The impact of social media interventions on adult influenza vaccination: A systematic review

Social media as a source of vaccination information and a communication tool to increase adherence to vaccination came into prominence with the introduction of the COVID-19 vaccine in 2020 (Benis et al, 2021; Baines et al, 2021). Around this time, there was emerging evidence that social media influencers could increase positive beliefs about influenza (‘flu) vaccines (Bonnevie et al, 2021), building on previous work that social media could improve vaccine acceptance beyond traditional campaigns. (Limaye et al, 2021; Moorhead at al 2013; Connolly and Reb, 2012; Frew et al, 2012). However, social media has purposefully misinformed and promoted anti-vaccination messages (Wawrzuta et al, 2021) which can be difficult to reverse (Steffens et al, 2019). Pro-vaccine messaging also attracts negative comments from highly organized vaccine opponents (Bonnevie et al, 2019).

Whilst informing individuals about the benefits of vaccination may not mean increased vaccine uptake, there is evidence that social media interventions are effective in improving other health behaviours in ‘at risk’ populations such as youth, older adults, low socioeconomic status and rural (Welch et al, 2016). People with few social networks tend to use social media more, thereby enabling social media interventions to reach people that may be difficult to engage in more traditional health promotion approaches, including community empowerment (Lai et al, 2014; Chang et al, 2013). As a result, there is a growing interest in using social media to improve ‘flu vaccine uptake (Benis et al, 2021). However, it has not yet been established if social media based promotional programmes can work to improve ‘flu vaccine acceptance. The aim of this systematic review is to look at the evidence for using social media to improve knowledge about ‘flu vaccine, social media's impact on intention to vaccinate amongst social media recipients and to assess whether there is any evidence that this could lead to an increase in ‘flu vaccination uptake.

Methods

This systematic review followed PRISMA guidelines (Prisma, 2020). Our inclusion criteria comprised of peer reviewed systematic reviews, intervention studies (randomised controlled trials, non-randomised controlled trials) and controlled observational studies about social media and adult ‘flu vaccinations. We used the comparator of usual care (i.e. any type of health promotion activity), no intervention or another intervention which may have a social media component.

The population was an adult population (aged 18 years and older) who were recipients of the social media intervention. We defined a social media intervention as a two-way communication between peers or between the website and users, both creating and reacting to Internet content (Obar and Wildman 2015; Capurro et al, 2014). Examples include Facebook, Twitter, Instagram or Tik Tok. We excluded mass media and any unidirectional forms of health communication. We also excluded any media platforms that were purposively built by researchers as we wanted to examine the effects of using existing social networks with commercial platforms. The social media intervention could be user generated content for adults in all settings with the primary outcome of changing intention to vaccinate or knowledge levels as defined by a comparable scale, and secondary outcomes of affecting vaccination appointment uptake and uptake and coverage levels.

The literature search was conducted in June and July 2021 using the following 8 electronic databases: Medline, AMED, CINAHL, Cochrane Library, EMBASE, HMIC database, PsycINFO and British Nursing Index. Internet searches, ‘snow balling’ and hand searches of journals were also employed to locate any unpublished study or other ‘grey literature’. There was an English restriction on the papers to be included in the review due to the language abilities of the researchers. No time restriction was placed on the search as this was expected to be self-limiting due to the search terms used.

The searches were performed using the following defined search terms: ‘social media’ OR ‘social network’ OR ‘social networking’ OR ‘Web 2.0’ OR ‘Facebook’ OR ‘Twitter’ OR ‘Instagram’ OR ‘Youtube’ Or ‘Snapchat’ Or ‘Twitter’ OR ‘Reddit’ AND ‘Adult’ AND ‘vaccination’ OR ‘vaccine’ OR ‘immunisation’ OR ‘immunization’ OR ‘vaccinator’. The initial search yielded 2381 items.

Two researchers independently reviewed and evaluated the studies using Covidence (www.covidence.org) and reached a consensus on the inclusion for the analysis. Any discrepancies were discussed with reference to the research objective until consensus was reached. Methodological quality was assessed using a system suggested by the Cochrane collaborative (Higgins and Green 2011). Key data that fulfilled the criteria of the study are presented in Table 1.

