priority topics important to Bradford citizens and decision makers.


| INTRODUC TI ON
Tackling the rise of misinformation, which we interpret as false or inaccurate information communicated irrespective of intent to deceive, is a central challenge of our age. 1 The abundance of misinformation, facilitated by social media, has the potential for severe adverse consequences, as people become not only misinformed but less able to believe in scientific facts and trust experts. 2,3 The rise and spread of misinformation is associated with periods of political and economic upheaval, and in the COVID-19 pandemic it poses a major threat to public health. 4 Trust in government, scientists and health professionals is seen as essential in preventing the spread of COVID-19 and implementing a successful vaccine programme. [5][6][7] Yet the spread of COVID-19 misinformation has contributed to what has been labelled as a 'crisis of trust'. 8 This decline in trust has been reinforced by legitimate criticism of government responses to the pandemic and the exacerbation of preexisting mistrust in governments and health services, particularly amongst marginalized groups. [9][10][11] There is evidence that fear and anxiety of catching and dying of COVID-19 is extensive, particularly by women, younger people and individuals who identified as being in recognized risk groups. 12 As well as health anxiety, research has indicated that people are experiencing greater financial anxiety, loneliness and mental health issues as a result of government measures to prevent the spread of COVID-19. [13][14][15] The constant news cycle around COVID-19 and the spread of misinformation is also reported to have exacerbated fear, anxiety and stress. 16,17 Within this climate, the UK government is attempting to roll out a mass vaccination programme.
On 2 December 2020, the Pfizer/BioNTech COVID-19 vaccine was approved for use in the UK, with the first vaccine administered six days later. Initially focused on people over 80, people who live or work in care homes and health-care workers at high risk, the vaccination programme is expected to be extended much more widely in 2021. 18 The results from multiple surveys in the UK have found that between 54% and 64% of respondents would definitely or are very likely to accept a COVID-19 vaccine. Between 4% and 9% reported that they would definitely not or were unlikely to accept it, suggesting many people are unsure. [19][20][21] There are already indications that certain population groups are more hesitant to receive a vaccine. A poll by the Royal Society of Public Health found that respondents from Black, Asian and minority ethnic (BAME) backgrounds were less likely to accept a COVID-19 vaccine (57%) compared with white respondents (79%). 22 Vaccine hesitancy refers to delay in acceptance or refusal of vaccination despite its availability. This is not a new phenomenon and has existed since vaccines were first introduced, and the reasons behind it are multifaceted and complex, with geographical variations. 6,23 In the last two decades, growth of vaccine hesitancy, particularly amongst new parents, has been assisted by a discredited study linking the MMR vaccine to autism and by the spread of misinformation and personal stories of alleged vaccine injury through the internet and social media. 24 In the context of COVID-19, there is emerging evidence that people's belief in misinformation about the virus and especially their views about the origin of COVID-19 (i.e. that it was manufactured) will make them less likely to accept a vaccine when it becomes widely available. 25,26 Large-scale surveys are helpful in identifying the general population's intentions towards a COVID-19 vaccine and some barriers to uptake, including exposure to misinformation. 4,27 Detailed qualitative work, however, enables us to further explore the interaction between misinformation and people's experiences of and reactions to the pandemic, building understanding as to why vaccine hesitancy varies across populations and places. The aim of our study was to explore people's COVID-19 beliefs, their interactions with health (mis) information during COVID-19 and attitudes towards a COVID-19 vaccine.

| Study design
This descriptive, inductive qualitative study was completed as part of a larger mixed-method, longitudinal research study to provide actionable intelligence to local decision makers, developed in response to community and stakeholder consultation processes described in Box 1. 28 We used in-depth interviews to explore citizen's health experiences and beliefs during COVID-19. University ethical approval was secured in July 2020.

| Study setting
Our study was conducted in Bradford, a city in the North of England. Bradford and its surrounding district is the fifth largest metropolitan district in England and is an area of high deprivation and ethnic diversity. Since March 2020, Bradford has experienced a relatively high number of COVID-19 cases compared with the rest of the UK, and stricter lockdown measures from July 2020 which remained in place until the introduction of the tier system in October 2020. 29 In the second wave of the pandemic, Bradford hospitals experienced patient numbers similar to the peak of the first wave in April and May. High rates of COVID-19 in areas like Bradford are likely to be due to greater deprivation, high population density and a higher than average number of multigenerational households. 30

