Participants and recruitment

This study used a qualitative design. Four groups were recruited from Liverpool, UK:

1. i). adult (aged 18+) members of the public who self-identified as being part of an ethnically minoritised group and eligible for a seasonal influenza vaccine (or had an eligible child),
2. ii). community engagement workers,
3. iii). primary healthcare staff,
4. iv). policy professionals (such as service commissioners and policymakers).

Potential participants were identified via various networks, including Primary Care Networks, Cheshire & Merseyside ICB, NIHR Applied Research Collaboration North West Coast and the Clinical Research Network, Liverpool City Council, and community organisations using purposive sampling.

Participants were recruited purposively across the four target groups. Policy professionals, healthcare staff, and community engagement workers were identified through existing collaborative links and face-to-face interactions at Liverpool City Council learning events, and were invited via email, newsletters, or direct conversation. Some members of the public were identified and approached by community organisations acting as gatekeepers, who translated and disseminated study information through in-person conversations, phone calls, and direct social media messaging. These organisations also supported participation by setting up Teams/Zoom links for remote meetings, providing rooms for in-person meetings, and arranging interpreters who spoke local dialects where needed. Study information was also provided in an easy-read format for community organisations. While many participants were directly identified and invited, some were recruited through open calls (e.g., newsletters and indirect social media posts). As noted in the later described limitations, the use of community organisations as gatekeepers to recruit members of the public likely introduced a positive bias towards these organisations.

Data were collected from 19^th January to 29^th May 2024. Ninety-six people took part (see Table 1 for participant characteristics): 55 members of the public, 14 community engagement workers, 20 primary healthcare staff, and 10 policy professionals. Three were both primary healthcare staff and policy professionals. Most were female (see Table 1 for demographics). Sample size was guided by the concept of information power (preferred to ‘data saturation’; Clarke and Braun [35]), which indicated that a larger sample was needed. The study aim was moderately focused, with purposive sampling used to ensure relevance and variation. While no established theory was applied, the approach was informed by existing literature. Interviews were conducted by researchers trained in qualitative methods and familiar with the topic, supporting high-quality dialogue. A descriptive cross-case analysis strategy, rather than case-by-case analysis, further justified a larger sample. Focus groups were conducted across 34 participants (with individuals from the same participant group), and consisted of between two and five people, while 62 individual interviews were conducted. Forty-one people not fluent in English were able to participate due to the support of community organisations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of participants.

https://doi.org/10.1371/journal.pone.0333602.t001

Ethical approval

This study was approved by the Health Research Authority (IRAS number: 334970) and the Liverpool John Moores University Research Ethics Committee (Reference: 23/PSY/081).

Procedure

Participants were given information on the study and if they agreed to take part a suitable time was arranged with either AP or DC, with interpretation as required. Most sessions were conducted by AP, the Principal Investigator (n = 81), who holds a PhD in Psychology and Neuroscience and has 11 years of experience in both qualitative and quantitative research. DC conducted the remaining interviews, and had been trained in qualitative research as part of MSc Stage 1 Health Psychologist training. To ensure consistency and data quality, DC was further prepared by AP through a structured process: DC initially shadowed several interviews, then conducted one interview under AP’s supervision, followed by a debrief and feedback. AP continued to monitor the quality of DC’s recordings throughout the data collection period. The choice of interview or focus-group was context-dependent and related to preferences of participants, interpreters and/or community organisations. For in-person data collection, written consent forms were signed at this meeting, prior to data collection occurring. For online (Zoom or Teams) data collection, a signed written consent form was returned via email, prior to the meeting. At the start of each session, the study overview was described, including confidentiality, the right to withdraw, and giving the opportunity for questions. Following the provision of informed consent, including to either audio- (in-person) or video- (online) record, sessions began with the collection of demographic data and then addressed the topics of interest. All sessions followed semi-structured schedules with open-ended questions, developed for each participant group by drawing on prior literature (see S1 File for topic guides). Participants received a shopping voucher with a value of 25 GBP (NIHR involve rate) as a thank you for their time.