Results

Titles and abstracts of 1471 papers were screened for relevance and duplication (see Figure 1 for inclusion flow diagram). Thirty-six studies that fulfilled the inclusion criteria were retrieved for more detailed evaluation.

A total of 8 studies were extracted for data synthesis. Six of these studies were USA based, (where flu vaccine is recommended for everyone over the age of 6 months) one in China and one in Australia. Participants in 7 of the studies were university based and 1 focused on pregnant women.

Due to the small number of studies and heterogeneity in design, we could not combine results to generate a meta-analysis. Instead, the results are reported qualitatively as a narrative synthesis. Reporting qualitatively allows for an exploration of differences and similarities of results found in different settings and avoids the possible biases associated with calculation of an overall measure of effect. Furthermore, a narrative synthesis can be more conducive as it allows for the identification of different initiatives and their relative effectiveness rather than a production of an overall effect of whether social media interventions work or not.

The main themes that emerged:

• social media influencers increase knowledge about adult ‘flu vaccinations;
• comments to posts reinforce existing attitudes;
• verified sources and multi-tasking increase intention to vaccinate
• social media only slightly improves vaccine uptake.

Social media influencers increase knowledge about adult ‘flu vaccinations

A controlled study by Bonnevie et al (2020) measured the impact of social media influencers in affecting the knowledge and positive attitudes towards the flu vaccine. The study recruited 117 influencers who generated 69,495 engagements with unbranded flu vaccine promotion, reaching a potential of 9.9 million individuals. To measure the impact of the campaign, a cross-sectional panel based online survey was conducted before and after the campaign in 8 states in the US where the campaign was delivered and compared with 8 control states. As the survey was panel based and used samples, it may not have been the same people questioned both times.

The sample of the campaign group had significantly better knowledge and attitudes based on 9 questions at follow-up vs baseline (e.g. side effects of ‘flu vaccine are worse than flu), compared to the control group's sample. The improvement seen in the campaign group was mostly mild (e.g. proportion who disagreed that healthy people didn't need the vaccine changed from 57% to 61%, p=0.012), whereas for the control group, results at follow-up were mostly not statistically significant or they showed an improvement of 0.1–0.2% or more negative attitudes. At follow-up, 14.1% of respondents sampled from the campaign region reported seeing positive posts from people the influencers. They were significantly more likely to receive the ‘flu vaccine in the past 6 months than those not exposed to the influencers (50.9% vs 43.3%, P=0.05), with a significantly higher proportion agreeing that the vaccine is effective (58.2% vs 46.8%, p<0.01), is the best way to protect others from the flu, (64.2% vs 55.5%, p=0,.01), and is worth the time and effort (67.1% vs 59.1%, P=0.03).

Comments to posts reinforce existing attitudes

Two related studies looked at the impact of comments to a post and how reactions and comments to posts can influence others.

Kim et al (2020) investigated how Facebook comments can influence one's behaviour towards vaccines. Two hundred and seventy-one college students were randomly assigned to one of four experimental conditions and instructed to read a Facebook post and its corresponding comments thread (with varying mixes of positive and negative opinions). They responded to questions including their perceived distribution of public opinion on the flu vaccine and attitude towards the flu vaccine. The study found that numerical dominance of user generated comments indirectly influenced vaccine-related attitudes and perceived distribution of public opinion on vaccines. This was greater amongst participants with lower levels of vaccine acceptance.

Kim et al (2021) examined how social media users are influenced by user-generated comments on social media. Two hundred and sixty-four participants reported their pre-existing attitude towards the influenza vaccine and were randomly assigned to one of four experimental conditions. In each condition they read different Facebook posts and comments. The participants’ cognitive responses in reaction to the comments were then collected, including levels of anger and intention to receive the influenza vaccine. When comments were mostly negative, participants who had more positive pre-existing attitudes, had statistically significantly positive intention to vaccinate (b=.31, p < .001). When comments were positive, people with more negative pre-existing attitudes were statistically significantly not intending to be vaccinated (b=-.23, p<0.5). The findings suggest reactions to comments reinforce existing beliefs.

