4

Assessing Global and Local Drivers of Vaccine Hesitancy

The second session of the workshop focused on assessing global and local drivers of vaccine hesitancy, with the objectives of (1) examining trends in hesitant attitudes toward vaccination and the effect on declining immunization rates; (2) evaluating the complex determinants and drivers of vaccine uptake, including sociocultural factors that influence perceptions, attitudes, and behaviors toward vaccination; and (3) exploring methods to monitor and measure vaccine hesitancy to better address concerns and to sustain confidence in vaccination. Stefan Flasche, associate professor at the London School of Hygiene & Tropical Medicine, described the burden of vaccine-attributable severe dengue in the Philippines and the effect of dengue vaccination campaigns on the national immunization program. Julie Leask, professor at the University of Sydney, Australia, discussed measurements of behavior and social drivers of vaccination. Julie Bettinger, associate professor at the Vaccine Evaluation Center at The University of British Columbia, Canada, explored the drivers and spectrum of vaccine hesitancy. Noel Brewer, professor at the University of North Carolina at Chapel Hill, presented on the Increasing Vaccination Model. He discussed propositions and next steps for changing vaccination behavior. The session was moderated by Alison Buttenheim from the University of Pennsylvania.

VACCINE-ATTRIBUTABLE SEVERE DENGUE IN THE PHILIPPINES AND THE IMPACT ON NATIONAL IMMUNIZATION PROGRAMS

Presented by Stefan Flasche, London School of Tropical Medicine & Hygiene

Flasche described the burden of vaccine-attributable severe dengue in the Philippines and the effect of mass vaccination with Dengvaxia—the first licensed dengue vaccine—on national immunization programs. He also made suggestions for approaching future coronavirus disease 2019 (COVID-19) vaccination efforts based on lessons learned from this case study in the Philippines.

Dengue and Dengvaxia

Dengue is a major global health priority, with 100–400 million annual infections worldwide.¹ Flasche said that many people with dengue require medical attention, with some developing severe symptoms including death (though this is relatively rare). The burden of dengue is growing rapidly due in part to the expanding range of the primary vector, the Aedes aegypti mosquito. Dengue has four serotypes. Typically, an infection with one serotype will trigger a pan-serotype immune response, providing short-lived immunity for about 6 months to 1 year; serotype-specific immunity can be longer lasting and even lifelong. The epidemiology of dengue is complicated by what is known as antibody-dependent enhancement of disease. That is, if a person has an initial infection with one serotype, a second infection with another serotype will be much more likely to result in severe disease owing to reactions among existing partial antibodies.

After decades of development, Dengvaxia (the first ever licensed dengue vaccine) showed efficacy against all serotypes in phase 3 clinical trials conducted with more than 30,000 participants across Latin America and Asia (Flasche et al., 2016). Flasche noted that the size of these trials was sufficient to address the primary outcome; all serotypes of clinical dengue were reduced by 50–70 percent among those who were vaccinated, including those with and without prior infection. During the follow-up phase in the third year of the trial, a potential safety signal was identified. In the control group of unvaccinated children aged 2–5 years, there was 1 case of hospitalized dengue while there were 15 cases among those vaccinated. Flasche noted that with a 2-to-1 randomization, this indicated that the risk

___________________

¹ More information about dengue is available at https://www.who.int/news-room/factsheets/detail/dengue-and-severe-dengue (accessed November 4, 2020).

of infection was seven-fold higher for vaccinated children than those in the control arm.

Researchers examined the potential origins of the infections observed in the trial. One hypothesis, Flasche explained, was that the increased susceptibility to severe disease was age related, because there was an age gradient to vaccine efficacy in which older cohorts displayed better protection. However, there could be a proxy for age—in this case, given the antibody-dependent enhancement of dengue, it could be that there was a higher proportion of dengue-naïve vaccinees in the intervention arm. Another hypothesis was that the vaccine was acting similarly to the natural infection. In unexposed individuals, it was possible that the vaccine was simulating an asymptomatic primary infection, thus increasing the likelihood of more severe secondary infection upon natural exposure. Yet the potential safety signal was only observed in children aged 2–5 years and was not observed during the trial in any participant aged 9 years or older. Flasche added that modeling suggested that even in dengue-naïve vaccine recipients, the lifetime net effect would be potentially positive despite the possibility of bringing forward risk (Ferguson et al., 2016).

Dengvaxia Uptake and Label Change

In spite of some uncertainty following the trials, Dengvaxia was licensed relatively quickly in about 20 countries for use in people aged 9–45 years, said Flasche. However, it was only widely used in parts of Brazil and the Philippines. He attributed this in part to a 2016 recommendation by the World Health Organization (WHO)² that restricted use of the vaccine to high-burden settings in which vaccinees averaged a 70 percent chance of having had dengue in the past. He suggested that only the Philippines and Brazil used the vaccine widely because it is difficult to calculate that risk percentage.

Sanofi, the manufacturer of Dengvaxia, announced a label change in November 2017. Additional studies enabled Sanofi to retroactively infer serostatus of vaccine recipients prior to vaccination. This provided a clearer picture of whether the potential risk was indeed age related or related to the serostatus of the vaccinees. Sanofi determined there was differential effect based on seropositive status; therefore, the label was changed to recommend the vaccine only for people who were seropositive due to previous dengue infection. However, more than 800,000 people in the Philippines and 300,000 people in Brazil had already been vaccinated with unknown serostatus, which Flasche asserted was a missed opportunity to better understand

___________________

² More information about WHO’s first dengue vaccine position paper is available at https://www.who.int/immunization/newsroom/press/dengue_first_position_paper/en (accessed November 4, 2020).

which vaccine recipients were likely to see substantial benefit and which would potentially see increased risk brought about by vaccination.

Uptake of Dengvaxia in the Philippines

Flasche and colleagues studied cohorts in parts of the Philippines in order to estimate the likely effect of Dengvaxia in the 5 years following widespread vaccination (Flasche et al., 2019). An extrapolation from the trials’ findings adapted to the Philippines cohort indicated dengue hospitalizations would decrease by an estimated 70 percent overall for the vaccinated cohort. Over the 5-year time period, they estimated that Dengvaxia vaccination would have averted (1) around 18 dengue hospitalizations among those who were seropositive at vaccination for each hospitalization among people who were dengue-naïve at vaccination and (2) about 10 cases of severe dengue among seropositive vaccine recipients for each case of severe dengue among those who were dengue-naïve at vaccination. Flasche said the balance of benefits and risks was therefore one-sided. Of the cases and hospitalized cases predicted to occur within 5 years of vaccination, an estimated 50 percent would be attributable to breakthrough infections among people who were seropositive and would therefore benefit from Dengvaxia vaccination. An additional 25 percent of cases would be those who were dengue-naïve at the time of vaccination, but who would have contracted the severe infection regardless. The remaining 25 percent would be vaccine-attributable with the risk brought forward.

