The COVID-19 Pandemic and Parental Attitudes Toward Routine Childhood Vaccines

We analyzed data from Health eMoms, a novel longitudinal survey program administered by the Colorado Department of Public Health and Environment, which randomly samples birthing parents in Colorado based on birth certificates and measures a variety of health topics. Birthing parents are recruited by mail to complete online surveys (in English or Spanish) from the time of recruitment through 3 years postpartum to gain a better understanding of the experiences of birthing parents and their babies. Embedded within the first Health eMoms survey, which is completed between 3 and 6 months postpartum, is the Parent Attitudes about Childhood Vaccines short scale questionnaire (PACV5), a 5-item short version of a validated survey tool on parental vaccine hesitancy that has been used in prior studies.^9,10  Respondent demographics, including birthing parent race or ethnicity, birthing parent age, previous live births, and educational attainment are obtained from birth certificate records. Other demographics included in this study were collected on the first Health eMoms survey.

All completed Health eMoms surveys from April 1, 2018 through August 31, 2021 were included in the analysis (n = 3594). Respondents were excluded if they did not select a PACV5 response (n = 5) or were surveyed during March 2020 (n = 36) as a washout period because of the rapidly evolving nature of the pandemic during that month. Health eMoms survey data are weighted on key demographic variables (age, race, ethnicity, education, geography, marital status, parity, and Medicaid status) using iterative proportional fitting to ensure it is representative of the target population, all births in Colorado.

Additional information about the Health eMoms Survey, including survey methodology, survey response rates by year, and weighting procedures, are available elsewhere.¹¹ 

The PACV5 vaccine hesitancy questionnaire embedded within the Health eMoms survey consists of 5 questions with a 5-point Likert scale of responses (Supplemental Fig 3). Consistent with prior studies, to measure overall vaccine hesitancy, each of the responses on the 5-item PACV5 questionnaire were scored from 0 to 2 with nonhesitant responses receiving a score of 0, equivocal responses (“not sure” or “neither agree nor disagree”) receiving a score of 1, and hesitant responses a score of 2. The individual scores were summed, and the cumulative score ranged from 0 (least hesitant) to 10 (most hesitant). Respondents were classified as vaccine hesitant if they had a total PACV5 score ≥4, as previously described.¹²  To measure individual PACV5 question responses, each of the responses on the 5-item PACV5 questionnaire were collapsed to 3 levels: nonhesitant responses (“nonhesitant” or “agree” responses); unsure responses (“not sure” or “neither agree nor disagree” responses); or hesitant responses (“hesitant” or “disagree” responses). PACV5 questions 2, 3, and 4 were reverse coded because of the polarity of the question stem (hesitant responses [“agree”] and nonhesitant responses [“disagree”]). COVID-19 time periods were categorized as prepandemic: April 1, 2018 to Feb 29, 2020; pandemic prevaccine: April 1, 2020 to December 31, 2020; or pandemic postvaccine: January 1, 2021 to August 31, 2021. COVID-19 pandemic periods were chosen based on COVID-19 vaccine approval. Birthing parent sociodemographic characteristics assessed included age (<30 or ≥30 years), race (American Indian or Alaska Native, Asian, Black, Chinese, Filipino, Japanese, Mixed Race, other Asian, other nonwhite, or white), ethnicity (Hispanic or Non-Hispanic), preferred language (English or Spanish), insurance status (public, private, or self-pay), education (less than high school education, high school education, more than high school education), region of residence (rural or metro county), and previous live births.

In the weighted sample, changes in the proportion of vaccine hesitant parents during different COVID-19 pandemic periods were compared using Rao-Scott χ² tests. Population survey weighted multivariable logistic regression was used to measure the association between the COVID-19 pandemic time periods and overall vaccine hesitancy. Multivariable population survey multinomial logistic regression was used to measure the association between the 3 levels of response (nonhesitant, unsure, hesitant) for each PACV5 question and COVID-19 pandemic periods. Regression models were adjusted for selected covariates determined a priori based on previously described differences in vaccine hesitancy and COVID-19 pandemic impact, including age, race, ethnicity, primary language, insurance status, education, region of residence, and previous live births. All tests were performed as 2-sided tests with a level of significance of 0.05. Data were analyzed using SAS statistical software version 9.4 (SAS Institute). The Colorado Multiple Institutional Review Board determined this study was exempt from IRB review (COMIRB# 22-1849).