Data analysis

Recordings were transcribed using the MS Teams transcription function or Otter AI and were then anonymised and checked for accuracy by the research team. The data was then analysed inductively in NVivo (R14.23.3), following the six steps of thematic analysis [36]. Data familiarisation, initial coding, and searching for themes was undertaken by AP, with a subsection of the data (n = 14, 14.58%) also processed by DC, to ensure similar interpretations of the data were being reached. Reflexivity was maintained through discussions between AP, who is White British, and DC, who is Indian and a member of an ethnically minoritised community in the UK, ensuring diverse perspectives were brought to the analysis. The initial set of themes were subsequently discussed amongst the research team, and explored further against the data, enabling them to be reviewed and refined until final themes were agreed upon, defined, and named. These final themes were then shared and agreed with the community organisations as a form of member checking, ensuring they reflected the conversations the organisations were witness to and involved in with their communities. Corresponding anonymised data can be made available upon reasonable request through the LJMU open access data repository via this link https://opendata.ljmu.ac.uk/.

1. Beliefs

2. Trust

3. Accessibility

4. Community

5. Healthcare (staff concerns)

Healthcare staff consistently cited lack of time to engage in in-depth conversations with patients about vaccines.

“…obviously more time would be the thing… Someone who might be persuaded, or who wasn’t sure. Then it’s about it’s about time, isn’t it?... the conversation about vaccine is often part of a consultation about something else, which has already taken more time than you have.” (P11, Primary Care)

They also felt time was an issue regarding the flu strategy, which was viewed as being released too late to support an impactful communication strategy.

“Yes, so I would like it to be sooner… So we do, we do have to wait until we get national guidance on eligible populations… So having to wait for that then determines what comms [communications] you are going to do.” (P1, Policy Professional)

Competition and lack of linkage between services was described negatively, as was the ordering process, with impacts on morale and finances.

“Pharmacies I guess… some of the challenges they [GPs] start to order thousands of flu vaccines and then more pharmacies coming on board to deliver, and then rightly or wrongly there’s a competition there, they’re competing for the audience, you know.” (P20, Policy Professional)

“…so you have to order flu like six months in advance. So you’re kind of guessing in springtime how many vaccinations you need for winter” (P2, Policy Professional)

While integration with the community was recognised as important, concerns were raised about resources for practices to do this.

“…it can require umm trusted people to work closely with those communities and understand what the barrier is for them, and what the concerns are for them and try to address that slowly… I think all of those things require, extra effort… and they require quite considerable resource and that resource would have to be provided… otherwise practices don’t have the resource to do that.” (P11, Primary Care)

Some staff described feeling unprepared to tailor information or their approach according to a patient’s culture, risk, or perception. Staff found responding to misinformation difficult.

“I don’t think I fully understand why some can and some can’t… You know why, why you need porcine free and why this other fella’s just told me that he’s a Muslim and he can have porcine… but you’re telling me you can’t? … I don’t fully understand the reasoning. If I did, I might be able to talk about it more with them” (P23, Primary care)

“I’ve phone call from a parent really abusive, down the phone asking me, how do I sleep at night?... I can’t remember now what she was saying something to do with the brain and swelling and I tried to see whether it’s from a reputable source, what she’d reasoned, but there’s just no, once they… get into that loop.” (P18, Policy Professional)

6. Alliance

Summary

Influenza outcomes are poorer in ethnically minoritised communities, partly due to low vaccine uptake [6]. This study explored influenza vaccine perceptions among community members, community engagement workers, primary healthcare staff, and policy professionals. Six inter-related themes emerged. Beliefs about vaccine safety, necessity, and efficacy; Trust in vaccine information and the healthcare system; Accessibility of accurate vaccine information; Community opinions/experience; Healthcare; and Alliance.

Comparison with existing literature

Several of these findings complement prior work, particularly around misinformation or lack of information, influenza being perceived as low risk, or of people believing alternative medicines or religious behaviours will adequately protect against it, low trust in the healthcare system, concerns about vaccine safety/efficacy, discrimination, and culturally-specific issues (e.g., halal status) [12,18,20]. Where the current study differs, is that while accessibility has been previously highlighted, this has primarily been around physical access such as transport and cost; Nagata, Hernández-Ramos [19], Bhanu, Gopal [21], whereas it was accessibility of information that was most emphasised here. This was found to impact vaccine beliefs and reduce trust due to feelings of low prioritisation. Similar language/literacy barriers were mentioned in the review by Nagata, Hernández-Ramos [19], however these were not major factors in the studies that mentioned them [37–39] with the exception of Schensul, Radda [40], who described a bilingual intervention. In the UK, influenza vaccines are free for eligible individuals, possibly explaining why direct costs are more emphasised in work conducted outside the UK. However, indirect costs such as time off work were mentioned in the current study as physical barriers for some, and have previously been understood as contributing to patients’ perceived cost-benefits of vaccination [41]. Those who described these indirect costs, also highlighted the importance of access to flexible vaccine appointments, particularly for people working variable hours, multiple jobs, and who may experience job insecurity, all of which are more likely in ethnically minoritised communities [42].

