Abstract
A growing tendency to refuse child vaccination is commonly regarded as a reason for concern. Attempts to promote vaccination by authorities often prove to be unsuccessful or even counter-productive. The aim of this study was to explore how parents perceived four messages used to promote vaccination. In eight focus groups (N = 73), we presented participants with messages and asked them to assess them. Using thematic analysis, we identified themes that our participants considered the most important. Messages that promote vaccination should be unambiguous, more balanced, not focus on repeating the negative effects of vaccine-preventable diseases and provide links to the evidence.
Objective
Refusing to vaccinate a child is a major public health issue. Central European countries are currently facing a paradigm change; in the Socialist system, vaccination used to be obligatory and was not a subject of discussion (Hnilicová, 2007). Societal changes, a shift towards individualism and the global trend of promoting patient autonomy and shared decision-making that arrived in the 1990s (Chin, 2002; Coulter and Collins, 2011) led to increased resistance against child vaccination. For instance, in Slovakia, the level of vaccination in the 2012 cohort dropped below 95 per cent in 4 regions and 36 districts (Slovak Public Health Authority, 2014).
One of the reasons for low levels of vaccination seems to be effective campaigning by anti-vaccination proponents and groups (Casiday, 2005). Their messages are effective in reaching parents who are making decisions about vaccinating their children (Brunson, 2013; Glanz et al., 2013).
However, maintaining high levels of vaccination is a vital public health priority; an argument for this could be the increased number of measles cases observed when more people opt out of measles vaccination (Centers for Disease Control and Prevention - CDC), 2012).
Public health authorities, such as the Centers for Disease Control and Prevention, are trying to counter these anti-vaccination messages with their own campaigns (Nyhan et al., 2014). As shown in studies by Nyhan et al. (2014) and Horne et al. (2015), not only do these interventions not work, but sometimes they are even counter-productive. Nyhan et al. (2014) tested four interventions (autism correction, disease risk, disease narrative and disease images; for details of these interventions, please see section ‘Material’), and none of them increased parental intentions to vaccinate a future child. In fact, one message even decreased the target group’s intention to vaccinate their children. The authors (Nyhan et al., 2014) explain this primarily by the fact that messages bring to mind other concerns about vaccination; furthermore, the authors point out the mixed effects of loss-framed messages and fear and a danger-priming effect. Health narratives that induce fear are less effective in changing beliefs and attitudes.
Horne et al. (2015) came to a similar conclusion when studying the same four interventions, yet they claim that the attitudes of parents could be influenced by focusing on the dangers of communicable diseases.
None of these studies, however, looked deeper into what makes these messages ineffective and how exactly the target group views them.
In this study, we re-visited the four pro-vaccination interventions used in the studies by Nyhan et al. (2014) and Horne et al. (2015) and analysed them in a focus group setting attended by members of the target group. Our aim was to see how the members of the target group view these messages and to establish exactly which parts they find ineffective.
Design
Our objective was to explore how parents perceive such messages. Using qualitative methods, we intended to provide in-depth insights into why the messages fail to persuade parents to have their children vaccinated.
Participants
The study involved two groups. Group 1 (Students group) consisted of female undergraduate university students (N = 39; mean age = 21.54, ranging from 19 to 26 years of age) divided into four focus groups. All were white Slovak nationals, and none of them had children. The majority of the participants came from the capital, Bratislava, and 16.66 per cent came from other municipalities in Slovakia.
Group 2 (Mothers group) comprised mothers with at least secondary education (N = 34; mean age = 33.54, ranging from 26 to 45 years of age) also divided into four focus groups. All were white Slovak nationals, and all had children who had passed the age of first vaccination but were not older than 5 years of age. The average number of children was 1.59. Out of the total number, 27 participants (79.41%) identified themselves as pro-vaccination and 7 participants (20.59%) as anti-vaccination when asked in a questionnaire before the groups began.