The publicity around Sanofi’s label change caused a social media frenzy, said Flasche. Parents were understandably scared of the possibility that their children may have received a potentially harmful vaccine. Even more confusing, it was impossible to determine whether a particular child’s vaccine would prove to be harmful or of substantial benefit. This was a statistical problem, he noted, and humans do not typically think statistically when it comes to risk. Shortly after Sanofi’s announcement, the Philippines suspended the entire Dengvaxia program and banned the vaccine in December 2017. Flasche added they even brought criminal charges against trial administrators and Department of Health officers. He pointed out that in Brazil, however, the response was markedly different. Although there was the potential for a similarly negative public response in Brazil, the announcement largely went unnoticed and controversy around Dengvaxia only took place in the Philippines.

Impact on the National Immunization Program

The fallout from the Dengvaxia vaccination campaign had a substantial negative impact on the national immunization program in the Philippines,

which Flasche characterized as a “massive blow.” In the wake of the Sanofi announcement, vaccine confidence plummeted in the Philippines (Larson et al., 2019). In 2015, about 82 percent of the population reported feeling that vaccines were generally safe; this dropped to 21 percent by 2018. Additionally, vaccination rates dropped in the national immunization program and particularly in the childhood program. For example, it took many years to achieve a relatively high level (88 percent) of measles vaccination coverage in the country by 2014, but this decreased to roughly 50 percent in 2019, with measles incidence rates now between 10–20 times higher than they were before the Dengvaxia controversy (Dyer, 2019; Lancet Editorial, 2019). This contributed to widespread transmission of measles, with measles partially overwhelming the national health care system. Furthermore, polio had largely been under control in the Philippines, but a recent polio outbreak in the country is a concern for global polio eradication, which is one of the most expensive global health efforts.³

Applying Lessons Learned to COVID-19

Flasche said there are lessons from the Dengvaxia controversy that can be applied to COVID-19 vaccination. “When we start vaccinating, the story is not over,” he cautioned. He advised nations to prepare for the likelihood of intense public scrutiny and the influence of personal agendas, which can be both helpful and harmful to vaccination efforts. Moreover, he noted that COVID-19 vaccination will involve age groups who do not typically receive routine vaccinations and that the temporal association with unexplained deaths may also pose a challenge.

Open access to vaccine-related data and the decision-making process can help to enable successful vaccination efforts, said Flasche. Conflicts of interest should be avoided, because they can contribute to the type of social media frenzy that occurred with Dengvaxia. To help assess safety in real time, it is important to understand background rates in order to respond to safety events with confidence. For instance, if a cohort of people aged 55 years is vaccinated, some will experience heart attacks that are temporally associated with the vaccine unless the background rate has been established. Without background rates and a strong public understanding thereof, it is difficult to confidently determine that a vaccine is safe. Systems should also be in place to monitor the effect of the vaccine in real time to facilitate risk/benefit analysis, particularly given the likelihood of a suboptimal number of doses being available for the initial rollout.

___________________

³ More information about the polio outbreak in the Philippines is available at https://www.who.int/westernpacific/emergencies/polio-outbreak-in-the-philippines (accessed November 4, 2020).

MEASURING BEHAVIORAL AND SOCIAL DRIVERS OF VACCINATION

Presented by Julie Leask, University of Sydney

Julie Leask, professor at the University of Sydney, Australia, described efforts by the Behavioral and Social Drivers (BeSD) working group at WHO to develop measures of the social and behavioral drivers of vaccination. The BeSD working group was formed to develop globally standardized tools to measure core components of the Increasing Vaccination Model, including people’s thoughts and feelings, social processes, motivation, and practical issues regarding vaccination. She also explored the rationale for developing new tools to understand the vaccination gap of approximately 20 million children worldwide.

Vaccine Rates and Contributing Factors

The news media frequently focus on the role of the anti-vaccination movement in reducing vaccine acceptance, but Leask explained that barriers to high vaccination coverage extend beyond negative messaging about vaccination. A commonly held belief is that anti-vaccination messages lead to reduced acceptance, which leads to reduced coverage, which causes outbreaks. Leask noted that WHO listed hesitancy as one of the top 10 threats to global health in 2019;⁴ however, WHO also included fragile health systems and weak primary care in the list of top threats. These factors also influence vaccine uptake.

Examining global measles cases over the past 2–3 years demonstrates the complexity of factors that affect vaccine coverage, said Leask. Many of the countries with the highest numbers of cases have experienced mixtures of vaccine hesitancy and inadequate or disrupted health systems. Among such hot spots, Venezuela has had a major disruption to its health care system, Brazil has underserved regions, and the Ukraine has both vaccine hesitancy and inadequate health services. The United States has high measles vaccine coverage, but pockets of low coverage and multiple importations. Madagascar—which had the highest number of measles cases of any country during that period—has a weak health system, she noted. Political factors, such as the conflict and displacement of children in Yemen, have also contributed to measles outbreaks. Leask added that the Philippines had many years of inadequate immunization services prior to the compounding effects of the Dengvaxia vaccine safety event in recent years.

___________________

⁴ More information about WHO’s top 10 threats to global health in 2019 is available at https://www.who.int/news-room/spotlight/ten-threats-to-global-health-in-2019 (accessed November 4, 2020).

Measuring the Vaccination Gap

The stagnation of vaccination coverage rates is a major global concern, said Leask. Global coverage of diphtheria, tetanus, and pertussis vaccine (DTP) dose 3 has remained at 85–86 percent for several years, leaving almost 20 million children either unvaccinated or undervaccinated.⁵ Determining the cause of this gap is an empirical question that warrants better measurement strategies. Leask noted several challenges with respect to current measures. One issue with current measures relates to the focus on measuring attitudes—such as what people think and how they feel about vaccination—without also considering important practical and logistical barriers to vaccination that also influence uptake. Further issues include the fact that many vaccination gap measures are not validated, with insufficient standardization to enable examination of change over time. Additionally, Leask noted that the binary notion of supply and demand oversimplifies the ways in which barriers to uptake are conceptualized, and that findings from the data already collected often go unused.

New Tools to Measure and Address Behavioral and Social Drivers of Vaccination

Leask emphasized that new measures are needed to understand the vaccination gap and its causes. To that end, WHO and the Vaccination Demand Hub,⁶ a network of partner organizations, have initiated an effort to develop globally standardized tools to measure the BeSD of vaccination. This involves examining various types of vaccine-related data: coverage, program performance, behavioral and social, and surveillance. Coverage data measure the rates and geographic distributions of those experiencing zero-dose, delayed, or undervaccination scenarios. Program data involves vaccine supply, wastage, policies, and legislation. Behavioral and social data can be used to identify barriers and drivers of vaccination per population group. Surveillance data measure disease burden. The objective of this effort to develop new tools to measure and address BeSD is to boost the availability, quality, and usability of local and global data on acceptance and uptake by (1) supporting assessments of undervaccination to inform policy making and planning; (2) informing the design and evaluation of targeted interventions; (3) tracking comparable trends over time, such as any decline in vaccine confidence; and (4) contributing to regional and global reporting processes

___________________

⁵ More information about progress and challenges with achieving universal immunization coverage is available at https://www.who.int/immunization/monitoring_surveillance/whoimmuniz.pdf?ua=1 (accessed November 4, 2020).

⁶ More information about Vaccination Demand Hub is available at https://www.demandhub.org (accessed November 4, 2020).

such as the Immunization Agenda 2030 (IA2030), the WHO/United Nations Children’s Fund (UNICEF) Joint Reporting Form, and Gavi 5.0.

