key: cord-0727858-ik34fwcs
authors: Shaw, Jana; Stewart, Telisa; Anderson, Kathryn B; Hanley, Samantha; Thomas, Stephen J; Salmon, Daniel A; Morley, Christopher
title: Assessment of U.S. health care personnel (HCP) attitudes towards COVID-19 vaccination in a large university health care system
date: 2021-01-25
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab054
sha: 3191178df2c8b1da069e226e0fcc1e14f759d859
doc_id: 727858
cord_uid: ik34fwcs

BACKGROUND: As a priority group, healthcare personnel (HCP) will be key to success of COVID-19 vaccination programs. The purpose of this study was to assess HCP willingness to get vaccinated and identify specific concerns that would undermine vaccination efforts. METHODS: We conducted a cross-sectional survey of HCP, including clinical and non-clinical staff, researchers, and trainees between November 23 (rd) ,2020 and December 5 (th) ,2020. The survey evaluated attitudes, beliefs and willingness to get vaccinated. RESULTS: A total of 5287 respondents had a mean age of 42.5 years (SD=13.56), and were 72.8% female (n=3842). Overall 57.5 % of individuals expressed intent to receive COVID-19 vaccine. 80.4% were physicians and scientists representing the largest group. 33.6% of registered nurses, 31.6% of allied health professionals, and 32% of master’s level clinicians were unsure they would take the vaccine (p<.001). Respondents who were older, males, White, or Asian were more likely to get vaccinated compared to other groups. Vaccine safety, potential adverse events, efficacy and speed of vaccine development dominated concerns listed by participants. Fewer (54.0%) providers of direct care vs. non-care providers (62.4%), and 52.0% of those who had provided care for COVID-19 patients (vs. 60.6% of those who had not) indicated they would take the vaccine if offered (p<.001). CONCLUSIONS: We observed that self-reported willingness to receive vaccination against COVID-19 differs by hospital roles, with physicians and research scientists showing the highest acceptance. These findings highlight important heterogeneity in personal attitudes among HCPs around COVID-19 vaccines and highlight a need for tailored communication strategies.

M a n u s c r i p t 4 tools hosted by the SUNY Upstate Medical University. The SUNY Upstate Institutional Review Board determined this project did not meet the definition of human subject research under the purview of the IRB according to federal regulations.

We developed a comprehensive survey to evaluate attitudes, beliefs and willingness to get vaccinated. We adopted a validated vaccine hesitancy survey developed by Opel et al [5] . The survey included 22 quantitative and qualitative questions. The survey included questions about basic demographic information, occupation, perceived risk and severity of COVID-19, history of prior influenza vaccination, intent to get vaccinated, who is most influential in decision to get vaccinated, and whether vaccination should be voluntary. Among subjects who expressed intent not to get vaccinated, we also inquired about their reasons. Two qualitative questions were included in the survey: QA) -Who do you think has the biggest influence on whether you will get the COVID-19 vaccine? If you chose -Someone else‖, tell us who has the biggest influence on whether you will get the vaccine?‖ and QB) -What concerns do you have?‖

The initial analytic plan assessed vaccine intent and barriers to uptake by respondents. The demographic distribution of the sample was assessed via descriptive statistics, to ascertain representativeness of the respondents compared to the overall population of the host institution. The distribution of several key variables pertaining to the perceived threat from COVID-19, and the intention to be vaccinated against COVID-10, were assessed across demographic and professional role indicators. For several variables, the five categories (Strongly Agree, Agree, Not Sure, Disagree, Strongly Disagree) were collapsed into three categories: Agree, Unsure, and Disagree. Differences in proportion were assessed via Χ2 analysis, and differences in mean age across response categories were assessed via analysis of variance (ANOVA). All quantitative analyses were performed using SPSS vs 27.

Data was downloaded from RedCap, exported into Microsoft Excel and uploaded into ATLAS.ti version 9. The two questions were analyzed using grounded theory, and were analyzed separately. First, the investigator (TS) independently reviewed a random sample of the data and openly coded the data. Once there was a strong sense of emerging theme, the investigator completed the rest of the coding using the machine learning features in ATLAS.ti. Once the machine learning was complete, the investigator randomly sampled the text with the codes to ensure the integrity of the machine learning coding. The preliminary coding scheme included 10 codes for question A, and included 9 codes for question B. For both questions, an -other‖ code was created to incorporate responses that were not identified by the machine learning coding, or did not fit with the current codes. The final high-level theme for was 11 for question A, and 10 for question B.

