Temporal Changes in Vaccine-Specific Willingness Across Race/Ethnicity Following Serious Adverse Event Reports

Vaccination for COVID-19 remains a major public health issue as the pandemic continues to cause hospitalizations and deaths across the world.¹ As of September 2023, an estimated 1.14 million deaths in the United States and 6.95 million across the world are attributable to COVID-19.² Despite such deaths, only 69.5% of the US population has completed the primary COVID-19 vaccine series, and only 17.0% of US adults have received the updated bivalent booster dose.³ There is strong scientific evidence for the public health benefit of the COVID-19 vaccination, which can be measured as averted cases, hospitalizations, and death, and continued vaccination is urgent.⁴ An improved understanding of COVID-19 vaccine hesitancy is vital to addressing the public health emergency and to reducing health disparities.⁵

Vaccine hesitancy has been described as a delay in acceptance or a refusal of vaccination despite availability of vaccination services.⁶ Despite early evidence of hesitancy among racial/ethnic minorities (i.e., Hispanic, Asian, African American, American Indian/Alaska Native, and Pacific Islander), the largest observed decreases in hesitancy have been among African American and Hispanic populations.⁴ Vaccine hesitancy, however, can be worsened by concerns about side effects, and the concern about a vaccine’s side effects may prompt vaccine hesitancy toward other vaccines.^7,8

By contrast, after the April 2021 “pause” of the Johnson and Johnson (J&J) vaccine because of vaccine-related serious adverse events, some investigators found no significant changes in trends of the proportion of the US adult population who were hesitant about getting a COVID-19 vaccine. They did find a significant increase in concerns about the safety and efficacy of COVID-19 vaccines among the already hesitant population.⁹ However, vaccine hesitancy can be complex and context specific and can vary over time and place and be specific to a particular vaccine. Because of shifts in vaccine hesitancy, there is a need to better characterize not only temporal changes in hesitancy but also the geographic and racial/ethnic differences through real-time data collection. We studied the statewide temporal changes in vaccine-specific willingness to be vaccinated in various races/ethnicities in Arizona following reports of serious adverse reactions to the J&J COVID-19 vaccine.

We administered a longitudinal survey developed by the National Institutes of Health–funded Community Engagement Alliance against COVID-19 Disparities to adult community residents. The recruitment for the survey was done via television and radio advertisements, distribution of flyers by community health workers, and e-mails to participants in a COVID-19 antibody surveillance program. We excluded individuals younger than 18 years, incarcerated persons, and persons with diminished mental capacity (see the Appendix, available as a supplement to the online version of this article at http://www.ajph.org). We surveyed respondents (n = 314) between April 2021 and April 2022. A majority (88.3%) indicated that they already had received a COVID-19 vaccination.

In mid-April 2021, the Food and Drug Administration (FDA) recommended a pause on the administration of the J&J vaccine after reports of rare blood clots and related deaths in women.¹⁰ In the latter part of April 2021, the FDA released the pause with black box warnings for administration in women who were pregnant, used oral contraceptives, or had a family history of blood clots.

We observed a decrease in willingness among non-Hispanic Whites and minorities to receive the J&J vaccine over time following reports of serious adverse events related to this vaccine (Figure 1). The willingness over time to receive the Pfizer or Moderna COVID-19 vaccine was higher than the willingness to receive the J&J vaccine in both non-Hispanic White and minority individuals (Figure 1). Among minorities, the decrease in J&J vaccine willingness was linear, dropping at a constant rate, but it was nonlinear among non-Hispanic White people; it was initially stable from July through about September 2021, sharply decreased until December 2021, and then remained stably low through April 2022.

FIGURE 1—

Nonlinear Smooths on the Fitted Values for the Johnson & Johnson, Moderna, and Pfizer Vaccines in (a) Non-Hispanic White, and (b) Minority Respondents: Arizona, July 2021–April 2022

Note. The y-axis represents the expected value of the latent variable on the real line. The dotted horizontal lines mark the cutoffs (thresholds) that separate the categories in the ordinal response variable willingness (1 = not at all willing; 7 = very willing; 8 = already vaccinated). The categories are indicated on the right side of each panel. Shaded bands represent the 95% simultaneous confidence interval.

Willingness to get the Pfizer vaccine also decreased slightly over time among minority ethnic groups, but the decrease was not as steep as that for the J&J vaccine. Among non-Hispanic White people, willingness to get vaccinated remained stable. Willingness to get the Moderna vaccine also remained stable for both minority and non-Hispanic White groups. Moreover, non-Hispanic White respondents expressed a higher willingness to receive the J&J vaccine than did minority respondents at the beginning of the study period (July–October 2021). Willingness to receive either the Moderna or Pfizer vaccine did not differ among races/ethnicities. We observed similar results in sensitivity analyses that removed or coded already vaccinated cases differently.

We observed that hesitancy for the J&J COVID-19 vaccine increased after reports of vaccine-related clots and death attributable to pulmonary embolism. However, such reports did not significantly influence respondents’ willingness to get the Pfizer or Moderna vaccines. Our study was limited by the lack of information regarding vaccine willingness between April 2021 and June 2021. However, it is noteworthy that even at baseline (July 2021), willingness to get the Pfizer or Moderna vaccine was generally higher than was willingness to get the J&J vaccine. Such a finding may be attributable to knowledge regarding relative efficacy of these vaccines to prevent COVID-19–caused severe disease, hospitalization, and death.¹¹

The spread of health information regarding vaccines is complex. We observed that the change in willingness to get the J&J vaccine continued to drop over many months after the initial reports of adverse events specific to this vaccine. Although such delays may be attributable to awareness, persuasion, and decisiveness of the individual respondent, the lag time of at least five months for both non-Hispanic White and minority individuals raises questions regarding the efficiency of disseminating vaccine-related health information in communities. Moreover, the observed lower ratings for vaccine-specific willingness in minorities—both at baseline observation and during the follow-up time-period—when compared with non-Hispanic Whites may reflect underlying cognitive biases toward vaccinations in general.¹² Such findings highlight the need for rapid and effective communication to promote public health that is performed in a culturally appropriate and scientific manner and is devoid of biases.

The greater representation of non-Hispanic Whites, women, and individuals with higher income in our survey falls short of adequate representation of various communities. Moreover, a greater representation of individuals who had already received vaccinations limits generalizability to unvaccinated populations, but such populations are also less likely to volunteer for research studies, possibly because of underlying inherent biases against science. Also, the period between April and July 2021 was not represented in our survey. Nevertheless, the observed changes in vaccine-specific willingness and the temporal changes over time in a predominantly vaccinated population reveal that vaccine-specific hesitancy may not always confer general vaccine hesitancy in the population we studied.

Our statewide data among a predominantly vaccinated population holds broader public health significance in two respects: (1) vaccine-specific hesitancy toward one brand or manufacturer did not influence hesitancy toward other COVID-19 vaccines, and (2) there was a notable time lag in vaccine willingness following initial reports of adverse events. These data demonstrate the need for rapid and effective communication of scientific information in a trustworthy and culturally appropriate manner that promotes and preserves public health. Moreover, real-time measurements of vaccine hesitancy may help medical and public health officials anticipate and possibly intervene dynamically with science-based information using trustworthy sources in future pandemics.