key: cord-0956826-huun83o5
authors: Allen, Jennifer D.; Feng, Wenhui; Corlin, Laura; Porteny, Thalia; Acevedo, Andrea; Schildkraut, Deborah; King, Erin; Ladin, Keren; Fu, Qiang; Stopka, Thomas J.
title: Why are Some People Reluctant to be Vaccinated for COVID-19? A Cross-Sectional Survey among U.S. Adults in May-June 2020
date: 2021-07-14
journal: Prev Med Rep
DOI: 10.1016/j.pmedr.2021.101494
sha: 5e4426b8e019fc3d51d7a0757c59bc3721ca23c0
doc_id: 956826
cord_uid: huun83o5

Understanding reasons for COVID-19 vaccine hesitancy is necessary to ensure maximum uptake, needed for herd immunity. We conducted a cross-sectional online survey between May 29-June 20, 2020 among a national sample of U.S. adults ages 18 years and over to assess cognitive, attitudinal and normative beliefs associated with not intending to get a COVID-19 vaccine. Of 1,219 respondents, 17.7% said that they would not get a vaccine and 24.2% were unsure. In multivariable analyses controlled for gender, age, income, education, religious affiliation, health insurance coverage, and political party affiliation, those who reported that they were unwilling be vaccinated (versus those who were willing) were less likely to agree that vaccines are safe/effective (Relative Risk Ratio (RRR): 0.45, 95% confidence interval (CI): 0.31, 0.66), that everyone has a responsibility to be vaccinated (RRR: 0.39, 95% CI: 0.30, 0.52), that public authorities should be able to mandate vaccination (RRR: 0.75, 95% CI: 0.58, 0.98), and that if everyone else were vaccinated they would not need a vaccine (RRR: 1.36, 95% CI: 1.04, 1.78). Our results suggest that health messages should emphasize the safety and efficacy of vaccines, as well as the fact that vaccinating oneself is important, even if the level of uptake in the community is high.

135 public authorities and vaccination, we adapted items from the Vaccine Confidence Index 26 to assess 136 respondents' level of agreement with two statements: "Public authorities decide about which vaccines to 137 recommend based on the best interest of the community" and "Public authorities should be able to 138 mandate that everyone be vaccinated" with responses on a 5-point Likert scale (strongly agree to strongly 139 disagree), with higher scores indicating greater agreement. In addition, we asked about personal and family 140 history of COVID-19 infection, as well as questions regarding underlying health conditions believed to be 141 associated with risk of severe COVID-19 consequences at the time of this survey 27 (see Appendix A). For 142 each individual, we calculated the total number of underlying health conditions associated with elevated 143 risk of severe COVID-19, and created three groups (0, 1-2, 3+ conditions) based on univariate distribution. 144 We drew socio-demographic measures, including age (continuous), gender (male/female), education (less 145 than high school/high school/some college/bachelor's degree or higher), race/ethnicity (non-Hispanic 154 We excluded 48 respondents with missing data for the dependent variable (intention to vaccinate) and key 155 independent variables (perceived vaccine safety and efficacy, individual/societal responsibility to vaccinate, 156 and trust in the role of public authorities), yielding a final analytic sample of 1,219. Since missing data only 157 accounted for 3.8% (48/1,219) of the sample, the parameter estimates were not likely to be biased enough 158 to substantially change the results. We assessed measures of central tendency and distributions for 159 continuous variables, as well as frequencies and proportions for categorical variables. We evaluated 160 bivariate associations between vaccination intention, key independent variables, and potential confounders 161 through Chi-square tests to select measures for consideration in final adjusted models. Variables that were 162 associated with vaccination intention at a p < 0.10 level were considered for inclusion in multivariable 163 models.

165 We found that two pairs of independent variables were highly correlated, so we created composite 166 measures of the average Likert score of each pair. Specifically, we found that "Most vaccines are very safe" 167 and "Most vaccines are very effective" had a high correlation (0.82). As a result, we combined them into a 168 new variable ("Most vaccines are safe/effective"), which represents their numeric average. We also found 169 that the correlation coefficient between the two items "I have a responsibility to get vaccinated because I 170 can protect others with a weaker immune system" and "vaccination is something everyone should do to 171 protect others in the community" was 0.89, so these items were combined into a new variable ("I/Everyone 172 should vaccinate"), which is the numeric average of their Likert scales.

174 We constructed multinomial logistic regression models to identify associations between our key 175 independent measures and vaccination intentions, adjusted for covariates. Given highly politicized views of 176 the pandemic response, we also estimated multinomial logistic regression models to identify associations 177 between our key independent measures and vaccination intentions stratified by political party affiliation 178 and adjusted for remaining covariates. Variance inflation factors for covariates were accepted if they were 179 less than< 4.0. We present regression results as relative-risk ratios (RRRs), as advised by the technical notes 180 in Stata 29 and similar practice. 30 The interpretation is that RRRs are equal to the ratio of odds ratios in binary 181 logistic regressions. The interpretation is that RRRs are equal to the ratio of odds ratios in binary logistic 230 who agreed that if everyone else were vaccinated they would not need to be vaccinated were significantly 231 more likely to report unwillingness to be vaccinated compared with willingness to get vaccinated (RRR: = 232 1.36, 95% CI: = 1.04, 1.78) or being unsure about vaccination intentions (RRR = 1.26, 95% CI = 1.05, 1.53).