Ethnicity and nationality is reflective of the population of the country. The focus on all-female participants was driven by our previous observation that children’s early age medical care is mostly considered the responsibility of the mother, who makes most of the decisions related to the child’s healthcare. However, in a separate on-going study, we are also trying to explore the specific role of fathers in making vaccination-related decisions.
The university participants (Group 1) were participants of a college course. The mothers (Group 2) were recruited by a survey agency using the criteria above.
To add some context, Slovakia is a Central European country and a member of the European Union. Vaccines are accessible, vaccination is obligatory, and failure to vaccinate a child may lead to financial penalties. The recent popularity of anti-vaccination movements and drops in vaccination rates are giving rise to general concerns of healthcare authorities (Slovak Public Health Authority, 2014).
Material
We presented the participants with the four interventions used in the original study by Nyhan et al. (2014) and then later by Horne et al. (2015). The four interventions were as follows:
Autism correction
A brief text explaining that although some parents may worry that the vaccine causes autism, many scientific studies have found no link between the measles, mumps and rubella (MMR) vaccine and autism. The text includes references to three studies from 2002 to 2008 that found no link between autism and vaccination.
Disease risk
A brief text explaining the seriousness of the three diseases covered by the MMR vaccine and listing the symptoms of MMR. The text explains that the MMR vaccine can prevent these diseases; most children who get their MMR shots will not get these diseases and many more children would get them if we stopped vaccinating.
Disease narrative
A brief narrative story of a Megan Campbell’s 10-month-old son, who fought for his life after contracting measles. The text includes a direct quote from the mother’s account of the story and a brief note saying that her son was exposed to measles when another mother brought her ill son into their paediatrician’s waiting room.
Disease images
This message states that all children should be vaccinated, claims that MMR are serious diseases and states that most people in the United States have never seen a child with these diseases. This is followed by three images of children, each with one of the three diseases.
The four messages focused on vaccination against MMR, so this was naturally the main focus of our discussions; in all groups, however, we also discussed other diseases and child vaccination in general, not only MMR.
Procedure
As young mothers regularly discuss various aspects of childcare in informal groups or online forums, the study used focus groups to simulate ‘the thinking society in miniature’ (Marková and Farr, 1995).
We conducted focus groups with university students in a small quiet lecture room and focus groups with mothers in a special focus group laboratory operated by the survey agency that recruited our participants. The data were collected in the second half of 2014. We asked the participants to read the first message, then to individually assign values indicating how understandable they thought the message was and how persuasive they thought the message was. Then, we asked them to provide arguments and a rationale for their choice and let them engage with each other to explore their arguments and positions. The facilitators asked supplementary questions and made sure the discussion stayed on topic. Then, we repeated this procedure with the second, third and fourth messages. In conclusion, the participants were asked to provide additional comments about the topic.
The facilitators were the authors of this study. Both had extensive experience in facilitating focus groups or therapy groups. We took every possible step to create a positive and safe environment for the participants. The feedback from the participants at the end of the sessions confirmed that they felt the discussions had been open and that the focus group environment had felt safe. The study design was reviewed by the Scientific Board of the Institute of Experimental Psychology and was found to be in accordance with ethical principles. All participants provided oral informed consent which was recorded at the beginning of every focus group, and the participants recruited by the professional agency (mothers) also signed informed consent.
Data analysis
We used thematic analysis as the conceptual framework. The recordings were transcribed and checked for accuracy, and we used the coding process to identify major themes. We coded the data independently and discussed any differences in codes and coded segments until we arrived at the consensual codebook. Throughout the process, we followed Braun and Clarke’s (2006) good practice guidelines.
Validity
The validity of the data collected in the focus groups could be tested by comparing the responses of several focus groups (Bender and Ewbank, 1994). All of our observations occurred repeatedly across groups, and themes identified in the two demographics did not seem to differ beyond the fact that the mothers were mostly building on their own experience, whereas the students were mostly relating stories from their surroundings, experience with vaccination of younger siblings or dilemmas faced by older siblings or friends who already have children. Moreover, as the authors (Bender and Ewbank, 1994) suggest, we compared qualitative answers with data from questionnaires. We also made every effort not to interfere with answers and did not provide our own opinions or take sides in any form. Finally, a strong case for validity was our sampling procedure. The participants were not recruited by the ‘snowball sampling’ procedure but by interviewing all students in a particular college course (Group 1) and by interviewing a sample composed by a professional survey agency based on quotas used in representative sampling (Group 2).
