Published in The British Medical Journal - 15th June 2021

A vaccine against COVID-19 is a ‘vital tool’ in the management of the current pandemic.1 Accordingly, extraordinary resources have been invested into vaccine development with unprecedented speed. Yet, even as approved vaccines become available, societies across the world still face another challenge: vaccine scepticism.

As of late 2020, researchers estimated that up to 82% of a country’s population may need to be vaccinated in order to reach herd immunity against SARS-CoV-2,2 3 and the emergence of new virus variants implies that individuals may need to get vaccinated repeatedly. However, general vaccine hesitancy has been on the rise in recent years in many countries.4 5 This has been the case for many non-COVID-19 vaccine programmes and is likely to pose a challenge for COVID-19 vaccines.6 7 Consistent with this, initial cross-national survey evidence suggests that substantially fewer people worldwide are willing to get vaccinated than would be necessary, and that some countries—for example, Russia, Poland and France—face strikingly high levels of scepticism.8 Thus, a key challenge for pandemic management is for health authorities across the world to encourage people to accept approved COVID-19 vaccines through careful approval procedures and effective health communication. This latter challenge emphasises the importance of understanding why people are hesitant about taking vaccines. Such knowledge is crucial for guiding communication in a way that increases vaccine acceptance and for understanding how to prepare for future health emergencies.

In this article, we first present descriptive analyses of the acceptance of a COVID-19 vaccine approved and recommended by health authorities across eight Western democracies. Second, we investigate individual-level predictors of vaccine acceptance. Third, we also explore macro-level correlations of vaccine acceptance. The study was conducted from the fall of 2020 to the winter of 2021. This data collection thus allows us to track levels and predictors of vaccine acceptance as vaccines were approved using large-scale cross-national surveys including a broad set of potential predictors, eg, political predictors, which are less often explored in traditional health research. Given the scale and broad impact of a pandemic crisis, however, such broader predictors may be particularly relevant to explore.

Potential predictors: who are expected to accept a COVID-19 vaccine?

To organise our expectations about the individual-level predictors of vaccine acceptance during the COVID-19 pandemic, we draw on one of the most comprehensive frameworks for understanding the antecedents of vaccine acceptance; the 5C model from Betsch et al.9 According to the 5C model, five psychological antecedents drive vaccine acceptance: confidence, constraints, complacency, calculation and collective responsibility. While we consider multiple predictors that are often not considered within this model, we strengthen the model’s coverage by theorising the link between the components of the model and the novel predictors that may be important for vaccine acceptance during the COVID-19 pandemic.

Confidence is defined as trust in (1) the effectiveness and safety of vaccines, (2) the system that delivers them and (3) the motivation of policy-makers who decide on the need for vaccines.9 10 Here, we consider two categories of predictors that reflect the underlying dimensions of confidence. First, we broadly tap into the second dimension of the definition by focusing on trust in a range of actors. Second, we investigate a range of political attitudes that broadly reflect the third dimension of the definition.

Empirically, trust is a crucial predictor of vaccine acceptance. Guay et al, for example, found that distrust in public health authorities is associated with general vaccine hesitancy.11 Similarly, people who trust official authorities were more likely to accept the human papillomavirus vaccine (HPV).12 Initial work on COVID-19 vaccines also demonstrates that those who have higher trust in scientists are more willing to get vaccinated.13

Furthermore, the literature on vaccine hesitancy has found that hesitancy is integrated into a broader set of political attitudes and perceptions. Political ideology has been found to be related to vaccine hesitancy as conservative individuals are less likely to trust authorities.14 Furthermore, it is a standard finding in political science that individuals are less likely to accept decisions from other political parties than the one they identify with or vote for.15 Thus, it is plausible that people who have voted for the government party/candidate are more likely to accept a vaccine, since the vaccine programme is a part of the governments’ response to the pandemic. In addition to these standard political attitudes, more extreme attitudes may also influence confidence in vaccines. Most prominently, people prone to conspiracy thinking are more likely to be hesitant about vaccines.16 17 In the context of the COVID-19 pandemic, higher levels of conspiratorial thinking have also been found to be associated with lower acceptance of future vaccines against COVID-19.18–20 Consequently, it can be expected that vaccine acceptance during the COVID-19 pandemic may relate to antisystemic sentiments such as conspiratorial thinking. We examine three levels of antisystemic sentiments and how they relate to vaccine acceptance, including (1) concern for democratic rights, (2) support for public protests against government policies and (3) beliefs in specific conspiracy theories related to COVID-19. Finally, we also examine the role of awareness of misinformation. From the literature, we know that susceptibility to misinformation negatively affects people’s acceptance of a vaccine against COVID-19.21 However, studies have also shown that prebunking can help cultivate ‘mental antibodies’ against misinformation.22 23 Thus, it is likely that awareness of misinformation is positively associated with vaccine acceptance.