Leask outlined the three new types of tools developed under this effort. The Childhood Immunisation Survey is designed for parents and caregivers to provide their perspective on what might be contributing to their children’s vaccination status.⁷ In-depth BeSD interview guides were developed to dig deeper into the perspectives of caregivers as well as vaccine providers, community stakeholders, and authorities across health and immunization systems. Implementation guidance has also been developed to inform local adaptation, testing, and use of tools. Informed by design-thinking principles, the guidance considers the personae of vaccine program managers and field researchers and what they need to know to use these tools and make changes based on their findings. Leask noted that a challenge in developing global tools is the need to consider the multiple contexts in which a single set of measures will be used, and thus the need to develop flexible tools that will suit variable contexts of delivery, provision, income level, and access.

Increasing Vaccination Model

The BeSD working group developed the Increasing Vaccination Model, which is adapted from work led by Noel Brewer (Brewer et al., 2017a). This model stipulates that motivation to vaccinate is influenced by beliefs and feelings about vaccination as well as by social processes. Leask explained that people’s thoughts and feelings about vaccination include confidence in vaccine benefits, vaccine safety, and in the provider, as well as religious beliefs. Social processes include the social influences of the provider, family, and community, in addition to gender equity and the decision-making autonomy that a woman in the family has. These come together to influence motivation, which is the point at which hesitancy is located within this model. Leask noted that hesitancy is conceptualized as a motivation in this framework, rather than as a behavior. Practical issues are those that affect whether a person is vaccinated, including awareness, availability, ease of access, affordability, service quality, and provider–patient relationships. When a person is sufficiently motivated and practical issues are not sufficient barriers to prevent vaccination, the child receives recommended vaccines.

To illustrate how the Increasing Vaccination Model can be used to help create measurement tools, Leask described some of the survey items. To better understand the role of gender equity in caregiver motivation to vaccinate,

___________________

⁷ The long form of the Childhood Immunisation Survey has 22 items; the short form has 5 items.

the survey asks: “In your family, who has the final say about vaccinating your child?” Answer options include “mother,” “father,” “both parents,” “grandparents,” “other caregiver,” “not sure,” and “declined.” The in-depth interview guide asks: “Tell me about how you decided to vaccinate (or not vaccinate) your child(ren); who else was involved in the decision?” Leask noted that open-ended questions such as this are analyzed differently than the multiple-choice survey questions. To examine ease of access issues, the survey includes this question: “How easy is it to get vaccination services for your child?” The in-depth interview states, “Walk me through what you do on the day of vaccination; start at the very beginning.”

Field Testing the Increasing Vaccination Model

Leask said that these new tools are being field tested in countries like Sierra Leone, where the BeSD working group partnered with Statistics Sierra Leone to integrate the tools into preparation for a national survey.⁸ This involved cognitive interviews in which researchers studied how people interpret the survey questions to ensure that the questions are phrased in a way they can be understood as intended. This process includes asking participants a question, then asking them questions about the question. For example, participants are asked, “How important do you think vaccines are for your child’s health?” They can select from “not at all important,” “a little important,” “moderately important,” or “very important.” Next, they are asked these questions: “What were you thinking about in deciding your answer?” “What do the words ‘important for your child’s health’ mean to you?” “Did the response options fit with the sort of answer you wanted to give?” This type of cognitive interviewing has been performed in both rural and poor urban communities in Sierra Leone.

Leask outlined three phases of in-country field testing (efforts were in the second phase as of August 2020). The first phase includes consultation with WHO’s Expanded Programme on Immunization managers and partners, analysis of existing tools and literature, and expert review and consultation. The second phase includes development of training and testing materials, testing, and local adaptation. The third phase involves implementation of scaled pilots for psychometric validation and integration. Leask noted that COVID-19 has stalled in-country field testing, so testing will be undertaken by a company with data collectors in-country. The goal is to complete data collection by April 2021 and develop a data repository for reports from countries so that metrics and indicators can be tracked for IA2030 and

___________________

⁸ More information about Statistics Sierra Leone is available at https://www.statistics.sl (accessed November 4, 2020).

Gavi 5.0.^9,10 WHO’s capacity-building efforts will include gathering data for analysis and translation, as well as for building social and behavioral research capacity in the field of vaccination. WHO is also planning feedback loops for continuous learning from the implementation of the new tools to inform revision work in the future, she added.

UNDERSTANDING DRIVERS OF VACCINE HESITANCY

Presented by Julie Bettinger, The University of British Columbia

Bettinger explored drivers of vaccine acceptance, hesitancy, and refusal. She described the characteristics common to different groups along the spectrum of vaccine hesitancy to explore why interventions that are effective in moving certain people toward vaccine acceptance may not work for others.

Immunization Rates and Trends in Canada and the United States

According to a UNICEF analysis on immunization, Canada ranked 28th out of 29 developed countries for immunization coverage of measles, polio, and DTP dose 3 in children aged 12–23 months.¹¹ Bettinger noted that Canada’s coverage rate of 84 percent is similar to some low- and middle-income nations in the world. The United States ranked 23rd out of 29, but remained in the bottom third of rich countries with a coverage rate of 93 percent. The analysis also compared rates of immunization coverage by antigen for pneumococcal infection; varicella; measles, mumps, and rubella (MMR); Haemophilus influenzae type B; polio; DTP; rotavirus; and hepatitis B.¹² Canada hit the target 90 percent coverage rate recommended by WHO for two of eight antigens (MMR and polio), whereas the United States had four antigens covered at the target rate—varicella, MMR, polio, and hepatitis B (PHAC, 2016).¹³ She pointed out that in comparing jurisdictions and coverage rates, data are often collected differently, resulting in “apples and oranges” scenarios in which true, parallel comparisons cannot be derived.

___________________

⁹ More information about IA2030 is available at https://www.who.int/immunization/IA2030_draft_4_WHA.pdf?ua=1 (accessed November 4, 2020).

¹⁰ More information about Gavi 5.0 is available at https://www.gavi.org/our-alliance/strategy/phase-5-2021-2025 (accessed November 4, 2020).

¹¹ More information about the Innocenti Report Card is available at http://www.unicef-irc.org/publications/pdf/rc11_eng.pdf (accessed November 5, 2020).

¹² More data from the National Center for Health Statistics are available at https://www.cdc.gov/nchs/hus/contents2017.htm (tables 012, 066, and 067) (accessed December 4, 2020).

¹³ Results from the 2017 Childhood National Immunization Coverage Survey conducted by Health Canada can be found at https://www.canada.ca/en/public-health/services/publications/healthy-living/2017-vaccine-uptake-canadian-children-survey.html (accessed February 25, 2021).

For example, these data from Canada were taken for children at 2 years of age, whereas the U.S. data include children from 19–34 months of age. Bettinger suggested that the Canadian data might be more similar to U.S. data if the same age range were used. This illustrates how lack of uniform data measurements can make it challenging to identify areas of low vaccine coverage on a global scale.

