COVID-19 Vaccine Acceptance Among Healthcare Workers in a United States Medical Center

INTRODUCTION

COVID-19 vaccines hold promise to minimize hospitalizations and deaths due to SARS-CoV-2, which has claimed the lives of more than 512,000 Americans to date.¹ The Federal Drug Administration (FDA) issued Emergency Use Authorizations (EUAs) for the Pfizer-BioNTech and Moderna COVID-19 vaccines in December 2020 and for the Johnson & Johnson vaccine in February 2020.^2,3 Recognizing that vaccine supply would be limited initially, the Centers for Disease Control and Prevention⁴ prioritized healthcare personnel for the first phase of vaccination in the United States to keep critical healthcare infrastructure open and functioning. By the end of December 2020, healthcare systems across the country began offering COVID-19 vaccines to employees.

However, access does not guarantee uptake,⁵ and it remains unclear to what degree healthcare workers will accept the COVID-19 vaccine as it is made available to them. In one study of long-term care facility staff, only 38% of eligible workers received COVID-19 vaccination in December 2020-January 2021.⁶ A survey of 1000 healthcare workers in January 2021 documented widespread vaccine hesitancy, with 1 in 4 respondents anticipating declining COVID-19 vaccination.^7,8 Moreover, prior work during the H1N1 influenza pandemic demonstrated variation in vaccine acceptance across roles of healthcare employees.⁹ It is currently unclear how COVID-19 vaccine acceptance and hesitancy might vary across the diverse healthcare employee ecosystem, which includes clinicians, other staff not providing clinical care but with clinical contact (e.g., unit clerks, hospitality staff, security staff), administrators, and researchers.

It is critical to address this knowledge gap to inform ongoing and future vaccine distribution efforts. When a healthcare worker declines a COVID-19 vaccine, it affects the herd immunity of the healthcare workforce and potentially the safety of patients and communities that workforce serves. Healthcare workers also constitute a trusted voice that influences others—vaccine hesitancy in this group could undermine efforts to widely vaccinate populations and achieve herd immunity, as has been modeled for seasonal influenza.¹⁰ Better understanding of healthcare workers’ attitudes and beliefs about COVID-19 vaccination could identify opportunities to create more effective vaccine messaging and policies and ultimately promote more widespread vaccine coverage and improved population health outcomes.

Our objective was to evaluate vaccine intentions and attitudes among a diverse sample of medical center employees eligible for COVID-19 vaccination and to explore variations in vaccine beliefs among employees with different occupational and demographic characteristics.

METHODS

RESULTS

A total of 39,259 individual and group email accounts received the study invitation and 11,387 individuals responded. This corresponds to a response rate of at least 29% of the employee population (the presence of group email accounts in the distribution list means that this calculation is an underestimate of the true response rate). Mirroring the demographics of the health system employee population,¹² participants were largely non-Hispanic White, evenly distributed across age groups, and reflective of a wide variety of professional roles within the healthcare system (Table 1).

As of the date of survey response, most respondents had received or were scheduled to receive the COVID-19 vaccine (n=9081, 79.8%) or planned to receive the COVID-19 vaccine as soon as possible (n=546, 4.8%). However, more than 11% of respondents were vaccine hesitant, either not intending to get it soon but might in the future (n=954, 8.4%) or not intending to ever get the vaccine (n=369, 3.2%). Vaccine intention differed significantly across professional role, high-risk unit, personal infection with COVID-19, gender, race/ethnicity, and age (Table 1). In univariate analysis, personal infection with COVID-19, female gender, non-Hispanic Black/Mixed/Other race/ethnicity, and younger age were significantly associated with decreased vaccination intentions.