Quantitative analysis: A total of 5,308 responses were collected in the survey, for an approximate response rate of 55%. Of those responses, 5,287 were available for the analyses as nominally complete at the time the survey closed. Missing variables resulted in the removal of cases from the analysis; Individual missing items affected fewer than 20 cases per question. Survey respondents had a mean age of 42.5 years (SD=13.56), and were 72.8% female (n=3842), 26.0% male (n=1374), 1.1% non-binary or non-disclosed by choice (n=61), and 0.2% missing (n=8). The sample was 85.1% (n=4482) self-reported White, 5.0% Black/African-American (n=263), 6.0% Asian (n=317), and 4.2% Native American, Native Hawaiian/Pacific Islander, Other, or missing (n=222). A total of 3,095 (58.5%) of respondents reported that they provided direct patient care, and 32.2% indicated M a n u s c r i p t 5 they provided care for patients with COVID-19. Additional demographic breakdown by role is presented in Table 1 .

Vaccine intention by professional role and demographics are summarized in Table 2 . An age gradient is apparent, with those who agree or strongly agree that they would accept a vaccine being older (mean=43.83 years) than those who were not sure (41.62 years) or who disagreed (38.67 years), with significant differences across all three answer categories (p<.001). Males agreed or strongly agreed (72.5%) that they would take the vaccine for more frequently (p<.001) than females (52.4%) or non-binary or non-disclosed individuals (41.0%). Although 80.4% of scientists and physicians agreed they would take the vaccine if offered, 33.6% of registered nurses, 31.6% of allied health professionals (physical, occupational, and respiratory therapists, radiology technicians), and 32% of Master's level clinicians (advanced practice nurses or physician assistants, registered dieticians, or Master's level social workers) were unsure whether they would take the vaccine. The observed difference across roles was significant (p<.001). These role differences, particularly at Bachelor's and Master's level roles, was manifested in fewer providers of direct care (54.0%) vs. non-care providers (62.4%), and 52.0% of those who had provided care for COVID-19 patients (vs. 60.6% of those who had not) indicated they would take the vaccine if offered.

Those agreeing that they were concerned about getting COVID-19 and who endorsed the safety of the vaccine when approved, tended to be older, whereas younger respondents were more concerned for household members or close contacts getting COVID-19. Individuals identifying as White were less concerned about their own (28.4%) safety from COVID-19 than Black/African-American (41.1%) or Asian (43.5%) respondents (p<.001), and this trend extended to concern for others (household or close contacts) as well. Conversely, White respondents more frequently agreed that a COVID-19 vaccine would be safe when approved (45.9%) than Black/African-American respondents (26.2%, p<.001). Additional comparisons are presented in Table 3 .

A majority of respondents, n=3462 (65.5%) preferred vaccination to be voluntary or were unsure when asked: -Do you think COVID-19 vaccination should be voluntary or mandated for health care workers?‖ Scientists and physicians were the only group where majority preferred mandatory COVID-19 vaccination, n=570 (55.2%). All other respondents leaned towards voluntary vaccination or were unsure vaccination should be mandatory (see supplemental online table).

Qualitative Analysis: In total, 1,332 individual responses were analyzed for question A, and 3,431 responses for question B. In Question B, 13 responses were removed leaving a total of 3,418 responses. Question A Analysis "Who do you think has the biggest influence…" A total of 11 high level themes were identified (see Table 4 ). The majority of participants felt that they were the sole contributor influencing their decision (theme: Influence by Me, n=510). Participants used language such as -I am my own advocate‖, and -the only person who will influence me is myself‖. Some participants said that their decision could be influenced by research (n=250) with the presence of -medical literature on efficacy and potential side effects‖. Also noted was influence by family (n=221), where participants mentioned -immediate family‖, as influencing their decision. Participants identified experts (n=160), employers (n=137), and the government (n=119) as also having some influence on their decision making. Question B Analysis -What concerns do you have?" A total of 10 high level themes were identified (see Table 5 ). The majority of participants expressed concerns about an adverse event and/or side effect from the vaccine (theme: Concerns about Adverse Event/Side Effects, n=2307). Participants expressed this by saying things such as they were concerned -about rare adverse events‖ and -long term effects and problems or allergies‖, -new vaccine no idea about the side effect or anything long term‖. M a n u s c r i p t 6 Many participants were concerned with the efficacy of the vaccine (n=720) and expressed this concern by questioning, -how is the effective is it with it being brand new vaccine?‖. Participants also noted concerns with the vaccine being rushed (n=529), and expressed concerns about the safety of the vaccine (n=534) as well as the newness of the vaccine (n=324). These concepts were expressed in comments like, -rushed to get approved‖, -limited long term safety data‖, and -how new the vaccine is worries me… ‖ respectively.