Bettinger reported that the United States and Canada had similar rates of influenza vaccination from 2018–2019 (PHAC, 2019).^14,15 In Canada, 42 percent of adults were vaccinated, compared to 45 percent in the United States. In both countries, more females were vaccinated than males and around 70 percent of older adults were vaccinated. She added that rates varied not only by age group but also by jurisdiction. For example, 34 percent of adults were vaccinated in Nevada versus 56 percent of adults in Rhode Island. This demonstrates that national estimates can be helpful in global comparisons, but examining data by age, jurisdiction, and gender is also valuable.

Estimates of non-vaccination rates also vary by location, said Bettinger. In Canada, the percentage of children who never received any vaccines across all age groups is estimated to be 1.5 percent nationwide, but provincial estimates range from 3–5 percent (Wilson et al., 2015). Rates vary further by age group, the community within a province, and by vaccine. For instance, in the province of Alberta, non-vaccination rates for DTP are 5.5 percent, compared to 15 percent for varicella. Even if a province’s coverage rate is much higher than the national average, community rates may be much lower, she added. Some communities in British Columbia have antigen-specific coverage rates lower than 50 percent, particularly for the human papillomavirus (HPV) vaccine. In the United States, the 2017 estimate for non-vaccinated children aged 3 years and under was 1.1 percent—which increased from 0.7 percent in 2013—but the estimated rate for uninsured children was much higher, at 7 percent (Mellerson et al., 2018). Bettinger pointed out one vaccine trend that is fairly consistent across Canada: a recent increase in exemptions. Ontario (which is one of only two provinces to mandate vaccination upon school entry) has had an increase in nonmedical exemptions from 0.4 percent in 1985 to 2.5 percent in 2005; this trend has continued to increase since then, she added. In the United States, the 2017 exemption rate for kin-

___________________

¹⁴ Results from the 2018–2019 Seasonal Influenza Vaccination Coverage Survey conducted by Health Canada are available at https://www.canada.ca/en/public-health/services/publications/healthy-living/2018-2019-influenza-flu-vaccine-coverage-survey-results.html (accessed February 25, 2021).

¹⁵ More information about the 2018–2019 influenza season vaccination coverage in the United States is available at https://www.cdc.gov/flu/fluvaxview/1819season.htm (accessed November 5, 2020).

dergarteners was 2.2 percent, which varied by jurisdiction (e.g., 0.1 percent in Mississippi versus 7.1 percent in Alaska).

Continuum of Vaccine Acceptance Model

Bettinger emphasized that vaccine coverage, non-coverage, and exemption rates are important, but “they do not tell the whole story.” These data allow experts to focus attention on particular vaccines, age groups, jurisdictions, and other demographics (e.g., the uninsured) to understand reasons for low vaccine coverage, but the data do not directly identify vaccine-hesitant individuals. She cautioned against conflating low vaccine coverage with vaccine hesitancy; the latter is just one of multiple factors that affect coverage rates. As evidenced in “anti-vaxxer” stereotypes, the media tends to propagate a certain image of vaccine-hesitant people as relatively privileged and philosophically opposed to vaccinating their children. However, vaccine hesitancy is actually a motivational state that a person might experience at different stages in life; it can also vary over time, place, and jurisdiction. For example, a woman who strongly believes in vaccines and is fully vaccinated might be hesitant toward an influenza immunization while she is pregnant.

Bettinger explained how the Continuum of Vaccine Acceptance Model developed by WHO’s SAGE Working Group on Vaccine Hesitancy and other academic researchers accounts for shifts in an individual’s vaccine hesitancy over time (Benin et al., 2006; Dubé et al., 2016; Leask et al., 2012; MacDonald et al., 2015; WHO, 2014).¹⁶ Generally, people fall into one of five categories along the spectrum of vaccine hesitancy:

1. those with no doubts or concerns,
2. those with minor doubts and concerns,
3. those with many doubts and concerns,
4. those who are late or selective vaccinators, and
5. those who refuse all vaccinations.

People with no doubts or concerns are individuals who have already made a decision about vaccination, so they do not qualify as hesitant. Likewise, people refusing all vaccinations have also already made a decision and are therefore not considered hesitant. It is the middle groups on the spectrum that are most aptly described as “vaccine hesitant” or in a “state of hesitancy.” Furthermore, even though individuals with minor or many doubts and concerns are considered hesitant, they typically still choose to receive all vaccinations. When looking at coverage rates, even those in the

___________________

¹⁶ Bettinger emphasized that while her focus in this presentation is on vaccine hesitancy, it is only one aspect of vaccine acceptance, and other factors play a role.

“many doubts and concerns” group tend to be among the 84 percent of Canadians and 93 percent of Americans who are vaccinated. However, there are individuals within that group who may move up or down the spectrum of vaccine acceptance.

The first group (those with “no doubts or concerns”) is composed of people who have accepted vaccines. Bettinger explained that these individuals typically consider vaccines safe and important. Often, they have a lot of trust in their health care provider and receive a strong recommendation from this provider. They are often heavily influenced by the social norm of vaccination—that is, their friends and families are vaccinated, and a positive vaccination message is reinforced in their community and among their contacts. The groups with “minor doubts and concerns” and “many doubts and concerns” have vaccine hesitancy. They tend to be more focused on vaccine risk and often have low perceived benefits of vaccination and low perceived risk for vaccine-preventable disease. They may not think they need to vaccinate and do not see the risk of not vaccinating, she added, noting that these groups can easily move up or down the continuum due to a variety of influences. For example, a health care provider they trust who provides a strong recommendation can influence them toward vaccinating, whereas a vaccine safety scare can shift them toward a decision not to vaccinate.

The fourth group, those who “vaccinate selectively and late,” are individuals who have significant doubts about the safety and necessity of vaccines, said Bettinger. They tend to actively seek information and may have conflicting feelings about whom to trust. People in this group may initially meet with doctors without bringing their children. Additionally, they may make multiple appointments at different clinics in order to weigh information from various sources before making vaccination decisions. This group also tends to be heavily influenced by social networks. However, unlike those with no doubts or concerns about vaccination, their social network may not be vaccinating. Friends or relatives can be a contributing factor to hesitancy when they tell an individual that vaccines are unsafe. Many of these individuals have had negative experiences with health care providers, the medical system, or vaccines. She noted that in some cases, these negative experiences have nothing to do with vaccines but were powerful enough to make people distrustful of any advice from health care providers.

The fifth group, those who “refuse all vaccinations,” are sometimes termed anti-vaxxers. Bettinger said that this group should not be considered as vaccine hesitant because they have made their decision not to vaccinate, often with strong conviction. Sometimes this is attributable to a “natural” attitude toward health, but she noted that strong religious or moral considerations drive such decisions just as often. Distrust in the medical community is also common in this group. She added that while it can be very challenging to move this group toward vaccine acceptance, it is not impossible.

However, the intervention that may work with someone with minor doubts and concerns will not necessarily work for someone who has decided not to vaccinate and has strong motivation behind that decision. For example, a presumptive communication approach in which a doctor says “Today you are going to be vaccinated” may be effective for someone who does not have any doubts or concerns around vaccination, but for someone who is a late or selective vaccinator, that approach could reduce trust and make the person even less likely to vaccinate. Bettinger added that people in this group often seek out a provider who will listen to them and discuss their concerns, so health providers should consider where patients are along the hesitancy spectrum in deciding what type of intervention to use.