Constraints refer to the structural and psychological barriers, impeding the implementation of vaccination intentions into behaviour.9 We consider the feeling of ‘pandemic fatigue’ as such a barrier and thus a potential correlate of vaccine acceptance. While the WHO has been warning about fatigue among populations in the fall of 2020,24 25 it is plausible that people who feel fatigued are willing to do what it takes to end the pandemic, including being vaccinated. However, fatigue could also generate an unwillingness or incapability to comply with further requirements, including vaccinations. Furthermore, we include the sense of having sufficient knowledge about behavioural recommendations as another psychological barrier. A sense of self-efficacy about proper behaviour was one of the best predictors of compliance with physical distancing policies during the first wave of the pandemic.26 Furthermore, perceived insufficient knowledge is significantly associated with general vaccine hesitancy.11 Finally, we assess remaining psychological constraints by assessing to what extent people report being able to change their behaviour in accordance with the recommendations from health authorities during the pandemic. This general measure of behaviour change should serve as a proxy for the range of constraints that may serve as a barrier for action over and beyond the directly assessed factors.

Complacency ‘exists where perceived risks of vaccine-preventable diseases are low and vaccination is not deemed a necessary preventive action’.9 10 Here we consider two types of predictors, including demographic factors and corona-specific risk perceptions.

First, a set of demographic predictors are expected to be associated with complacency. Thus, prior studies have found that men are more likely than women to accept a potential COVID-19 vaccine,27–29 potentially due to sex-based differences in COVID-19 mortality.27 Likewise, older people are expected to be more willing to take a vaccination due to higher risks of severe infections. This is supported by Lazarus et al8 and Hacquin et al,29 while neither Dror et al27 nor Wong et al28 found any age differences in vaccine acceptance. As a final demographic variable, we also consider education, even though this variable may influence vaccine acceptance through other dimensions than complacency (eg, confidence). The findings of prior studies on vaccine hesitancy are mixed with regard to education, indicating that that the association between education and vaccine hesitancy is context specific.30 To illustrate, whereas Guay et al found that lower education was associated with general vaccine hesitancy in Canada,11 Wagner et al found that educational level was not associated with general vaccine hesitancy across five low-income and middle-income countries.30 Similarly, Bertoncello et al found that while low parent education was significantly associated with general vaccine hesitancy, it was not associated with hesitancy in the context of child vaccine programmes in Italy.31 In the context of COVID-19, studies have found that higher education is associated with higher levels of vaccine acceptance.8 29

Second, we also investigate the role of personal risk perceptions. Several studies have found that self-perceived risks of COVID-19 positively predicts acceptance of potential COVID-19 vaccines.19 27 29 Likewise, Wong et al argued that perceived susceptibility to infection predicted the intention to take a future COVID-19 vaccine.28 Thus, we expect that personal risk perception predicts vaccine acceptance.

Collective responsibility is defined as the willingness to protect others by one’s own vaccination by means of herd immunity.9 32 We consider three groups of predictors to be relevant for this category of vaccine antecedents: (1) prosocial concerns (ie, concern for others), (2) support for pandemic restrictions and (3) interpersonal trust.

Focusing first on prosocial concerns, we measure a range of concerns over the disease’s impact on society, including hospitals’ ability to help the sick, society’s ability to help the disadvantaged, social unrest and crime, and the country’s economy. These concerns clearly tap into the collective responsibility and can be expected to positively predict vaccine acceptance, given that vaccine uptake can be viewed as a form of other directed behaviour that protects individuals beyond the self.

Second, we examine the association between compliance during the COVID-19 pandemic and vaccine acceptance. Protective behaviour thus might be viewed as a collective good, implying that compliance with health advice might reflect the willingness to protect others rather than being individually rational to protect oneself.33 Here, we specifically investigate support for non-pharmaceutical interventions, that is, government restrictions to stop infection spread as a direct measure of the acceptance of collective responsibility.

Third, interpersonal trust may be a key predictor of the willingness to contribute to collective action during the COVID-19 pandemic.33 Vaccination is a form of collective action, where herd immunity is produced via the collective participation in vaccination programmes,34 and people may be more likely to participate if they trust others to do the same.

Table 1 in the Methods section shows the specific operationalisation of each of these predictors and summarises how these predictors are related to the 5C model. As is evident, we do not include measures that reflect the calculation component of the 5C model. From a communication perspective, however, this component is less important as it refers not to the content of the individual’s considerations but to more stable individual differences in decision-making style (ie, extensive cost–benefit analyses of pros and cons of vaccination and infection).9

Table 1

Main measures in the study

Supplemental material