Results
First, the students and mothers were asked to rate two parameters of the messages: whether they were understandable and whether they found them persuasive. All four messages were rated as understandable (mean value of 1.36 on a scale where 1 = excellent and 5 = inadequate); however, the persuasive rating was much more negative (a mean value of 3.05 on the same scale). For ratings of individual messages, please see Table 1. The ratings were recorded individually in writing before the group discussions. The findings were similar to those reported by Nyhan et al. (2014).
Understandability and persuasiveness of the four messages as reported by our participants before discussion.
We further explored the reasons provided by our participants for their statements. Under the following four headings, we list the main themes identified in the focus group discussions. The themes are ordered based on how often they were mentioned, starting with the one mentioned the most frequently for the given category. Subheadings in italics attempt to summarize the most important themes. Quotations are used to illustrate a point which was usually expressed in slightly different words by multiple participants.
Autism correction
Message does not provide enough information
The participants repeatedly mentioned the need for graphs, links, citations and methodology details. The studies cited within the message were considered to be outdated (the most recent study in the information was from 2008; students and mothers, particularly those with some academic background, indicated that they would need more recent material to be persuaded).
Information was considered to be one-sided
The message only covered autism and did not mention other possible side-effects of vaccination. Such a one-sided focus may trigger ambiguity aversion heuristics (Hertwig et al., 2013; Ritov and Baron, 1990; Zimmerman et al., 2005), where the target group dismisses an argument as being just one of several possible outlooks on the problem (participants repeatedly phrased this as ‘some experts say this, other experts say that’).
Studies cited in the messages were cherry-picked by pharmaceutical companies
Our focus groups repeatedly came back to the ‘Big Pharma’ narrative of corrupt pharmaceutical companies using financial incentives to manipulate healthcare professionals into prescribing unnecessary medications. Because of this general mistrust, institutions are often not seen as credible.
Authors are trying too hard
After being exposed to this message, many participants strongly expressed the idea that they were being manipulated. As one participant remarked, the eagerness of the authorities to challenge the link between vaccination and autism was considered suspicious: ‘They repeat the message so many times that I have a terrible feeling of being manipulated. (…) Why do they repeat it so much? As if they strongly wanted us to believe them’ (Students Group 1, 20-year-old participant).
Text does not provide reasons for vaccination
Although the text clearly states that ‘all children should be vaccinated for MMR’ in its first sentence, the participants repeatedly suggested they had not been told specific reasons why they should have their children vaccinated. They understand that the link between vaccination and autism is dubious, but the text does not give them a positive reason for vaccinating: ‘If they want to persuade someone, they should say why it is beneficial, but this text does not include such information’ (Mothers Group 3, 33-year-old participant with two children of 2 years and 7 months).
Disease risks
Oversimplifying
This theme was always triggered by the word ‘most’ in the sentence ‘Most children who get their MMR shots will not get these diseases’. The participants picked up on this in almost every group, and thought it was unfair to simplify this. As one participant stated, the statement did not make it clear if ‘most’ was 60 or 99 per cent.
Participants asked for more details
Just as with the previous message, the participants would have liked to see exact percentages, illustrations, references to studies and so on.
Intention of authors was seen as too obvious
The message was almost generally perceived to be an attempt to induce fear. One person remarked, ‘I found it completely pathetic. The person who wrote this was grasping at straws […] Like, you don’t get vaccinated, you get sick. A 100 per cent chance. He mentioned all the worst things that could happen’ (Students Group 2, 21-year-old participant). This person thought the message was ‘grasping at straws’, because the organization seemed desperate to use any means possible to push the pro-vaccination message.