THE INCREASING VACCINATION MODEL

Presented by Noel Brewer, University of North Carolina at Chapel Hill

Brewer provided more detail about the Increasing Vaccination Model and its propositions about how to influence vaccine behavior. He described the relative effect of different types of interventions based on those propositions and suggested ways forward to apply the model toward influencing vaccination behavior. In collaboration with Julie Leask, Gretchen Chapman, Alex Rothman, and Allie Kempe, Brewer developed the Increasing Vaccination Model to help address low vaccination uptake (Brewer et al., 2017a). He noted that uptake is not the only factor affecting coverage rates—delayed vaccination and instability in vaccination also play a role. The premise of the model is that certain factors influence individuals to schedule, consent to, delay, or refuse vaccinations. The model groups the factors that motivate vaccination into three categories: what people think and feel, social processes, and direct behavior change.

Proposition 1: Thoughts and Feelings Influence Vaccination Behavior

The first proposition of the model is that people’s thoughts and feelings influence their vaccination behavior, said Brewer. Thoughts and feelings can encompass disease risk appraisal, vaccine confidence, and motivation. For example, an individual may express risk appraisal by thinking, “I am concerned about getting pneumonia.” That person’s vaccine confidence might be that the pneumonia vaccine is effective and safe. The risk appraisal and vaccine confidence would inform the individual’s motivation—in this case, the intention to receive the pneumonia vaccine—that leads the person to vaccinate. In cases where disease risk appraisal and vaccine confidence lead to vaccination hesitancy rather than motivation, the result may be vaccination refusal or delay.

Although non-experimental and correlational studies provide support for this proposition, available evidence from randomized trials does not. Presently there is little evidence to suggest interventions that target people’s thoughts and feelings affect vaccination behavior (Brewer et al., 2017a). For example, the following interventions were found to have minimal or no likely effect on participants: messages that increase disease risk appraisals, education campaigns that increase vaccination confidence, decision aids, and motivational interviewing. Brewer did note that there is some evidence to suggest messaging for the purpose of increasing disease risk appraisal may have some effect, but a recent meta-analysis did not find this to be an effective method for increasing vaccine uptake (Parsons et al., 2018). He added that, owing to a lack of reliable study data, the effectiveness of educational campaigns to increase vaccine confidence is unclear. Interventions to increase confidence may eventually be shown to indicate increased vaccine uptake, but so far no randomized controlled trials have been conducted. As a result, evidence that confidence is a vaccination driver is lacking. Brewer maintained that decision aids are not effective in increasing uptake. Motivational interviewing has some promise, but there is still no randomized controlled trial evidence available to support it.

Proposition 2: Social Processes Influence Vaccination Behavior

The second proposition of the Increasing Vaccination Model is that social processes influence vaccination. Brewer explained that social processes begin with an individual’s virtual or in-person social network (e.g., family, friends, colleagues, neighbors). Homophily, or the tendency to seek out people similar to oneself, often shapes individuals’ social networks. Ideas propagate through one’s social network and generate social norms in people’s minds that can guide vaccination behavior. Social processes also include social preferences. For example, a social preference for altruism can motivate a person to vaccinate in order to help other people avoid becoming ill. Alternatively, a person’s social preference for “free-riding”—in this case, assuming that everyone else is going to get vaccinated—can be associated with vaccine refusal.

Support for the proposition that social processes are central to vaccination behavior comes from both correlational and experimental studies, but there is little evidence from randomized controlled trials (Brewer et al., 2017a). Although the available evidence suggests that there is minimal or no likely effect of messaging designed to change altruism or free-riding beliefs, other types of interventions focused on social process interventions seem to have a modest effect on vaccination behavior. Descriptive norm messages are promising and have been effective in other areas, but strong evidence is not yet available for their use in vaccination efforts. Similarly, Brewer noted

that social network interventions that build on contagion have worked well in other areas and are promising in the area of vaccination, but trials have not yet been published.

Proposition 3: Direct Behavior Change Influences Vaccination Behavior

The third proposition is that vaccination uptake can be affected by changing behavior directly, rather than changing how people think and feel. Brewer explained that this involves building on favorable intentions without attempting to change underlying thoughts and feelings. To build on favorable intentions, the use of reminders, prompts, and primers can help keep vaccination on people’s minds. Another useful strategy is to reduce barriers with logistics or behavioral defaults. For people who do not have favorable intentions toward vaccination, behavior can be shaped by offering incentives, implementing sanctions, and perhaps even requiring vaccinations by mandate.

Clear evidence from intervention studies supports the proposition of changing behavior directly, said Brewer (Brewer et al., 2017a). All types of behavior-change interventions indicate a modest or substantial likely effect on vaccination behavior. For instance, presumptive health care provider recommendations, onsite vaccination, default appointments, incentives, and vaccination requirements all show a substantial likely impact. Reminders and callbacks showed a modest effect, which Brewer surmised would be greater if provided consistently with appropriate follow-through in primary care. The available data also suggest that these reminders are particularly effective when they are centralized and identified as “clinical centralized reminders [callbacks]” (Kempe et al., 2015). Brewer added that the other direct behavior-change efforts showing modest likely effect on vaccine uptake were implementation intention interventions and mere measurement interventions. Onsite vaccination, default appointments, incentives, sanctions, and vaccination requirements were all found to have a substantial effect.

Presumptive Recommendations

To test the hypothesis that presumptive recommendations—which use language that assumes parents are ready to vaccinate when introducing the vaccine—would increase uptake, Brewer and colleagues trained providers in 30 clinics in North Carolina and found that presumptive recommendations were associated with increased vaccination and reduced time to vaccinate (Brewer et al., 2017b; Opel et al., 2013). In the intervention arm, the physicians started appointments by announcing that the child was due for an HPV vaccine and then pivoted to a more consultative interaction if the reaction

warranted. This resulted in a 5 percent increase in vaccine uptake. By contrast, when physicians began appointments with a consultative interaction, there was no statistically significant increase in vaccine uptake.

Another direct behavior-change intervention study on default vaccination appointments found that people who are automatically scheduled for an appointment for a seasonal flu vaccine (without first requesting it) are more likely to come in for the vaccination (Chapman et al., 2016). This opt-out method resulted in 27 percent of patients being vaccinated. When an opt-in method was used, the vaccination rate dropped to 18 percent, which was only slightly higher than the control group who received no letter regarding vaccination (17 percent). Brewer emphasized that this study used a registry and medical records to account for vaccinations provided at both doctors’ offices and at flu clinics, so these figures do not represent a displacement of vaccinations from one location to another, but rather an actual increase in overall uptake.

In a study on incentives used in India, researchers randomly assigned villages to a control group, a monthly vaccination camp, or a monthly vaccination camp with incentives (Banerjee et al., 2010). The incentives provided were a kilogram of lentils per shot, worth approximately 75 percent of 1 day’s wage, as well as a set of metal thali plates (food serving platters) at completion. The study found that onsite vaccination increased uptake. At the end of the intervention, 6 percent of 18-month-olds in the control group were fully vaccinated, compared with 18 percent in the monthly vaccination camp; the rate increased to 39 percent in the villages that provided monthly vaccination camps with incentives.

