Difficulties faced by physicians from four European countries in rebutting antivaccination arguments: a cross-sectional study

Difficulties in dealing with vaccine resistance

Discussions about vaccines often do not occur in isolation—they are part of a wider healthcare system that builds on the ongoing relationship between physicians and their patients. This relationship can be helpful for encouraging vaccine recommendations, as it positions physicians as a trusted source of information.3 19 By the same token, physicians may find it more difficult to discuss vaccines if they worry that it will affect a longer-term relationship with their patients.8 20 Previous surveys of physicians responsible for delivery of the human papillomavirus (HPV) vaccination programmes in Australia and the USA found that physicians felt conversations with strongly hesitant patients (or their caregivers) to be especially challenging8 and many healthcare professionals (HCPs) felt they could not change their patients’/caregivers’ minds.16 Difficulties with vaccine conversations can influence vaccine recommendation behaviours.3 16 For example, healthcare providers who felt less confident to effectively address their patients’ concerns were less likely to routinely recommend the HPV vaccine to vaccine-eligible young people.16

While it is impossible to know in advance what concerns a specific patient may raise, research has shown that arguments opposing vaccines tend to cluster around a finite set of themes.21–23 ‘Arguments’ in this context refer to the propositions put forth by patients as a rationale for not having a vaccine. Many different studies have documented these arguments, resulting in a rich literature on reasons for vaccine rejection.18 However, although these arguments include those that physicians in previous studies had identified as their patients’ concerns24–26, there has yet to be an analysis of how physicians perceive the different arguments they could encounter.

There is a good reason to believe that physicians’ difficulty with rebutting arguments against vaccines would vary across different arguments. Physicians are trained to provide reassurance by giving their patients scientific facts about vaccines.27 This can be effective if the patient’s concern stems from a lack of knowledge, and the patient trusts the physician to provide that knowledge.28 However, there are many documented arguments against vaccination that scientific facts cannot directly address.18 Rather, these may be philosophical (eg, rejection of the epistemic basis of scientific knowledge29–31) or political (eg, rejection of vaccines along partisan political lines32–34), or may reflect an aversion to being told what to do (ie, reactance35). Moreover, even if a concern should in principle be assuaged with the correct knowledge, just providing facts is not always effective at dislodging misinformed beliefs.13 Providing facts can occasionally even backfire,12 particularly if an individual is motivated to interpret new information as supporting their strongly held belief.36 If a patient is motivated to reject a scientific counterargument, rather than explaining the science, physicians would need to address that underlying motivation to effectively deal with this type of concern.17 Physicians would likely find antivaccination arguments more difficult to address when the facts they have been trained to provide are insufficient as a counter.

Attitude roots of vaccine resistance

A patient’s stated reason to reject a vaccine can be conceptualised as the manifestation, or expression, of their underlying motivations for that rejection. These are likely linked to a number of psychological factors that consistently predict vaccine hesitancy, such as conspiracist ideation (ie, the tendency to believe in conspiracy theories35 37–40), distrust (eg, of healthcare systems40–42) and reactance (ie, the tendency to push back against a perceived imposition35 40), among others. The strength of the relationship between a certain attitude root and vaccine hesitancy may vary across countries—for example, conspiracist ideation was shown to be a consistently strong predictor of hesitancy across 24 different countries, whereas reactance was a predictor in some countries but not others.35 Nonetheless, these psychological factors shape and constrain people’s beliefs, attitudes and the expression of those, without the person necessarily being aware of it—leading to the terminology ‘attitude roots’ to describe these underlying motivations for people’s resistance to vaccination.17

In a systematic literature review, Fasce et al 18 analysed over 2000 documented arguments against vaccination and identified 11 attitude roots: conspiratorial ideation, distrust, unwarranted beliefs, worldview and politics, religious concerns, moral concerns, fear and phobias, distorted risk perception, selfishness, epistemic relativism, and reactance (see table 1 for definitions and example arguments of each attitude root). This 11-root taxonomy forms a comprehensive hierarchy to group different arguments against vaccines and opens an avenue to assessing whether arguments arising from different attitude roots generate varying degrees of difficulty for rebuttal by physicians. Such an assessment would help to ascertain where there may be communication skills gaps that could be addressed with tailored training.

The present study

The objective of our research was to understand how difficult physicians perceived it to rebut antivaccination arguments with different attitude roots, and how the perceived difficulty of rebuttal was associated with their vaccine recommendation behaviours and their preparedness to discuss vaccines. We compared the physicians’ perceived difficulties to rebut 33 different prototypical arguments that represented each of the 11 attitude roots in a taxonomy of antivaccination argumentation.18 We hypothesised that physicians would report differences in the perceived difficulty of rebutting arguments from different roots. In addition, since the strength of the relationships between attitude roots and vaccine hesitancy can vary among countries,35 it is worth considering whether the patterns in physicians’ difficulties with arguments of different attitude roots persist across countries. The majority of studies on antivaccination arguments have been done in English, but vaccine opposition—and indeed antivaccination misinformation—is known to persist among non-English-speaking populations as well.18 Therefore, we could expect there to be a general correlation among different countries in how physicians perceive the difficulty of rebutting arguments, but also some country-specific differences in difficulties with certain attitude roots.