It is notable that Trust was such a key issue in this research as Liverpool has a complex history with regards to ethnicity and diversity, which has affected the way in which people in Liverpool interact with “authority” figures. Liverpool gained much of its wealth through being the largest European transatlantic slave-trading port in the 18^th Century [43], and communities in Liverpool have been subject to significant historic and current racism, such as the violent race riots in 1919, primarily against Black residents [44], and the more recent Southport riots in 2024, against Muslims, asylum seekers, and other ethnically minoritised groups [45]. Liverpool is also home to possibly the oldest Black and Chinese communities in Europe, with the former originating partially from African sailors and freed people who had been enslaved [46], and the latter from seamen arriving in the 19^th Century [47]. Liverpool also has a substantial Irish population due to migration from Ireland during the Great Hunger of the 1840s [48], during which successful crops were exported from Ireland to England while many Irish people were starving [49]. Additionally, a considerable number of asylum seekers and refugees reside in Liverpool [50]. This indicates a highly unique combination of cultures, values, experiences, and needs, across the individuals who make up the fabric of the city, and historical events which have eroded trust between communities and government.

Indeed, regarding the previous finding that barriers can be specific to communities [22], while the current study identified overlapping barriers across marginalised groups in Liverpool, certain aspects were more prevalent in certain groups. Roma and Romanian individuals were particularly concerned about vaccine safety and trustworthiness of unfamiliar sources, were vulnerable to misinformation due to low literacy and chosen information sources (social media and family members), preferred video/audio information in their languages, perceived religious behaviours as protective, valued free will (feeling distrustful of being told ‘what to do’), and feared needles. The Chinese and Hong Kong communities preferred written information in Chinese, perceived alternative medicine as protective, and were wary of phone calls and text messages from unrecognised sources due to concerns about scams. People who had recently arrived in the UK/Liverpool (e.g., asylum seekers) were less sure about how to access the vaccine and feared needles as some had no prior experience of injections. Muslim communities were concerned about halal status of vaccines, and vaccine acceptability in their community related to this. Finally, Polish people were more likely to view influenza vaccination as low priority or inaccessible due to working long hours or multiple jobs. Some also travelled out of Liverpool to access services in Manchester (which has a Polish speaking healthcare centre) or Poland, where the private healthcare system was described as efficient, responsive, and with long appointments with familiar clinicians – which, alongside poor NHS recording of Polish ethnicity (historically “White Other”) indicates their UK vaccine uptake data may be inaccurate. Importantly, the groups described here participated in larger numbers, revealing recurring patterns, but patterns for other groups likely remain unidentified. Interestingly, there were no identified patterns relating to age-groups, which is somewhat surprising due to the vaccine’s eligibility status being related to this, and prior research indicating that this is important [24,25].

The findings of this study are consistent with the broader literature on vaccine hesitancy, particularly regarding COVID-19 and MMR vaccines. Common themes such as mistrust in healthcare systems, concerns about vaccine safety and necessity, and the influence of misinformation are prominent with influenza, MMR and COVID-19 vaccine hesitancy. Similar to our participants’ uncertainty about influenza vaccine safety and halal status, concerns about COVID-19 vaccine safety, speed of development, and religious acceptability have been well documented [20,22]. Likewise, the MMR vaccine controversy, which is rooted in the discredited claims of an association with autism, shows parallels in how misinformation can persist within communities, particularly when amplified by trusted social networks [51].