Ways to Apply the Model to Influence Vaccination Behavior

Based on his review of this data, Brewer said that the proposition that people’s thoughts and feelings affect vaccination behavior is not as promising in trying to improve vaccine uptake. Social processes are a promising area, but they are currently understudied and poorly understood with respect to vaccination, underscoring the need for additional funding and research in this area. Direct behavior change has been shown to increase uptake, but those approaches remain underutilized. Brewer noted that most of the data he presented come from high-income countries (e.g., Australia, the United States, countries in Western Europe), highlighting the need for more data from low- and middle-income settings, which have strengths that could be brought to countries of all income levels.

Brewer highlighted that policies and programs utilizing direct behavior change are effective. He hypothesized that, despite the current lack of data, the underlying reasons why such direct behavior-change initiatives work are confidence in vaccination and the vaccination system, as well as trust that

vaccinations are wholesome and will benefit both children’s health and the health of the general population. He posited that this level of confidence may be a prerequisite for effectively increasing vaccine uptake. If this is the case, it would indicate that confidence—something that people think and feel—may have a role in the efficacy of direct behavior change. Thus, building confidence could help if used as a tool to encourage the public and policy makers to engage in the programs, projects, and policies that are effective in increasing vaccination. He noted that strong policies are needed to overcome some of the recent instability in vaccine coverage. For instance, HPV vaccine coverage is low in the United States and is an area that warrants urgent intervention. Seasonal influenza vaccination is another area that warrants immediate focus during the COVID-19 pandemic, because (1) seasonal influenza vaccination in 2020–2021 can be used as a model for delivering the COVID-19 vaccine, and (2) controlling seasonal influenza rates may help avoid dual global pandemics. He suggested that when a vaccine for COVID-19 becomes available, tools to increase vaccine uptake should be used within programs built on strong logistics and based on concretely effective strategies, rather than focusing exclusively on media campaigns.

DISCUSSION

Prioritizing Vaccination Behavior Driver Types

Vaccine promotion often focuses on educating, convincing, and persuading people to vaccinate their children and themselves, said Buttenheim. She noted that Brewer’s presentation deemphasized those types of interventions because of a lack of evidence of effectiveness, and instead he prioritized behavioral interventions. Furthermore, he hypothesized that confidence may be a required precedent for behavior change. Buttenheim asked panelists to discuss the tension or dichotomy often indicated between a focus on education or persuasion and a focus on behavior changes. Brewer said that the strongest motivator of vaccination is a provider recommendation, be it from a physician, a nurse, or anyone on the primary care team. He noted that a provider recommendation is a powerful tool for increasing uptake, but the reason for its effectiveness is not yet understood, although the availability of the vaccine and the social contract could be at play. Thus, while provider persuasion can work, it is not yet known which process that persuasion is operating through, he said.

Leask speculated that regardless of the level of vaccine confidence and motivation, the greatest barriers to vaccination may be at the population level around practical issues. This could account for why system-level changes are effective. She suggested that interventions around thinking and feeling that help to communicate to the public about vaccination programs

are important components of the logic model of causal change through which vaccination programs can become successful. She added that the confidence of a politician, leader, or program manager is among the different layers needed for vaccination programs to be successful. Although the ways people think and feel about vaccines are an important part of the mechanism, and communicating about vaccination is important, she contended that such interventions alone are not sufficient to substantially improve vaccine coverage.

Bettinger agreed that a focus on structural interventions such as mandates and incentives (which are more common in the United States and Australia than in Canada) can be effective by sending a clear message that vaccinations are considered important at a structural level. Furthermore, this has the effect of encouraging individuals who may not vaccinate simply because of inconvenience. In addition to these behavior-change interventions, she emphasized that the social process of altruism can impact vaccination behavior. It is difficult to quantify concepts like altruism, but Bettinger has performed qualitative research with vaccine-hesitant parents who were moved toward vaccinating. Such parents often question why vaccination matters and make statements such as: “It’s my decision,” “Why do you care?” and “It’s my child.” However, in Bettinger’s experience, some vaccine-hesitant parents were motivated to vaccinate by learning more about herd immunity and how their vaccination decision actually has an effect on their community. She added that in British Columbia, there are some religious communities that will not vaccinate but are willing to participate in other types of control and containment measures. This demonstrates the importance of understanding the context in tailoring work with a community or individual.

Flasche emphasized that in the unprecedented challenge of the COVID-19 pandemic, experts should be aware that the factors involved in vaccine hesitancy, delay, and refusal may shuffle as the global pandemic impacts more and more aspects of society. He predicted that there may be a range of responses from extreme vaccine demand in communities wanting to end social distancing as quickly as possible to skepticism about the need for a vaccine from parts of the world that have not been as heavily impacted by COVID-19. Therefore, the relative contribution of factors driving vaccination behavior may be different in this context compared to past situations. Furthermore, Flasche added that the conditions of the pandemic will be different between and within various parts of the world.

Brewer said that organizations will be remiss if they choose not to be a part of the conversation on social media. Current communication around COVID-19 vaccination needs improvement, especially the messaging coming from agencies and policy makers. So little is currently known about the virus that the public naturally wants to fill in the gaps in that knowledge.

He cautioned against framing vaccines as “new” because of the unprecedented speed of development. Instead, he suggested identifying a single spokesperson, or multiple spokespeople representing diverse communities, to communicate regularly with the public about what is known about the vaccines. That way, Brewer explained, when COVID-19 vaccines eventually arrive, the public will be familiar with them as a guaranteed eventuality. Just as people have grown accustomed to the reality of COVID-19, they can grow accustomed to the idea of a vaccine.

Recommendations for COVID-19 Vaccine Acceptance

Buttenheim asked about first-priority strategies to increase acceptance of the upcoming COVID-19 vaccine. Flasche reiterated that a provider recommendation is potentially the biggest factor influencing a person’s likelihood of getting vaccinated; therefore, strong recommendations from providers could have the largest possible impact. He added that including clear and digestible information on potential benefits and risks, as well as a quick overview of what is known and yet to be known about the vaccine, could be included with the recommendation. Leask said that based on vaccination conversation research, presumptive communication could be influential (Randall et al., 2020). If a patient is eligible for a vaccine that is available with demonstrated safety and effectiveness, Leask suggested that the provider could say: “You are eligible for the COVID-19 vaccine. I think it would be good if you had it. Do you have some questions?” Ideally, the patient would have received some information before the visit, even in the doctor’s waiting room. Leask noted that this approach to the vaccination conversation includes presumptive communication as well as space for the patient to ask questions, simultaneously honoring the need for valid consent. She continued that because the vaccine is new, people will likely have questions, but the conversation can still be framed around a recommendation. If a patient is particularly hesitant, a slightly different pathway could be used. She added that in her leadership role with the Sharing Knowledge About Immunisation project, she developed a primary care provider communication package for childhood vaccination that includes pathways aligned with differing positions on vaccination and immunization knowledge-sharing tools.¹⁷ Brewer said that he and his colleagues have developed the “announcement approach,” in which the medical care provider announces that the child is due for vaccination.¹⁸ He said

___________________

¹⁷ More information about the Sharing Knowledge About Immunisation project can be found at https://www.ncirs.org.au/our-work/sharing-knowledge-about-immunisation (accessed February 25, 2021).