To our knowledge, this is the largest survey of HCP attitudes towards COVID-19 vaccination. In our population, 57.5% of individuals expressed intent to receive COVID-19 vaccine, including 80.4% of physicians and scientists. These findings are consistent with a prior anecdotal report showing that 80% of New York State's community and hospital-based physicians plan to become immunized with the COVID-19 vaccine [6] . Contrary to our findings, 66.5% of HCP at UCLA intended to delay vaccination and similarly to our observations, the intent varied by hospital role with nurses having greater odds of delay compared to physicians [3] . Physicians' acceptance of COVID-19 vaccine is important as physicians often serve as a trusted source of vaccine information. Their vaccination may positively influence staff, who may be unsure and want to delay vaccination until others are vaccinated. Among all respondents, 920 (60%) listed themselves, research or expert as an influential source on vaccination guidance. Participants' feeling of influence as being -Me, myself, and I‖ was most significant in deciding about getting vaccinated. People in individualistic cultures like the U.S. perceive themselves as independent, and individual benefit is important in driving personal choices [7] . Therefore, health messages should target personal preferences, and stress risk of COVID-19 for the individual. In order to enhance vaccination, we must give people time to talk with experts (medical, public health, and other) to ensure they feel they made an informed decision that aligns with their religious, educational, philosophical or other views.

Nearly half or more of nurses, Master's level clinicians, allied health professionals, and ancillary service personnel were not sure whether the vaccine will work and protect them from COVID-19; slightly lower but similar levels of uncertainty were expressed by the same groups about vaccine safety, and nearly a third of each group was unsure whether they would take a vaccine for COVID-19 if offered for free. This high level of vaccination intent uncertainty raises concern about whether adequate vaccination coverage can be achieved through vaccination campaigns and education, without a state or federal mandate. Relevant research shows that influenza vaccination coverage is significantly higher among institutions that require influenza vaccine for employment [8] . Although prior vaccination behavior predicts a vaccine uptake, acceptance of the COVID-19 vaccine is likely to be different due to its rapid development, lack of long-term safety data, and its use under EUA, given those attributes were associated with lower vaccination intents among the U.S. public [4, 9, 10] . While we do not advocate for COVID-19 vaccine mandates, we recognize that our own high institutional vaccination influenza coverage (99.6%) is unlikely to be achieved with the COVID-19 vaccine in the absence of a mandate. Efforts to enhance vaccine acceptance among staff that is -unsure‖ will require careful, individually tailored communication strategies. They should specifically address concerns about a perception of a safety concerns due to speed of development, and potential for political influence on the review process. Communication, advocacy, and training should focus on concerns about -trials being rushed‖, safety, and efficacy, in a manner that builds trust with authorities and confidence in COVID-19 vaccines. Consideration should be given to interpersonal channels that deliver leadership by example, and leverage influencers in the institution to encourage wider adoption of vaccination. M a n u s c r i p t 7 Similarly to a recently published poll, we have observed gender differences based upon safety concerns; females were more concerned about the vaccine and the disease, compared to males [11] . This finding is not surprising as women tend to research more health-related information, make approximately 80 percent of health care decisions for their children, and become caregivers for family members [12] . Driving both clinical concerns and gender differences, nurses (86% female), other Master's and Bachelor's level clinicians, and allied health professionals, appear to manifest levels of concern about the vaccine that are likely due to expedited vaccine development, and limited vaccine information available at the time of the survey. The survey was conducted shortly before FDA hearing and public release of Pfizer-BioNTech data. Therefore, it is possible that this level of uncertainty is more reflective of a need for more information, rather than a resistance to vaccination that is exhibited along an educational gradient and by those with clinical roles.

Similar to prior report, we showed that older age was associated with intent to get vaccinated. Perceived risk of infection associated with increased age, or increased hesitancy among younger people, could explain this finding and deserves further research.

In our population, Black/African American respondents were least likely to want to get vaccinated. This is particularly concerning since African Americans are disproportionately affected by COVID-19. According to a recent survey, less than one-fifth of African Americans -mostly or completely trust" that a COVID-19 vaccine will be safe or effective [13] . Given medical mistrust in the Black community, culturally sensitive vaccination campaigns, coupled with individually tailored communication strategies, will be key to development of trust and vaccine acceptance.