Finally, we hypothesised that greater perceived difficulty in rebutting arguments would be negatively correlated with the frequency with which physicians recommend vaccines to patients and with physicians’ ‘proactive efficacy’—defined as how prepared they felt and how proactive they were during vaccine discussions.43

Participants

We collected data as part of an international cross-sectional survey on physicians’ vaccine confidence (conducted online between March and June 2022). Each country sent invitations to participate to physicians who had vaccination roles. Vaccination duties varied across countries and this is reflected in the distribution of physician types in the study (see table 2). Here, we report data (total n=2718 after excluding 157 incomplete responses to the questions investigated in this paper) from four European countries: France (n=1162), Finland (n=389), Portugal (n=560) and Germany (n=607). These countries were part of the JITSUVAX project submitted to and funded by the European Commission in 202044 and represent a broad spectrum of vaccine hesitancy across member states of the European Union (EU). A current report on vaccine hesitancy among European countries revealed that 58.9%–60.4% of the general public in Germany and Finland agree that vaccines are important, safe, effective and compatible with their beliefs.45 This agreement was only 46.8% in France but 75.3% in Portugal. Moreover, only 67.3% of HCPs in France agreed that vaccines are important, safe, effective and compatible with their beliefs while agreement ranged between 93.9% and 95.7% in Germany, Portugal and Finland.45 Differences are also observable for disease burden and vaccine uptake rates. For example, Germany and France were among the five countries that accounted for 77% of all measles cases in the EU in 2022 while Finland reported a single case and Portugal no cases over the last three reporting periods.46 In addition, only Portugal was among the five countries that reported a coverage of ≥95% for the second dose of a measles-containing vaccine in 2021 (Germany: 93%; Finland: 93%; France: 86%). We assumed that these differences in country profiles also indicated differences in physicians’ difficulty ratings in dealing with arguments against vaccination. The demographic characteristics of the survey sample are reported in table 2. We provide analyses of the overall sample as well as per country.

Materials and procedure

We collected data as part of an international questionnaire on physicians’ vaccination attitudes and behaviours. Perceived difficulty to rebut antivaccination arguments was our main variable of interest. It was presented at the end of the questionnaire. Participants provided ratings of 33 arguments (3 arguments per attitude root), in response to the following question:

Below is a list of anti-vaccination arguments. All the arguments are false or misleading and have been repeatedly debunked.

Please read the messages below and indicate for each message how easy you would find it to rebut the message while interacting with a patient.

The arguments were then presented in a matrix format, in a random order for each participant. Participants responded to each argument on a 5-point Likert scale (1: I would find it very easy, 2: I would find it rather easy, 3: undecided, 4: I would find it rather difficult and 5: I would find it very difficult). The full list of prototypical arguments for each attitude root is shown in table 1. We calculated the average difficulty rating for arguments within each attitude root, as well as an overall difficulty rating across all arguments. Means and SD of ratings for all arguments by physicians in each of the countries, and correlations between each country’s average difficulty ratings of each argument as well as correlations between participants’ average difficulty rating per attitude root among the 11 attitude roots can be found in online supplemental tables S1–S5. Overall, there were strong correlations between ratings from each country (r=0.75–0.90) and for the different attitude roots (r=0.52–0.84).

The overall questionnaire included the I-Pro-VC-Be (a series of questions on vaccination recommendation behaviour, own vaccination status and determinants of vaccination attitudes47), and several other scale measures related to other research questions not investigated in this paper. This questionnaire was validated in all four countries47 and the full wording is provided asonline supplemental file 2. We aimed to determine the association of physicians’ difficulty rebutting arguments with physicians’ vaccine recommendation behaviours and preparedness to discuss vaccines. Therefore, we preregistered analyses with two key variables from the I-Pro-VC-Be that were targeted at measuring these validated constructs,47 as described below.

Analysis strategy

To assess differences in difficulty to rebut arguments between countries and attitude roots, we preregistered a between-within analysis of variance on the perceived difficulty variable, with attitude root as a within-subjects factor and country as a between-subjects factor. In addition, to assess the contribution of difficulty ratings for the antivaccination arguments towards physicians’ vaccine recommendation behaviours and proactive efficacy around these behaviours, we analysed the average difficulty rating across all the arguments as a predictor of two outcome variables: average recommendation behaviour (frequency or intention) and proactive self-efficacy. For each of these outcomes, we ran a mixed-effects linear model using the lme4 package in R48 that included the rebuttal difficulty score as a fixed effect and country as a random effect. This model deviated slightly from our pre-registered plan to analyse the contribution of difficulty ratings for each root towards physicians’ vaccine recommendation behaviours and proactive efficacy around vaccination. The issue of multicollinearity (see online supplemental table S4) prevented use of the preregistered models. We thus used an overall rebuttal difficulty score.