Vaccine intentions also demonstrated notable shifts across professional roles, especially when stratified by gender (Figure 1) and race/ethnicity (Figure 2). Physicians, nurse practitioners/midwives/physician assistants, and trainees demonstrated the highest vaccine acceptance, while nurses, other staff providing clinical care, those without clinical responsibilities but with clinical contact, and those with other professional roles not involving clinical contact demonstrated lower vaccine acceptance. In these professional groups with lower vaccine acceptance, there was a pattern of greater vaccine hesitancy among females, who demonstrated higher proportions of “not now” and “not ever” intentions compared to males (Figure 1). When employment roles were stratified by race/ethnicity (Figure 2), physicians had extremely high vaccine acceptance across all racial groups, with only five physicians in the whole sample reporting vaccine hesitancy. There were no Non-Hispanic Black/Mixed/Other physicians who reported vaccine hesitancy. In contrast, in nearly all other employment groups, the highest rates of vaccine hesitancy were observed among Non-Hispanic Black/Mixed/Other employees. For example, “not now” and “not ever” intentions were reported by more than 1 in 4 other staff providing clinical care (“not now,” 22.8%; “not ever,” 12.6%) and staff not providing clinical care but with clinical contact (“not now,” 20.0%; “not ever,” 6.4%). Employees who self-identified as Hispanic also demonstrated patterns of higher vaccine hesitancy, with 19% of Hispanic nurses reporting “not now” and 8.7% reporting “not ever” vaccine intentions. Across all employee groups (with the exception of physicians, who demonstrated globally high rates of acceptance), employees who self-identified as Asian had the highest rates of vaccine acceptance compared to all other racial groups. Formal statistical testing for interaction between gender groups and role and racial-ethnic groups and role could not be reliably performed due to small overall numbers in some subgroups.

• Download figure
• Open in new tab

Figure 1. Not Now and Not Ever vaccine intentions according to medical center employee role stratified by gender

Not Now=I do not intend to get the vaccine soon but might get it in the future; Not Ever=I do not intend to ever get the vaccine; Other Staff Providing Clinical Care=phlebotomist, radiology/imaging technician, medical assistant, etc.; Other Staff Not Providing Clinical Care but with Clinical Contact=unit clerk, check-in/check-out, food service, etc.; Trainee=medical student, graduate student, post-doctoral student; NP=Nurse Practitioner; PA=Physician Assistant

• Download figure
• Open in new tab

Figure 2. Not Now and Not Ever vaccine intentions according to medical center employee role stratified by race/ethnicity

Not Now=I do not intend to get the vaccine soon but might get it in the future; Not Ever=I do not intend to ever get the vaccine; Other Staff Providing Clinical Care=phlebotomist, radiology/imaging technician, medical assistant, etc.; Other Staff Not Providing Clinical Care but with Clinical Contact=unit clerk, check-in/check-out, food service, etc.; Trainee=medical student, graduate student, post-doctoral student; NP=Nurse Practitioner; PA=Physician Assistant

In multivariable logistic regression models predicting vaccine receipt/intention, there were strong associations with professional role (Table 2). Compared to nurses, physicians (aOR 22.2, 9.1-54.3), trainees (aOR 5.9, 3.0-11.4), and nurse practitioners/nurse midwives/physician assistants (aOR 1.9, 1.2-3.0) were significantly more likely to demonstrate vaccine acceptance, whereas other clinical staff were significantly less likely (aOR 0.8, 0.6-0.9) and all other employees marginally less likely (aOR 0.8, 0.6-1.0). Working in a high-risk clinical unit versus not was marginally associated with a higher rate of vaccine acceptance (aOR 1.2, 1.0-1.4) after controlling for other covariates. Prior infection with COVID-19 was associated with significantly lower vaccine acceptance (aOR 0.4, 0.3-0.4). Male employees were almost twice as likely to accept vaccines compared to female employees (aOR 1.9, 1.6-2.4). Compared to non-Hispanic White respondents, non-Hispanic Asian respondents had significantly higher odds of vaccine acceptance (aOR 2.3, 1.5-3.4), while non-Hispanic Black/Mixed/Other respondents had significantly lower odds of vaccine acceptance after controlling for other covariates (aOR 0.4, 0.4-0.5). Increasing age was significantly associated with vaccine acceptance, with increased acceptance across age groups compared to the reference group of respondents ≤34 years of age. In subgroup analyses, associations between sociodemographic and occupational characteristics and vaccine acceptance persisted among only female respondents, but diverged among only male respondents. Specifically, compared to male nurses, only male physicians had higher odds of vaccine acceptance, while other roles all demonstrated statistically similar odds. Respondents with prior infection with COVID-19 and non-Hispanic Black/Mixed/Other respondents continued to demonstrate decreased odds compared to peer groups, but age was no longer strongly predictive of vaccine intention.