¹⁸ More information about the announcement approach is available at www.hpviq.org (accessed November 5, 2020).

the key is to refer to vaccines in a similar manner to all of the other services offered in the clinic or medical office. If parents have questions, the process slows down to meet their needs. He suggested first identifying the parent’s main concern so it can be addressed.

Targeting Specific Populations

On the topic of tailoring vaccination programs to target specific populations, Buttenheim asked about data on coverage rates for specific racial and ethnic groups or homeless populations. Brewer replied that in the United States, the federally funded Vaccines for Children Program pays for approximately half of all vaccines for children aged 18 years and older.¹⁹ Although Vaccines for Children charges an administration fee of up to approximately $17, all vaccines are free and families are not charged this fee if they cannot afford it. He said there are some options for families that are homeless or without a permanent place to live, the most widely available being a visit to the local health department. Understanding the logistical barriers to homeless families’ ability to get free and low-cost vaccines is important, Brewer said. He acknowledged that vaccine access for homeless adults is more complicated and stated that he is less familiar with that population.

Brewer noted the complexities involved in understanding why different ethnic groups have higher or lower vaccine coverage. For instance, African American adults are less likely to get the influenza vaccine than white adults, yet African American families are more likely to get the first dose of HPV vaccine for their children than their white counterparts. Brewer said the mechanisms behind this are not clear. Emphasizing that rates of the HPV and flu vaccine in the United States are low, Brewer highlighted the need for coverage rates to increase across all groups of Americans. He suggested that lower interest among African American people in getting the COVID-19 vaccine may be a durable problem, as is uptake of the seasonal influenza vaccine. If programs can work now to address problems with uptake of the seasonal influenza vaccine, it could be used as a model for encouraging uptake of COVID-19 vaccines when they become available.

Bettinger highlighted the need to better understand racial inequities in the United States. Some of the disparity is likely related to social determinants of health, but other factors are likely involved that are not yet understood. She noted that Canada has similar health disparities within its aboriginal population that cannot be completely explained by health inequities. In Canada, a history of racism and oppression toward the aboriginal population carries into the present day. It influences how various aboriginal

___________________

¹⁹ More information about the Vaccines for Children Program is available at https://www.cdc.gov/vaccines/programs/vfc/index.html (accessed November 5, 2020).

communities view Western medicine and health care providers, as well as the level of trust they have in allowing health care interventions to come into their communities. She added that local community contexts should be explored to understand whether trust issues, access issues, or other types of issues are contributing to low vaccination rates.

Leask said that there was once a vaccination gap between Indigenous and non-Indigenous populations in Australia, but targeted efforts have successfully eliminated this gap in certain age groups, with coverage rates for 5-year-old children in Indigenous communities now higher than the rate for non-Indigenous 5-year-old children.²⁰ She noted that good recordkeeping is partly responsible—Australia has a national registry that enables vaccination data for every child to be tracked. Culturally appropriate, respectful services administered through reminder systems also increased vaccination rates. For example, in a region of New South Wales, two Aboriginal health workers were employed to contact every single parent of a newborn identified as Aboriginal or Torres Strait Islander, obtain their mobile phone numbers, and send them a short message service (SMS) just before their babies’ vaccinations were due (Cashman et al., 2016). Leask said this initiative closed the immunization gap for 12-month-olds in this district and inspired the state government to employ an Aboriginal immunization coordinator in each local health district. Ensuring that Aboriginal people are employed to support immunization programs and influence their communities is key, she added. Australia closed the gap in Aboriginal children’s immunization rates in New South Wales partly through that program and partly from broad, upstream legislative changes that have resulted in more children getting vaccinated on time all over Australia. Leask concluded that the success stories in Australia can be attributed to ensuring cultural respect, understanding barriers, and comprehensively addressing those barriers.

Brewer also suggested that in designing systems to deliver vaccines, context should be at the forefront to avoid overemphasis on hesitancy. Furthermore, instead of assuming that people are disinterested or unwilling to vaccinate, focus ought to be shifted toward understanding how certain groups of people are systematically excluded from access. For instance, the reasons why rural areas of the United States tend to have lower vaccination rates are unclear, which warrants investigation into how people access vaccinations in rural health systems. With respect to people experiencing homelessness, Leask suggested working to increase vaccine uptake by leveraging

___________________

²⁰ More information on vaccine coverage in Australian Aboriginal and Torres Strait Islander peoples is available at https://www.health.gov.au/health-topics/immunisation/childhoodimmunisation-coverage/immunisation-coverage-rates-for-aboriginal-and-torres-straitislander-children (accessed May 20, 2021) and https://www.health.gov.au/health-topics/immunisation/childhood-immunisation-coverage/immunisation-coverage-rates-for-all-children (accessed December 18, 2020).

existing outreach services on the ground to bring vaccines to the places those populations are staying.

Addressing Vaccination Misinformation

Buttenheim asked about strategies that laypeople can use to combat misinformation when talking to others who are vaccine hesitant. Referencing an article she wrote on this topic called “Four Ways to Talk to Vaccine Skeptics,”²¹ Leask outlined several options, beginning with choosing one’s battles (Leask and Steffens, 2019). “If somebody is dead-set against vaccines, you may be wasting your time,” she said. When dealing with a family member or close friend, for example, she suggested having a respectful conversation in which both parties can share their concerns and position. Sometimes it will be necessary to agree to disagree, but if someone has had a particularly negative vaccination experience, they should be listened to and potentially referred to specialist immunization clinicians who can listen to their story and help them sort through what happened. However, Leask noted this approach can feel like “procedural justice” for the person who had the negative experience. For people who are only slightly hesitant about vaccines, it may be worth investing some time with them. Evidence suggests that providing information to correct misguided thinking can be more effective than doing nothing at all. However, she pointed out that this can backfire if confirmation bias affects the way a person with strong beliefs around vaccination receives a message.

Brewer suggested “taking the long view.” First, he said that it is the job of doctors to talk to people about vaccination; it is not the layperson’s job to convince people they know to vaccinate, and they should not be expected to have expert knowledge. For those who want to have the conversation, he recommended following the advice in Leask’s article. He added that many comments on social media are not actually made by people; they are generated by bots that can be ignored and blocked. He added that when addressing vaccine hesitancy or misinformation, it is helpful to identify one’s role (e.g., health care provider, friend, or someone speaking to the public who is trying to protect oneself in the process). For public figures who are addressing vaccine misinformation, Brewer recommended guidance issued by WHO titled “How to Respond to Vocal Vaccine Deniers in Public,”²² which provides two effective options: (1) debunking myths by identifying misinformation

___________________

²¹ Read the article on The Conversation at https://theconversation.com/4-ways-to-talk-with-vaccine-skeptics-125142 (accessed February 25, 2021).

²² WHO’s guidance on how to respond to vocal vaccine deniers in public is available at https://www.who.int/immunization/sage/meetings/2016/october/8_Best-practice-guidancerespond-vocal-vaccine-deniers-public.pdf (accessed November 10, 2020).

point by point and (2) identifying the person’s conversational ploys, such as “moving the goal post” or denying evidence. Buttenheim added that these two approaches are called “inoculation theory,” a communications strategy in which a small dose of misinformation, identified as such, is sandwiched between the correct information. She discussed another resource for people engaging on social media around vaccinations: a group called Shots Heard Round the World.²³ Formed by two American pediatricians, this group has been effective at countering social media attacks on health care providers and other immunization supporters. She added that lending support on social media is a useful way to counteract some of the misinformation activity on those platforms.