Patient and public involvement

As the relevant public group to involve in the conduct of the research, HCPs from all the participating countries were involved in designing the study questionnaire through one-on-one interviews (n=28) and a pilot test (n=272). Once the study has been published, participants will be informed of the results through a dedicated website (https://sks.to/jitsuvax).

Arguments of different attitude roots pose different levels of difficulty for physicians

Within each attitude root, difficulty ratings for arguments had good reliability (α>0.75). We calculated an average difficulty rating per attitude root per country. As shown in figure 1, physicians across all countries rated arguments from some roots as more difficult than others. Across the full sample, arguments based on ‘religious concerns’ and ‘reactance’ were rated as the most difficult roots to rebut, while arguments based on ‘distorted risk perception’ were the easiest. This pattern was consistent across all countries except Germany, where the most difficult attitude root to rebut was ‘distrust’.

Request permission

Difficulty ratings differed significantly between attitude roots, F(7, 18137)=201.75, p<0.001, η^{_p 2}=0.07. Difficulty ratings also varied significantly between countries F(3, 2714)=10.99, p<0.001, η^{_p 2}=0.01. On average, French physicians indicated the most difficulty with arguments (M=2.29, SD=0.78), followed by German physicians (M=2.24, SD=0.90), Portuguese physicians (M=2.18, SD=0.98) and finally Finnish physicians (M=2.00, SD=0.78). All follow-up pairwise comparisons among countries were significant at p<0.01 after applying a Bonferroni correction (see online supplemental table S1).

The interaction between attitude root and country was significant, indicating that rebuttal difficulty across arguments varied among countries F(20, 18137)=13.41, p<0.001, η^{_p 2}=0.02. To assess differences in difficulties with each attitude root for each country, we used one-sample t-tests to compare average ratings for each of the 11 attitude roots in a country to the mean difficulty rating of all arguments in that country. We used this approach because within each country, it is of more practical importance to know whether particular attitude roots stand out as significantly more or less difficult than the average argument difficulty in that country, rather than being concerned with specific comparisons among roots. Because this resulted in 11 comparisons per country, we applied a Bonferroni correction of α/11.

As shown in table 3, some patterns were similar across all countries: distorted risk perception arguments were on average reported as significantly easier to rebut compared with the country mean, whereas reactance arguments were on average reported as significantly more difficult to rebut compared with the country mean. French physicians showed the greatest variation in difficulty, with all mean root ratings differing significantly from the overall country mean. While the larger sample in France enabled greater power to detect significant effects, the effect sizes for the significant effects are still comparable to those of significant effects in the other countries. Physicians in Finland and Portugal also reported religious-concerns arguments as significantly more difficult and perceived self-interest arguments as significantly less difficult. German physicians reported distrust as the only other significantly more difficult (and indeed, the most difficult) attitude root.

Difficulty rebutting arguments is associated with lower proactive efficacy and recommendation of vaccines

Physicians who found arguments more difficult to rebut tended to recommend vaccinations significantly less, β=−0.18, p<0.001 (95% CI −0.22 to –0.15), R²=0.03. Physicians who perceived greater difficulty rebutting arguments also reported significantly lower proactive self-efficacy, that is, lower commitment to vaccination and self-efficacy, β=−0.23, p<0.001 (95% CI −0.26 to –0.19), R²=0.05. These associations remained significant in robustness checks that controlled for other I-Pro-VC-Be variables and analysed the individual recommendation and intentions variables independently (see online supplemental tables S6 and S7).

Limitations

There are some limitations to this work. First, all data are self-reported quantitative measurements. Thus, the absolute values of physicians’ difficulty ratings may be biased due to the overestimation or underestimation of their own abilities in dealing with patients’ arguments, or physicians’ discomfort in reporting difficulties refuting certain arguments about vaccination. Therefore, we mainly discuss the differences in difficulty ratings across countries and attitude roots, rather than the absolute values of the difficulty ratings. Future work could include qualitative assessments, which may bring up further insights and additional strategies physicians usually used to counter antivaccination arguments from patients.

Due to time constraints, we did not assess the frequency with which physicians encountered each argument (or one similar to it), so we cannot be sure that the difficulty ratings were not an indirect proxy of the frequency with which physicians are confronted with certain arguments. Physicians also did not indicate the type of rapport they have with their patients, which may also affect how difficult they find it to counter certain arguments. The results of this study can, therefore, only be a building block to assess which attitude roots are most relevant for the day-to-day work of physicians in the four European countries, and future work may wish to assess other factors that could potentially affect physicians’ difficulty and boost their self-efficacy in rebutting antivaccination arguments.

Finally, our sample of physicians was recruited through convenience sampling in only the four targeted countries, and over-represented general practitioners, which may have meant we do not represent all doctors in each country. It would be good for future research to extend our findings with wider groups of HCPs (eg, nurses, midwives) who also have vaccination roles, and in other countries. Studying HCPs who may also specialise in different subgroups of patients could also give an indication of what arguments HCPs find harder to rebut among these patient groups.