Verified sources and multi-tasking increase intention to vaccinate

Two studies looked at the effects of social media on the outcome of intention to vaccinate. One examined use of a social media platform to engage with others around an online article and whether that promoted more flu preventative behaviours (including vaccination) whilst the other looked at the difference between verifiable and nonverifiable sources on intention to vaccinate.

Kononova et al (2017) studied how health protective behavioural intentions are influenced by media multitasking (i.e. the concurrent use of multiple digital media streams). One hundred and twenty-one students read an online article about influenza which either included strong or weak suggestions to engage in flu-preventative behaviours. One group of participants read the article and checked Facebook (multi-tasking), and the other group just read the article. Participants then rated their intentions to follow flu-preventative suggestions provided in the article. The findings showed that participants in the media multitasking condition were more likely to follow the article flu-preventative suggestions than participants in the control condition.

Chen et al (2020) examined the effectiveness of fearinduced health campaigns on social media by verified and non-verified sources. Five hundred and thirty-four college students received one of 12 pictures containing influenza related fear appeal information from a visible source and retweeted on either a public (post wall) or private setting (chat frame). Participants who viewed a picture from a verified visible source developed a greater intention to receive the influenza vaccination [mean (M) = 5.08, SD = 1.41] than those who viewed a picture from an unverified source [M = 4.70, SD = 1.54, p < 0.01]. However, whether this was in a private or public setting, and whether the account retweeting the information was verified or unverified was not significant.

Social media only slightly improves vaccination uptake

Only one study looked at the direct impact of social media engagement on vaccine uptake (Bonnevie et al, 2019). Three studies reported on the added value (or not) of social media engagement to a web-based intervention that was designed to improve vaccine acceptance and uptake (Huang et al, 2018; Lau et al, 2012; O'Leary et al, 2019).

Bonnevie et al (2019) found that despite the use of campaigns by social media influencers resulting in a significant increase in specific knowledge and positive attitude towards the influenza vaccine, this was accompanied by a slight but not significant increase in people reporting receiving the vaccine in the campaign group. The vaccination rate was 45% in the campaign region, compared to 42% in the control group.

Huang et al (2018) assessed the impact of a campus community health worker programme on student influenza vaccination at a university with an enrolment of 6650 students over a period of five years. The initial intervention was an in-person peer outreach to promote the flu vaccine which showed a 66% increase in vaccinations. This was followed by the second intervention which was the addition of a personalised dormitory-specific social media campaign, which showed an 85% increase in vaccinations.

O'Leary et al (2019) conducted a randomised controlled trial of 1,093 participants to test the efficacy of an online vaccine resource in increasing the uptake of influenza vaccine in pregnant women. Participants were randomised into 3 groups – website with vaccine information and interactive social media components, website with vaccine information only, and usual care only. This study found that the addition of an interactive social media component only slightly increased the uptake beyond the web-based intervention - 57% of women received the influenza vaccine in the groups with vaccine information and interactive social media components, compared to 55% of the vaccine information only group. Only 36% of women in the usual care group had the vaccine.

Lau et al (2012) did a randomised controlled trial of 742 university students and staff. One group had access to a personally controlled health management system (which contained consumer care pathways, social forums, messaging links with a health provider service, and an online appointment booking system) and the second group were put on a 6 month waiting list for accessing the personally controlled health management system. They found that participants assigned to the online system were 6.7% more likely to receive an influenza vaccine than the waitlist group, and therefore the online system had a significant effect on influenza vaccination rates.

Discussion

To our knowledge, this is the only study to undertake a systematic review of the impact of social media interventions in improving acceptance and uptake of the adult ‘flu vaccination. The use of social media interventions to affect vaccine acceptance is a new but growing area of study. How a communication message is framed, who delivers the message, and network structure are critical for affecting the vaccine decision-making process. The studies in this review show that use of influencers, multi-tasking and the use of a verified source for information can impact on vaccine knowledge and intention to vaccinate. Interactions or comments on postings tended to reinforce preconceived ideas rather than change attitudes.

Vaccine campaigns on social media platforms have the potential of negatively impacting knowledge about immunisations and intention to vaccinate and therefore need to be strategic and carefully managed. Reactance needs to be considered when generating positive messages regarding vaccination as it could result in someone with pre-existing negative beliefs about ‘flu vaccines being more determined about not getting vaccinated. Anti-vaccine arguments, misinformation, and negative opinions in user-generated comments can impact an individual's perceived distribution of public opinion on vaccination. This could be detrimental to the persuasive potential of the message.