Influenza and COVID-19 Vaccination

Buttenheim asked whether 2020 seasonal influenza vaccination programs should be tailored to anticipate the release of the COVID-19 vaccine. She queried further if the anticipated public interest in a COVID-19 vaccine might be used to promote and boost seasonal influenza vaccine coverage. Bettinger replied that she is using the fall seasonal influenza vaccine program as a trial run for COVID-19 vaccines. Similar populations will be targeted for these vaccines, and there are many pandemic-related logistical issues to address for both. Measures such as social distancing need to be in place for influenza vaccine delivery, she noted, and communication strategies developed for the influenza vaccine can be applied to the forthcoming COVID-19 vaccine. If enough educational information is available, then one-on-one communication can take place between providers and patients. Buttenheim asked Bettinger which factors have had the greatest effect on coverage rates—hesitancy and motivation challenges or logistics, hassle, procrastination, and challenges related to behavior change. Bettinger responded that both are substantial challenges, but the extent to which each contributes separately to vaccine coverage issues is unclear. An additional challenge in Canada is that the seasonal influenza vaccine is not universally covered in all provinces and territories, which further complicates efforts to measure the drivers of low coverage.

Brewer pointed out that if families do not feel safe getting a flu vaccine, it will pose a barrier to the delivery of a COVID-19 vaccine. He emphasized that vaccination must be put “back on track” after the COVID-19 pandemic. While childhood doses for children aged 0–6 years are largely at pre-pandemic levels, adolescent vaccine rates continue to lag substantially. He surmised that adult vaccination rates are also down. The fall seasonal

___________________

²³ More information about Shots Heard Round the World is available at https://www.shotsheard.org (accessed November 5, 2020).

influenza vaccination season could provide an opportunity to help restore vaccinations to typical levels, but this will be contingent on parents and families trusting health care providers to provide a safe experience to reduce the risk of contracting COVID-19. He noted that not all providers are clearly communicating the safety measures they are taking and why it is safe to get vaccines from them. Having spent years working on improving seasonal influenza vaccination rates in children and adults, Leask said she hopes that a COVID-19 vaccine will lead to lasting improvement in seasonal influenza vaccination rates. She predicted that communities and governments will likely be motivated to have high vaccination coverage in order to return to pre-pandemic life. Flasche pointed out that in many parts of the world, especially in areas where COVID-19 has had a large impact, there is no seasonal influenza vaccination program. In such places the delivery of the COVID-19 vaccine could be useful in creating a mechanism for improving adult vaccination rates over the longer term, but this will require planning, he added.

Policy Interventions

Buttenheim asked for suggestions about policy interventions to promote vaccine coverage. Leask said that federal governments can enact policies to ensure that vaccination systems are well supported financially and structurally. She noted that mandates can be effective and are perceived as major policy levers for federal or state governments, citing an article by Saad Omer and colleagues on COVID-19 vaccine mandates and when they may or may not be appropriate (Mello et al., 2020; Omer et al., 2019). Additionally, policy makers can create national registries. For example, in Australia, extending the national registry for vaccination of children to a lifelong registry equipped vaccine providers with more data. Consistently logging vaccines each time they are administered to adults enables health care providers to know whether the patient they are seeing is fully vaccinated or not, and it provides a mechanism for reminding them. Therefore, Leask asserted that registries and multifactorial programs to improve coverage, including support and perhaps even incentives for providers, are potential policy levers.

Brewer highlighted several features of effective vaccination programs that ought to pertain to a vaccination campaign for COVID-19: vaccines should be free, safe, available, and easy. Vaccines should be entirely free, because even a small copay is a disincentive for people to act. Secondly, the settings where people receive vaccines must be made completely safe, with transparent social distancing and other safety measures so people do not fear exposure while receiving the vaccine. The third aspect is availability; people need to understand how and when they can access the vaccine once it becomes available. Furthermore, Brewer emphasized that vaccination must be made easy to access through a system that is as user-friendly as possible.

The capacity of primary care may be overwhelmed by handling vaccination for an entire country, so he suggested that pediatric vaccines be provided through pediatricians, then additional options for adults should be explored. He also suggested using the existing infrastructure for vaccines that has already been developed by CDC.

Bettinger described the mandate lever as a “very blunt instrument.” Although they can be effective, mandates can also cause backlash. She suggested that prior to deploying the mandate option, efforts should be made to strengthen immunization systems and offer compensation for providers to adequately deliver vaccines. Brewer cited an article that provides guidance about when and under what conditions to require vaccines (Omer et al., 2019). He contended that many factors would need to be in place before a mandate is instituted, including vaccine availability and measures to ensure equity. Brewer stated that it could be “disastrous” to mandate vaccination in the United States before first setting up the proper infrastructure.

Reflections on Session 2

The second session of the workshop concluded with reflections from Matthew Zahn, medical director at the Orange County Health Care Agency’s Division of Epidemiology and Assessment, and Walter Orenstein, professor at Emory University and associate director of the Emory Vaccine Center. Orenstein remarked, “Vaccines do not save lives; vaccinations save lives.” He contended that substantial resources are needed not only for vaccine development research, but also for vaccine implementation research.

Zahn noted that the dengue vaccine controversy in the Philippines illustrates the value of accurate messaging and providing the public with information on the front end to help families understand the risks of vaccines and shape public perceptions toward vaccine uptake. Orenstein added that when the COVID-19 vaccine is rolled out, serious adverse events will occur that may be either causally or coincidentally related to the vaccine. Therefore, researchers should prepare in advance to evaluate such events for causality and plan the communication strategy to address such concerns when they do occur.

Zahn remarked that gaps in immunization coverage should not be automatically attributed to vaccine hesitancy. To ensure that communities are well served by immunization systems that make it as easy as possible to get vaccinated, it is important to measure immunization rates at the community level. Orenstein added that broader, national level immunization estimates can be misleading because they miss substantial pockets of underimmunization. Detecting underimmunized subpopulations can illuminate how financial and access barriers contribute to low immunization coverage, thus highlighting the need for these barriers to be addressed when designing

immunization systems. Zahn said that the COVID-19 pandemic has led to greater outreach efforts to ensure people are informed about symptoms and the available options for getting tested. He noted this outreach work could be translated into addressing the vaccination and other health needs of specific communities, including racial or ethnic subgroups and people experiencing homelessness.

Referring to the spectrum of vaccine hesitancy, Zahn remarked on the importance of reaching out to communities around the world to learn about their perceptions, concerns, and community-specific opportunities and challenges. Given the crucial role of provider recommendations in influencing people to get vaccinated, Zahn underscored the need for providers to communicate a clear and consistent message to their patients. Orenstein said that direct sources of immunization information may be useful, but most studies have shown that primary care providers are often the most trusted source for vaccination information. Quoting a former director of communications of the U.S. Immunization Program, Orenstein said, “You need the right message delivered by the right messenger through the right communications channel.”

This page intentionally left blank.