K E Y W O R D S
Bradford, COVID-19, misinformation, qualitative, vaccine hesitancy

| Sampling and data collection
We conducted in-depth interviews with 20 people in different communities and different areas of Bradford via a maximum variation sample, with our key sampling focus being diversity of ethnicity and age. Nine 'community influencers' were contacted (three people from each major ethnic group in Bradford -South Asian, White British and Eastern European) and invited to take part in an interview or identify others who would. This method was favoured because community influencers, people embedded in community settings through their paid or voluntary work (eg advice worker, school and nursery community liaison, community councillor), are more likely to be trusted by their peers and people with whom they engage. The initial community influencers were identified by the second author, a community-based researcher with significant local knowledge.
Snowball sampling was used to recruit further participants. When 15 interviews had been completed, demographic and geographical gaps were identified, and additional participants were recruited via contact with volunteers at a community organization. Conventional recruitment methods such as adverts on social media may not have attracted the diversity of respondents we were seeking.
Fieldwork took place between September and October 2020.
Eleven women and nine men participated, ranging from 20 to 85 years old, but most were aged between 25 and 54 years. In terms of ethnic group, they identified as Asian or Asian British (Pakistani,

In March 2020, Bradford's COVID-19 Scientific Advisory
Group was formed to support policy and decision makers in Bradford and the UK to deliver an effective urgent response and to better understand the wider societal impacts of COVID-19. As part of these aims, a rapid community and stakeholder engagement process was used to identify priority topics important to both citizens in Bradford and local decision makers. This process took place in April 2020 via the following engagement activities:   Indian and Bangladeshi) (10) White British (6), White Other (Eastern European, Gypsy or Irish Traveller) (4) (see Table 1

| Interview questioning
Headline topic guide questioning was derived from the areas that the consultation process (Box 1) identified as important to explore.
The format of the topic guide and interview questioning was flexible to allow participants to voice what they considered to be important.
The initial topic guide contained open-ended questions and was piloted and iterated several times particularly in relation to data generated in early interviews regarding people's beliefs about a COVID-19 vaccine. We included all data in the analysis and did not discard meaningful data gathered during piloting.

| Analysis
We undertook the analysis using the principles of reflexive thematic analysis. 31 Five interview transcripts, chosen based on their representativeness of the whole data set, were analysed independently by the first and last authors. We held an analysis session to identify commonalities and differences in the interview narratives and worked towards ordering the data into loose themes. The resultant coding framework was then refined further by the first author and applied to the rest of the data set. The first author then coded all interviews and conducted further interpretive work to write up the findings, sense checking with the last author as necessary. The analysis conducted was manual without the use of a software package.
The analysis was wholly inductive, and, as such, we did not structure it on any existing theoretical frameworks.

| Confusion, distress and mistrust
The avalanche of information surrounding COVID-19 had left many of the people we interviewed feeling overwhelmed and confused. (Jackie).
Participants felt the government response had been particularly bewildering. No one we spoke to seemed to be staunchly antigovernment, and they expressed a desire to follow the rules and restrictions, but they also felt that the national government communication had been poor and their decisions were contradictory or hypocritical: The government aren't being clear and they're saying one thing but then they're saying other things, and basically what they're trying to do, they're trying to please everybody all of the time, it doesn't happen. (Hasan).
There was also mistrust for some traditional news outlets that people felt were a mouthpiece of the government. This state of con- (Ambreen) Participants underlined how quickly social media stories were shared, with Tariq stating 'they just forward it straight away and then it just spreads like wildfire'. The more controversial or dramatic the posts or videos were, the more they spread. Individuals in these videos were (or were posing as) trusted professionals, such as a teacher, nurse or doctor. Being able to deliver a video in multiple languages indicated higher levels of education and trustworthiness.
They're just so passionate the way they talk, they grab your attention and they've got you and the way they're speaking and the terminology they're using and they give you the facts and the figures and then you just get drawn and locked into it.

(Tariq)
When participants talked about their interactions with false news they distanced themselves, referring to them as something they were dismayed at or amused by, scrolled past or ignored.

| Safety concerns
The safety of a potential COVID-19 vaccine was a concern, even for those who were very willing to have it. Some felt reassured by the medical establishment testing process in the UK. Angela commented 'in England we're very good at testing stuff, aren't we?' Louise was less sure: I think I'd have to know that it was a safe… I mean, they wouldn't be doing an unsafe vaccine anyway would they, you know, but I think I'd have to have some confidence that it was a good vaccine and that it was quite safe. (Louise) A major issue for people was how quickly any potential vaccine would have been produced and that the vaccine makers would not know all the side effects as yet. Sofija was worried it had not had time to be fully tested and Tariq wanted to wait three to six months to see what the effects of the vaccine were on others before he would be happy to take it. Some participants were afraid of very severe side effects, and it was clear that these worries had been exacerbated by engagement with social media stories: People are saying they don't know how safe it is, plus they've made it so quick we don't know the sideeffects it's going to have in the future. I mean it's probably safe because they wouldn't be allowed obviously to give it to us otherwise, or maybe they would you know, sometimes they don't care, but you just don't know if it could cause infertility, it could cause cancer in the future. (Rebecca)