We also found that direct patient care or care for COVID-19 patients is associated with lower vaccination intent. This interesting and counterintuitive observation should be explored further to provide insight into complex vaccination decisions among HCPs.

This cross-sectional survey was conducted in a single health care system in the northeastern United States. Thus, generalizability to other regions of the U.S. and other countries may be limited. Nevertheless, we do believe that there are likely lessons learned which can be more broadly applied. This survey was also voluntary, raising the possibility of selection bias among respondents opting to participate. It is unclear whether those feeling positively, uncertain, or negatively would be more likely to respond. However, our response rate was rather high for this type of survey. The similar demographic profiles of respondents and the institution as a whole is somewhat reassuring. The baseline characteristics were comparable between respondents and non-respondents, except for African Americans and Asians who were under-represented, and Whites who were over-represented (<8 percentage point difference for each race). Finally, this survey was conducted at a single point in time amidst a dynamic pandemic, wherein information, options, and perceptions are rapidly changing and one's willingness to get vaccinated may change over time, as vaccine decisions are multifactorial. However, this was an incredibly important point in time, almost immediately preceding vaccine availability in this population. We envision this survey as a starting point for ongoing discussion and engagement within our institution and our community, in the months and years ahead. M a n u s c r i p t 8

In summary, we have shown that self-reported willingness to receive vaccination against COVID-19 differs by age, gender, race and hospital role, with physicians and research scientists showing the highest acceptance. Vaccine intent and concerns about safety were concern among nurses, pharmacists and other clinical staff. Vaccine safety, adverse events, efficacy and vaccine development time dominated respondents' concerns. Vaccine hesitancy was most prevalent among those at Master's, Bachelor's, and lower education levels. We believe our findings may help guide public health campaigns to enhance acceptance of COVID-19 vaccine. M a n u s c r i p t 9

Acknowledgements: This research would not have been possible without the exceptional support of Kathleen Froio Pace, Mark Zeman and Susan Murphy. Their enthusiasm, flexibility and generous support to ensure that survey reaches as many Upstate Medical University staff was invaluable and deserves special recognition. We are also grateful to all Upstate staff who participated in our survey and shared their views to inform our vaccination strategy.

None of the authors have received any funding for the presented 1 = includes research staff and medical students; 2 = includes educational support and other; 3 = includes clerical, dietary, phlebotomy, unit support, registration, clinical support, environmental services; 4 = includes laboratory, informational technology, and pharmacy services ; 5 = includes physical therapy, occupational therapy, radiology, and respiratory services ; 6 = includes nurse practitioners, physician's assistants, social workers, and registered dieticians * = includes American Indian or Alaska Native, Hawaiian/Pacific Islander, or Other. Excludes missing.

M a n u s c r i p t 13 a = Chi-Square or ANOVA p-value 1 = includes research staff and medical students; 2 = includes educational support and other; 3 = includes clerical, dietary, phlebotomy, unit support, registration, clinical support, environmental services; 4 = includes laboratory, informational technology, and pharmacy services; 5 = includes physical therapy, occupational therapy, radiology, and respiratory services; 6 = includes nurse practitioners, physician's assistants, social workers, and registered dieticians M a n u s c r i p t 18 

COVID-19 Map -Johns Hopkins Coronavirus Resource Center. 2020. Available at

COVID-19 Vaccination Program Interim Operational Guidance for Jurisdictions Centers for Disease Control and Prevention (CDC) COVID-19 Vaccination Program Interim Playbook for Jurisdiction Operations

Assessment of COVID-19 vaccine acceptance among healthcare workers in Los Angeles

Factors Associated With US Adults' Likelihood of Accepting COVID-19 Vaccination

Validity and reliability of a survey to identify vaccine-hesitant parents

Medical Society of the State of New York. 80 Percent of NY Physicians Plan to Get Immunized When Vaccine Is Widely Available

Implications for Cognition, Emotion, and Motivation

Influenza Vaccination Coverage Among Health Care Personnel -United States

Importance of vaccination habit and vaccine choice on influenza vaccination among healthy working adults

Mandating COVID-19 Vaccines

More women than men in U.S. nervous about fast rollout of COVID vaccine, and that's a problem: Reuters/Ipsos poll | Reuters

National Association for the Advancement of Colored People, COVID Collaborative. Coronavirus Vaccine Hesitancy in Black and Latinx Communities

M a n u s c r i p t 17