Overall, 29.6% of respondents reported at least one concern about COVID-19 vaccination, with vaccine concerns varying by vaccine intention (Table 3). Among those with vaccine hesitancy, the most commonly reported concerns were mistrust in the vaccine because of how quickly it was developed, insufficient safety and effectiveness data, disbelief that the vaccine would protect against COVID-19 infection, and concerns about serious side effects from vaccination. Even in respondents accepting the COVID-19 vaccine, a notable proportion also harbored vaccine concerns, with 21.2% responding affirmatively to at least one concern, including 6% concerned about how quickly the vaccine was developed, 6% concerned about lack of safety and effectiveness data, and 7% concerned about the potential for bad reactions to the vaccine.

DISCUSSION

During a pandemic, health systems must rapidly respond to keep critical infrastructure operating. Employee vaccination is central to ensuring a healthy workforce, but vaccine concerns may limit widespread uptake of vaccines. In a timely assessment of over 11,000 medical center employees’ COVID-19 vaccine intentions conducted just as vaccination was occurring within a healthcare system, we document important patterns in COVID-19 vaccine acceptance and hesitancy. Despite high vaccine acceptance overall, there was notable variation across professional roles, with physicians, nurse practitioners/nurse midwives/physician assistants, and medical trainees demonstrating significantly higher vaccine acceptance than nurses, other staff with clinical roles, other staff without clinical responsibilities but with clinical contact, and non-clinical faculty and staff. Women, individuals identifying as non-Hispanic Black/Mixed/Other, those with prior COVID-19 infection, and younger adults were also significantly less likely to have accepted COVID-19 vaccination.

Prior studies, conducted before Phase III clinical trial results were released and FDA EUA was granted, hinted at discrepancies in hypothetical COVID-19 vaccine acceptability by healthcare worker role, with higher hesitancy among non-physician personnel.^13,14 Another recent study documented higher vaccine acceptance among clinical compared to non-clinical health system workers.¹⁵ Our study significantly advances this prior work by newly documenting real-world COVID-19 vaccine acceptance among medical center employees and significant variation in vaccine delay/deferral across professional role and demographic characteristics—especially gender, race/ethnicity, and age. Our findings call for nuanced, nimble communication strategies that account for the multiple intersecting characteristics that may affect a medical center employee’s risk of vaccine acceptance, recognizing, as others have noted, that no demographic group is a monolithic group with a single, collective viewpoint.^16,17

Our findings advance other studies suggesting that vaccine hesitancy among many healthcare workers is less driven by overall vaccine skepticism and more by specific concerns about the lack of long-term data on safety, efficacy, and potential side effects of the COVID-19 vaccines.^13,14,18,19 Many concerns were cross-cutting, noted even among some of those accepting COVID-19 vaccination. These concerns constitute potential targets for communication interventions to reduce vaccine concerns. Clear, consistent messaging about the rigorous clinical trials process, ongoing safety data, and FDA approval process may help mitigate hesitancy. However, decades of anthropological research caution against simply providing “more and better” information without also leveraging empathy.^20,21 Acknowledging the legitimacy of vaccine concerns and fear of the unknown may be a critical tool to help dismantle vaccine hesitancy. In fact, as Lisa Rosenbaum has suggested,²¹ healthcare workers with vaccine concerns may be potent vaccine champions, uniquely poised to understand patients who are reluctant to get vaccinated, to communicate that what is known about the vaccine is more important than what is unknown, and to change hearts and minds to increase vaccine uptake.

Our findings hold important implications for clinical practice, public health efforts, health policy, and research, both immediately during the COVID-19 pandemic and more broadly to guide efforts in future infectious disease outbreaks. More robust communication efforts are needed to address vaccine concerns among medical center employees, particularly among non-physician groups, women, and non-Hispanic Black/Mixed/Other individuals. Increasing vaccine confidence among healthcare workers is foundational to successfully instilling vaccine acceptance in other populations and achieving vaccine coverage sufficient to stop viral spread.^22,23 Emerging research suggests that “Be a protector” messaging, which emphasizes empathy and being a protector for loved ones, may better reduce vaccine resistance than traditional “get the vaccine” messages, which may fuel resistance by eliciting shame and anger about being told what to do.²⁴ “Protector” schema messages may be particularly salient for healthcare workers, who have a shared goal and obligation to protect their patients and community. Tailored messaging that targets the informational needs and preferences of specific groups may also be helpful.

Contributors

The authors thank Sarah Block for tireless assistance with manuscript preparation.

Funding

None

Prior presentations

None