Disease narrative
No perceived connection with vaccination
The message did not make sense to most of our participants: it was a narrative about a child fighting a serious disease, yet it was unclear to them how this was connected with vaccination. It was not clear to our participants if vaccination would have prevented the condition, or if it would have made the symptoms less severe. The message, moreover, does not explicitly give a clear recommendation to vaccinate.
Emotional blackmail
The story was found to be too artificial, not credible and reminiscent of the tabloid press or television shows that exploit emotions. It was considered a cold-blooded attempt to manipulate; one respondent even suggested that the story was not real and that the protagonists were paid financial compensation to allow the use of their names: ‘I think this is just horrible. This style of advertisement. It does not work for me. […] I take it those people were paid money to say that’ (Students Group 4, 20-year-old participant). The negative response to this and the following message can also be partially explained by the danger-priming effect (Nyhan et al., 2014).
This is just a single story that does not prove anything
Participants did not understand why the story should be generalizable – it was a specific case and it was not clear how it was linked to vaccination. Several mothers also reported familiarity with the symptoms, because their child had been in a similar state when fighting a disease which was not vaccine-preventable.
Several participants also observed a secondary unintended message: that the child in the narrative was infected when visiting a doctor. This led them to the conclusion that if the child had not visited a doctor, he would not have contracted the condition – which to our participants meant that the medical system should admit responsibility and not try to blame parents who had opted out of vaccination.
Disease images
Emotional blackmail
Similar to the previous message, the images were considered a crude attempt to manipulate emotions.
The message contains no facts
The participants did not like the idea that the images were supposed to speak for themselves. The participants thought the main point should be clearly verbalized and argued.
The source of the pictures is dubious
To make this message effective, it would be necessary to communicate to the target group that the pictures are real-life cases of unvaccinated children who caught vaccine-preventable diseases. However, this premise seems to fail on three levels. First, the participants are familiar with the manipulation technique of using images out of context to prove a point, and this message is often considered a prime example of such manipulation – as one participant suggested, it was ‘… as if someone had downloaded images from Google and presented them like this’ (Students Group 1, 24-year-old participant). Second, people seem to be desensitized to ‘shocking’ and frightening images, and they seem to react in a detached and sometimes cynical manner. And third, several mothers suggested that they had repeatedly seen rashes or swelling in images as a consequence of a different disease. One mother even suggested that her child had looked very similar to the children in the picture when suffering from the side-effects of vaccination.
Conclusion
Our respondents generally demonstrated serious concerns in relation to vaccination. Pro-vaccination messages did not eliminate these concerns and even seemed to make them worse.
Vaccination induces fear, mostly in relation to negative side-effects and false immunological claims (Dannetun et al., 2005; Poland and Jacobson, 2012; Poland and Spier, 2010). Moreover, the counter-intuitive nature of vaccination (Miton and Mercier, 2015) seems to bring in multiple cognitive mechanisms that tend to problematize vaccination beliefs.
Hobson-West (2003) notes that having an understanding of basic categories of health and disease is crucial to accepting or resisting immunization. The concept of vaccination is difficult to understand, and it is liable to biases, heuristics and questioning. Confirmation bias, the familiarity effect of anti-vaccination messages and the absence of risk make it difficult to judge arguments (both for and against) correctly (Graves, 2015; Horne et al., 2015; Lord et al., 1979). Many parents do not understand why they should vaccinate, and medical professionals often fail to provide them with good enough reasons. The chief factor, however, seemed to be a lack of trust in the authorities that deliver these messages.
Previous research suggests that it matters who provides the vaccination information and to whom, when and in what manner (Fiks and Jimenez, 2010; Ołpiński, 2012). Our participants were suspicious about the intentions of the messengers and the overall role of the pharmaceutical and medical systems. Given doubts about the intentions of the medical authorities, even well-meant pro-vaccination messages may be neutralized because of low trust in the medical system. Studies repeatedly suggest that most parents trust their child’s paediatrician to communicate information about safe vaccination to them, but there is another group of parents who trust sources other than their healthcare providers – for example, other providers, other experts, family, friends or even celebrities (Freed et al., 2011; Gellin et al., 2000; Ołpiński, 2012).