233 Those who agreed with the statement that public authorities should be able to mandate vaccinations were 234 also less likely to say they would not get the vaccine than would get the vaccine (RRR = 0.75, 95% CI = 0.58, 235 0.98) or than being unsure about whether to vaccinate (RRR = 0.82, 95% CI = 0.71, 0.95).

238 Political party affiliation was strongly associated with vaccine intentions in bivariate analyses. Due to this 239 partisan difference, as well as recent findings from other polls and studies, 32 we repeated the primary 240 multivariable analyses stratified by party affiliation. Overall trends in stratified models were generally 241 consistent with the primary analysis, and we did not observe substantial evidence for effect modification by 242 party affiliation on the relative scale (Appendix 2). One exception was for the relationship between 243 agreement with the statement "public authorities should be able to mandate that everybody be 244 vaccinated" and vaccine intentions. Those who agreed with this statement were less likely to report that 253 unsure. Those who reported that they would not be vaccinated were more likely to have concerns about 254 vaccine safety and efficacy, to believe that there is not an individual or societal responsibility to be 255 vaccinated, and to oppose vaccine mandates compared to those who reported that they would get a 256 vaccine. Moreover, those who said they would not be vaccinated were more likely to believe that they 257 would not need a vaccine if everyone else were vaccinated. These findings can help to inform interventions 258 designed to address the concerns and attitudes of those Americans least receptive to taking the vaccine 259 and could help to inform public health messaging as vaccination continues to be rolled out in 2021. 318 vaccination. In addition, our finding that those who did not intend to be vaccinated were more likely to 319 endorse the idea that they would not need to be vaccinated if others were, suggests a lack of 320 understanding about the importance of herd immunity. Stressing the importance of this, both for 321 protection of others and to limit opportunities for viral mutation may be key. At the interpersonal and 322 community levels, development of interventions to create injunctive norms about vaccination could be 323 particularly helpful in increasing vaccination rates. For example, messaging that appeals to the desire to 324 protect one's family members (e.g., "do this for grandmother") and a sense of collectivism (e.g., "we are all 326 to do our part").

328 Our findings suggest that those opposed to vaccination are less trusting of public authorities. Across a 329 variety of (non-COVID) vaccines, having a healthcare provider recommendation has been shown to be the 330 most influential factor in individual decision-making. 32,50 Therefore, ensuring that vaccine 331 recommendations are consistently made by healthcare providers is vital. Prior studies on other vaccine 332 types find that offering vaccines at all visits ("points of care") using a presumptive approach (i.e., assumes 333 that a patient will be vaccinated) is more effective than inquiring about vaccine hesitancy, 51 although an 334 informed decision-making approach, which stresses providing information for an individual to make their 335 own decision, has also been recommended. 52 Given partisan attitudes toward control of the pandemic 336 across the U.S., it may be important for political leaders of both parties to convey the importance of 

All analyses applied sample weights to be more representative of the U.S. population. More 183 information of sample weighting and survey design is available elsewhere

The analytic sample (n = 1,219) was 52.1% female, with a mean age of 48.1 years (linearized standard error

At the time of initial panel entry (78% in 2019), 65.6% were employed. Among those who 192 did not refuse or have missing data on religion, nearly three-quarters (73.3%) reported being affiliated with 193 a religious denomination

9%) or reported that a family member had been diagnosed with the infection

Those who strongly opposed this statement were most likely to 209 report an intention to vaccinate (74%), while those who strongly agreed were most likely to indicate they 210 would not get the vaccine (45.3%). Those who were neutral about this statement were the most likely to 211 report that they were unsure about vaccination (42.4%). Regarding trust of public authorities, half (49%) of 212 those who strongly disagreed that public authorities decide about which vaccines to

Although there was variation in vaccine intentions 219 by race/ethnicity, with 15.8% of non-Hispanic White, 28.8% of non-Hispanic Black, 19% of Hispanics

Table 3), those who believe that most 225 vaccines are safe and/or effective were less likely to say they do not intend to be vaccinated

Similarly, people who agreed that they (themselves) and everyone else should be vaccinated 228 were less likely to report that they would not be vaccinated versus being vaccinated (RRR = 0.39, 95% CI = 229 0.30, 0.52) or being unsure about vaccination intention (RRR = 0.60, 95% CI = 0.47, 0.75). Conversely, those 343 References 344 1. CDC. COVID Data Tracker

Herd Immunity to COVID-19: Alluring and Elusive

The false promise of herd immunity for COVID-19

Challenges in creating herd immunity to SARS-351

CoV-2 infection by mass vaccination. The Lancet

COVID-19 herd immunity: where are we?