Where our findings diverge from COVID-19 and MMR vaccine hesitancy is the centrality of “accessibility of accurate information” rather than physical access. While access issues during the COVID-19 pandemic often focused on vaccine availability, rollout logistics, and structural inequalities [52], in the present study, language and literacy barriers were more common. This distinction highlights how, for influenza vaccination, especially in ethnically minoritised communities, communication gaps may outweigh logistical barriers. Another contrast lies in perceived disease severity. COVID-19 was widely viewed as a high-risk illness, which motivated some to accept vaccination despite hesitancy [18]. In contrast, influenza was often perceived as low risk, echoing findings from MMR research where the disease itself was seen as relatively harmless compared to perceived risks of the vaccine (Brown et al., 2012). This perception of low threat reduces motivation to vaccinate and amplifies the influence of misinformation.

Strengths and limitations

This study was conducted across a large and diverse sample, including often under-represented groups in research such as Romanian and Roma communities. As one of our participants felt, it can certainly appear that some ethnically minoritised groups have “been marked as the hard to reach, but nobody really wanted to reach them” (P10, Community Worker). This emphasises the importance of working with established community organisations to support recruitment and data collection of marginalised groups, and of not excluding individuals who do not speak English, particularly from conversations about access to healthcare. Additionally, a variety of stakeholders took part in this research, increasing the utility of the findings. Due to the importance of understanding this topic through the lens of the ethnically minoritised groups, the analysis prioritised the viewpoints of members of the public and the community workers, to ensure that the analysis most reflected their experiences. Views of healthcare and policy professionals were used to compare against and complement the views of communities. Finally, while data collection, coding, and analysis was primarily conducted by AP, 14.58% of this was conducted by DC, who reached a similar interpretation of the data before it was integrated into the primary analysis, thus reducing the impact of any preconceptions by AP.

A limitation of this study is that while the focus was to understand a localised population, this does mean that the results may not reflect other geographical areas. However, as the results largely complement prior work, this suggests that the findings may generalise, though specific details will be more applicable to certain communities. Additionally, while many communities did participate, it was not possible to recruit a member of every ethnically minoritised group in Liverpool, or even to speak to multiple members of each group who did take part to establish group-specific patterns. An example of a key group not reflected is the Liverpool Irish Traveller community, who could not be recruited due to time and capacity of the supporting organisations. When recruiting from one community organisation, we extended our inclusion criteria to include people who were eligible for influenza vaccine or those who had an eligible child. This decision was made because during our engagement with the organisation, they felt that it would reduce trust if we only targeted people who were eligible for influenza vaccine. The groups in this organisation were predominantly women, so the decision was made to extend the criteria to include those women who would be consenting for flu vaccination for their children (N = 8 who were parents, of which N = 4 were also eligible themselves). While this could potentially affect the results as the reasons for consenting to a childhood vaccination may differ to those for an adult vaccination, our interview questions did not focus on childhood vaccination programmes and were centred around the adult seasonal flu vaccination programme, so we do not believe that this will affect the results. Moreover, as noted above, there are some similarities between vaccine hesitance in vaccination for MMR (given predominantly to children in the UK) and influenza, so inclusion of a small number of people with eligible children would not be likely to affect the robustness of our results. Finally, it should be noted that community members were primarily recruited through community organisations, and it is unsurprising that active users of these services trust and value them. A challenge therefore is to consider how to understand perceptions of marginalised individuals who are not involved with such organisations, who may be difficult to access and to plan intervention for.

Implications for research and/or practice

This study emphasises the need for stakeholders to collaborate in improving access to reliable information through trusted sources (such as local community organisations), while ensuring flexible appointments are easily available and effectively communicated. Communication strategies should be tailored to each community, using the most appropriate language and format for that group, and addressing relevant concerns and misinformation in a culturally appropriate manner. Culturally informed training for healthcare staff would enable them to support this. Other factors contributing to mistrust (e.g., racism and medical scandals) should be effectively and transparently addressed, at both an institutional and interpersonal level. Similarly, stakeholders should foster trust and improve influenza communication by collaborating with and supporting community organisations, particularly regarding their long-term stability. Such support could include funding, up to date/accurate information and support to share this in various formats, training for community workers in how to disseminate influenza information, medical language training for local interpreters, and stability of funded roles. Finally, stakeholders should seek to further encourage ongoing relationships between healthcare and communities, including by improving access to reliable translators, hiring community members, and increasing the consistency of access to familiar healthcare staff. Future research should seek to develop and pilot localised intervention based on these findings. Table 2 may serve as a practical guide to the settings in which specific barriers occur, thereby supporting the design of tailored intervention strategies. Researchers should also consider how to access and support marginalised individuals not affiliated with a supporting organisation.