Using influencers is a promising strategy for communicating health information and increasing knowledge. Individuals are likely to engage in a positive way to a message if it is presented by an individual that they already admire. The use of influencers may be particularly useful when trying to reach specific at-risk groups with lower vaccination rates (Moorhead et al, 2013). Care is needed in selecting influencers for a general campaign as influencers may work for some cohorts and not for others and one person's influencer may be another's non-influencer. Media multi-tasking is a widespread habit, particularly of young adults. Processing different forms of media simultaneously can enhance learning (Wang et al, 2015). This suggests that social media should be used in conjunction with a planned campaign to improve reach. More exploration on the impact of media multi-tasking on vaccine acceptance is warranted.

In relation to use of verifiable sources, a systematic review by Limaye et al (2021) found that health information should be directly released by verified accounts in public way to increase individuals’ intention to perform health behaviours. However, we only found one study relating to that and so more research is needed on this.

None of the studies disaggregated results to show variation between different socio-economic groups or at-risk groups. However, this review has thrown up where health inequities persist. Wanting to have a vaccination is dependent upon accessibility of vaccination services and health care professionals’ advocacy for vaccinations. No study addressed how effective social media is in informing health care providers about adult vaccination programmes, and in turn how they can reach the under-served groups. In addition, the four studies that addressed improvement in vaccine uptake demonstrated that while social media campaigns can increase vaccine acceptance and intention to vaccinate, increasing vaccine uptake relies on improving access to vaccination services, e.g. having a convenient user-friendly online system which can improve engagement with the healthcare system. This then means that whilst you may be able to reach people through social media networks about ‘flu vaccine, their access to vaccination services remains determined by existing inequities in the healthcare system.

A strength of our review is we used a broad approach to uncover the impact of social media interventions on ‘flu vaccine acceptance and uptake. We were limited by the quality of studies and the lack of randomised controlled trials. There was bias in how the participants were selected. Participants were often students, who are not representative of the general population, particularly in relation to their use of technology and social media. Further bias was added through vaccine uptake being based on personal report and the use of online portal to sign up participants, whereby people have signed up to be included in studies.

More research needs to be done to investigate the role of web-based campaigns and social media in individuals with different levels of both consumer health literacy and technological literacy. Interventions were often heterogenous, making it difficult to determine whether the intervention of social media produced a large or significant enough effect.

We undertook a rapid systematic review approach and therefore screening of the studies was done by one author whilst two authors undertook the quality selection. We do not feel that this limitation affects the results of our review. However, our study is limited in its relevance to the UK setting where ‘flu vaccine is routinely offered to older adults and at-risk groups and not to university students. UK studies are warranted on this topic.

Conclusion

When linked to a web-based intervention, social media can improve ‘flu vaccine knowledge and intention to vaccinate. However, it needs to be strategically managed due to the many factors that may influence an individual's response to a social media campaign. This is a new and emerging field of study and higher quality evidence studies are needed to investigate the role of social media on vaccine knowledge in different demographics, and how to best optimise its role. Access to vaccination services remains a key determinant of vaccine uptake.

Reflective questions

• How can you use these study findings in your role?
• Are there applications your practice or organisation should consider when engaging with patients about ‘flu vaccine this winter?
• If you were to design a social media intervention on ‘flu vaccine (or another adult vaccination), what should you consider doing?

Key points

• There is little but growing evidence on social media interventions improving knowledge about ‘flu vaccine and intention to vaccinate.
• Promotion of ‘flu vaccine by influencers can positively influence vaccine knowledge.
• Communicating (e.g. tweeting) a verified source of information is likely to positively influence knowledge and intention to vaccinate.
• Interactions through comments on social media postings enforced preconceived ideas rather than change attitudes.
• Wide use of social media to support a ‘flu vaccine campaign has the potential to harness users’ multi-tasking ability to improve their knowledge.
• Whilst social media is unlikely to lead to improved vaccine uptake, access to services is key especially in terms of ease of booking appointments and outreach.