Herd Immunity and Implications for SARS-CoV-2 Control

Will an imperfect vaccine curtail the COVID-19 pandemic in the 358

Herd Immunity: Understanding COVID-19

Bureau UC. How Do COVID-19 Vaccination and Vaccine Hesitancy Rates Vary Over Time? The 362 United States Census Bureau

COVID-19 vaccine hesitancy worldwide: a systematic review of vaccine acceptance rates. 366 medRxiv

Strategies for addressing vaccine 368 hesitancy -A systematic review

NW 1615 L. St, Suite 800Washington, Inquiries D

Public Trust in Government

The Strengthening of Partisan Affect

NW 1615 L. St, Suite 800Washington, Inquiries D 20036USA202-419-4300 | M-857-8562 | F

The Partisan Divide on Political Values Grows Even Wider

The rush to create a covid-19 vaccine may do more harm than good

Poll: Most Americans Worry Political Pressure Will Lead to Premature Approval of a 383 COVID-19 Vaccine; Half Say They Would Not Get a Free Vaccine Approved Before Election Day

Social media and vaccine hesitancy

COVID-19 vaccine rumors and conspiracy theories: The need 393 for cognitive inoculation against misinformation to improve vaccine adherence

It's a minefield': COVID vaccine safety poses unique communication challenge

Poll: Large Majorities Now Say They Wear Masks Regularly and Can Continue Social 398

Distancing for At Least Six Months if Needed, though Republicans Remain Less Likely to Take Such 399 Precautions

The Accuracy of Measurements with Probability and 404 Nonprobability Survey Samples: Replication and Extension

Comparing the Accuracy of RDD Telephone Surveys

Internet Surveys Conducted with Probability and Non-Probability Samples

Address-based sampling may provide alternatives for surveys that require contacts with representative 410 samples of households | Articles | Quirks

Beyond confidence: Development 416 of a measure assessing the 5C psychological antecedents of vaccination

CDC. COVID-19 and Your Health

Acceptability of a COVID-19 vaccine among adults in the United 425 States: How many people would get vaccinated? Vaccine

Equity in Health, Wealth, and Civic Engagement: A nationally 428 representative survey

Timing of COVID-19 vaccine approval 430 and endorsement by public figures

Covid Vaccine Hesitancy Drops Among All Americans, New Survey Shows. Kaiser 432 Health News

Attitudes Toward a Potential 435 SARS-CoV-2 Vaccine

A National Survey Assessing SARS-CoV-2 437 Vaccination Intentions: Implications for Future Public Health Communication Efforts

Determinants of COVID-19 vaccine acceptance in the 440 US

Willingness to get the COVID-19 vaccine with and 442 without emergency use authorization

COVID-19 vaccine hesitancy and resistance: Correlates in a 445 nationally representative longitudinal survey of the Australian population

Psychosocial and demographic characteristics relating to vaccine 448 attitudes in Australia

Psychological characteristics associated with COVID-19 451 vaccine hesitancy and resistance in Ireland and the United Kingdom

Joining the herd? U.S. public opinion and vaccination requirements across educational 454 settings during the COVID-19 pandemic

COVID-19 and vaccine hesitancy: A longitudinal study

Converting the maybes: Crucial for a 459 successful COVID-19 vaccination strategy

Latest Data on COVID-19 Vaccinations Race/Ethnicity

Poverty and Covid-19: Rates of Incidence and Deaths in the United 464 States During the First 10 Weeks of the Pandemic

Centers for Disease Control and Prevention

Latest Data on COVID-19 Vaccinations Race/Ethnicity

An ecological perspective on health promotion 472 programs

The Health Belief Model: a decade later

Predictors of HPV vaccine uptake 476 among women aged 19-26: importance of a physician's recommendation

How health care 479 providers should address vaccine hesitancy in the clinical setting: Evidence for presumptive language in 480 making a strong recommendation

New Poll: Trump Voters Want COVID Vaccine Information from Doctors, Not Politicians. de 485 Beaumont Foundation

The public's role in COVID-19 vaccination: Human-489 centered recommendations to enhance pandemic vaccine awareness, access, and acceptance in the 490 United States. Vaccine

Associations with vaccination intentions, United States

RRR: Relative Risk Ratio; CI: confidence interval

Covariates in both models include: gender (men, women); age (continuous); race/ethnic status

Hispanic, multi-racial or other); household income 530 as a percentage of Federal Poverty Level 27 ; education (less than high school, high school, some college, 531 bachelor's degree or higher); religion (Catholic, Protestant, other religion

532 health insurance (employer provided, governmental insurance or marketplace, no insuranc 533 534 Highlights 535  Nearly one-in-five U.S. adults would not get a COVID-19 vaccine

attitudinal, and normative factors are associated with COVID-19 vaccine intentions 537  Campaigns should emphasize safety/efficacy, promote trust in authorities

All authors have approved the submitted version and agree 543 both to be personally accountable for the author's own contributions and to ensure that questions related 544 to the accuracy or integrity of any part of the work, even ones in which the author was not personally 545 involved, are appropriately investigated, resolved, and the resolution documented in the literature