With regard to the content, personal narratives and persuasive texts by anti-vaccination promoters and by the authorities seem to have different effects. Our participants were very critical of the parts of the messages that contained stories, images and emotional accounts. They requested more detailed information about the sources and doubted the legitimacy of the information. While social reference generally tends to give the narrative an air of credibility, when the authorities use this technique, they fail and may even be counter-productive. Such observations are in line with findings about strategies for processing distorted information (Cook and Lewandowsky, 2011; Lewandowsky et al., 2012). Decision-making in mothers includes eliminating risks and making informed decisions (Reich, 2014). This could be the reason why narrative information is evaluated more strictly when published in the media or on the Internet or when provided in pro-vaccination leaflets (compared to informal narratives, which are often taken at face value).
The lack of trust seems to be apparent on the level of institutions as a whole (authorities, medical facilities, the pharmaceutical industry) but also when dealing with practitioners. As confirmed in other studies, medical professionals seem to lack the knowledge and skills to debunk vaccination myths (Cook and Lewandowsky, 2011; Lewandowsky et al., 2012). Medical paternalism, where a physician takes over responsibility for the patient (Gert and Culver, 1979), is currently being replaced by shared decision-making, which leads to the emergence of new types of patients and the need for a new type of practitioner (Edwards and Elwyn, 2001; Makoul and Clayman, 2006). Many medical professionals, however, still act from a position of power, which may alienate patients and increase the vaccination rejection rate (Graves, 2015). Shared decision-making obviously has its limits (Fiks and Jimenez, 2010; Kravitz and Melnikow, 2001). As parents try to balance the array of potential risks of vaccinating (or not vaccinating), patronizing responses by the medical community may drive them towards using confirmation bias and ambiguity aversion when making decisions and ultimately towards adopting anti-vaccination attitudes (Asch et al., 1994; Ritov and Baron, 1990).
Moving from authorities to opposition groups, it is often presumed that one of the reasons for decreasing vaccination rates is the existence of anti-vaccination groups and their messages (Blume, 2006). When we contrast the way in which pro-vaccination and anti-vaccination messages are communicated, the latter seem to empower their audience by asking them to avoid being manipulated, by focusing on positive values such as freedom, or by helping them uncover links between medicine and the pharmaceutical business (Penta and Baban, 2014). This places the individual in a position of strength: by doing something as relatively simple as rejecting vaccination, they can actually grow stronger in their position of relative vulnerability against the powerful and sometimes confusing position of the state in vaccination policy (Colgrove and Bayer, 2005; Larson et al., 2011). In contrast, pro-vaccination messages tend to do the opposite: their messages are felt to be manipulative, patronizing, condescending, and even to belittle the recipient who may have doubts about vaccination. Our participants repeatedly confirmed this by clearly stating that pro-vaccination messages should not try to hide any associated risks (e.g. side-effects). Betsch and Sachse (2013) found that messages strongly indicating that there is ‘no risk’ led to a higher perceived vaccination risk than weak negations. As a result, moderate formulations could be more effective than extreme formulations.
In summary, society is in need of new approaches to deliver pro-vaccination messages to offset effective anti-vaccination narratives. Messages that promote vaccination should be more balanced and not focus on repeating the negative effects of vaccine-preventable diseases. They should talk about both sides of the story, the benefits as well as the side-effects. They should provide links to the evidence and arguments, and additional evidence should always follow the main message. The recommendation to vaccinate must be clear to avoid ambiguity. However, what is perhaps most important is that pro-vaccination messages should strive to empower their target group instead of doing the opposite.
Footnotes
Acknowledgements
The authors would like to thank Jana Bašnáková for inspiring discussions that led to the idea for this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by VEGA grant No. 2/0154/13 – ‘Social Influences in Individual Decision Making’ awarded by the Ministry of Education, Science, Research and Sports of the Slovak Republic and the Slovak Academy of Sciences.