| Negative stories and misinformation focusing on the vaccine
Those hesitant about having the vaccine felt confused by the negative stories about it, rather than being resolutely against it. There was one exception, Faiza, who had joined live social media broadcasts where people were revealing the 'truth' about the negative side effects of the vaccines that they said was being hidden from the general public. Other people's engagement with misinformation around the vaccine was more passive. Rebecca did not actively seek out these types of videos but said they often auto-played when she was watching YouTube and she described feeling confused after watching one. Tariq knew that watching these videos was impacting on his feelings about the vaccine and vouched to stop watching them. Alongside videos which claimed that the COVID-19 vaccines were unsafe, there were also rumours that certain communities and ethnic groups were being targeted to test the vaccine, or the vaccine was being used as a way to harm them: I think what the community are saying is that the vaccine is testing people, they're just using people as the guinea pigs… we experience discrimination for many years, and if we've been focused for, if the Slovakian authorities we are focused especially on the Roma, and the focus is they will be testing them because they were noting who could be spreading all this coronavirus, they may think the same thing now why are we going to offer immunisation, because they're going to trial it out on us. (Kristof) It's all about keeping the population, like controlling the population, and that's the only two I've heard and

| D ISCUSS I ON
This study aimed to explore the health beliefs of citizens in Bradford regarding COVID-19, alongside their attitudes towards a vaccine.
We found that participants encountered a range of misinformation, usually through social media sources. This led to confusion, distress and mistrust in participants' everyday lives and beliefs about government institutions and health services, a more general phenomenon. 33 Vaccine hesitancy could be attributed to: safety concerns, negative stories and personal knowledge, all of which had been amplified by recent exposure to misinformation via social media. We found that the more confused, distressed and mistrusting the participants felt during COVID-19, the more likely they were to be hesitant about uptake of the COVID-19 vaccine.
In February 2020, the World Health Organization stated that the COVID-19 outbreak and response has been accompanied by a massive 'infodemic'. 34 The abundance of information about COVID-19 transmitted across a multitude of platforms has continued, making it difficult to discern what is accurate and what is not. 35 We do not yet know exactly how far-reaching and impactful misinformation about COVID-19 has been or will be, but early evidence suggests that susceptibility to misinformation about virus. Our concern is that these risks will not be as well mitigated by the rollout of widespread vaccine programme, as they may be in other parts of the country, due to the impact of misinformation and vaccine hesitancy. If this is the case, we may see a deepening of health inequalities both within Bradford and in comparison to the rest of the UK.

| IMPLI C ATI ON S FOR P OLI C Y AND PR AC TI CE
Local decision makers have the ability to counter misinformation by implementing targeted local responses. To do this successfully and quickly, there has to be systematic monitoring of the circulation of misinformation on social media. Participants' trust in national government, health professionals and NHS organizations appeared to have weakened over the duration of the pandemic but there was evidence that they did trust people within community support roles that they had frequent contact with, for example teachers, nursery workers and advice workers. Effectively harnessing these connections, through trusted community networks and providing information in languages spoken locally, will be central to ensuring the spread of correct information and providing reassurance.
Hesitancy around the COVID-19 vaccine appears to be rooted in anxiety fuelled by misinformation, and there is a need for this to be mediated by clear, honest and responsive information that is sensitively framed and non-judgemental. It would be prudent for health, social and community workers to be provided with an updated summary of locally circulating misinformation with helpful resources to help them counter concerns and provide informed reassurance.

| Limitations
Interviews took place before announcements about efficacious COVID-19 vaccines were made in November 2020. Participants' discussions about a COVID-19 vaccine were therefore hypothetical.
We do not know how these announcements, and the subsequent surge of information and misinformation, will have impacted on acceptability of a COVID-19 vaccine.
The research was conducted in one place with specific population demographics and may, therefore, not be widely generalizable.
However, there are places all over the UK which have multi-ethnic communities, similar levels of deprivation and population density, and have experienced comparable rates of COVID-19 cases and deaths, which we predict will have similar problems with misinformation spread.

| CON CLUS ION
Our study found that there is an intensity of misinformation being spread about COVID-19 in Bradford and this has impacted on participant's lives by evoking confusion, distress and mistrust during the pandemic. Heightened levels of confusion, distress and mistrust are related to a lower proclivity towards COVID-19 vaccine uptake. As is often found with the inverse law of care, the people most likely to be affected by COVID-19 are those who are most hesitant towards the vaccine. Of critical importance to decision makers is the ability to understand misinformation in its local context and countering it in a sensitive and non-judgemental way via trusted local people whose opinion is valued in their community.

ACK N OWLED G EM ENTS
The authors would like to thank Professor Neil Small, Dr Sufyan Finally, we would like to thank all our interview participants who took the time to be part of our study in a difficult year.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.