The overall response rate for the Health eMoms survey during the study period was 43.8%. A sample of 3553 respondents was analyzed, representing a weighted population of 205 496 birthing parents in Colorado. Demographics of unweighted and weighted survey populations along with bivariate analyses of demographics by overall parental vaccine hesitancy (PACV5 score ≥4) are outlined in Table 1. In the weighted sample, 21.7% (n = 44 518) of all birthing parents were vaccine hesitant. There was no significant difference in the proportion of parents who were vaccine hesitant during the prepandemic period compared with pandemic prevaccine (pre COVID-19 vaccine availability) and late pandemic (post COVID-19 vaccine availability) time periods (22.0% versus 19.4% and 23.3% respectively; P = .2). In unadjusted bivariate analyses, significant differences were found between vaccine hesitant and nonhesitant parents in regard to race, insurance status, education, region of residence, multiparous birthing parent, and age (Table 1).

In the weighted multivariable adjusted model, there were no significant differences in the odds of overall parental vaccine hesitancy in the pandemic prevaccine period compared with the prepandemic period (adjusted odds ratio [aOR] = 0.82, 95% confidence interval [CI] = 0.65–1.04) or in the later pandemic postvaccine period compared with the prepandemic period (aOR = 1.07, 95% CI = 0.85–1.34) (Table 2). Trends in the proportion of overall vaccine hesitant parents and predicted values across different COVID-19 pandemic periods using the final weighted multivariable model are presented in Fig 1. There were significant differences in the odds of parental vaccine hesitancy by race, preferred language, insurance status, education level, and previous live births. Specifically, there was a higher likelihood of vaccine hesitancy in parents who self-identified as Asian or Black compared with white ([aOR = 1.58; 95% CI = 1.03–2.42] and [aOR = 1.63; 95% CI = 1.05, 2.51], respectively), parents who preferred English language compared with Spanish (aOR = 2.52; 95% CI = 1.38–4.58), parents with public or self-pay insurance compared with private insurance ([aOR = 1.57; 95% CI = 1.26–1.96] and [aOR = 3.13; 95% CI = 2.02–4.86], respectively), parents with only a high school education compared with more than high school education (aOR = 1.53; 95% CI = 1.18–1.97), and those with previous live births (aOR = 1.26; 95% CI = 1.04–1.52).

In the weighted sample, responses to each individual PACV5 question were assessed using a weighted multivariable multinominal adjusted model. Respondents were more likely to be unsure compared with agreeing (nonhesitant response) with the statement “I trust the information I receive about shots” in both pandemic time periods compared with the prepandemic period (pandemic prevaccine: aOR = 1.68; 95% CI = 1.19–2.36; Pandemic postvaccine: aOR = 2.14; 95% CI = 1.55–2.96) (Fig 2). Respondents were less likely to agree (hesitant response) compared with disagree (non-hesitant response) with the statement “It is better for my child to develop immunity by getting sick than to get a shot” in the pandemic prevaccine time period compared with the prepandemic period (aOR = 0.65; 95% CI = 0.46–0.92), which reversed in the later pandemic postvaccine time period (aOR = 1.51; 95% CI = 1.13–2.01). Furthermore, respondents were less likely to agree (hesitant response) compared with disagree (nonhesitant response) with the statement “Children get more shots than are good for them” in both pandemic time periods compared with the prepandemic period (Pandemic prevaccine: aOR = 0.67; 95% CI = 0.52–0.88; pandemic postvaccine: aOR = 0.68; 95% CI = 0.52–0.89). Finally, respondents were less likely to have an unsure response compared with a nonhesitant response to the statement “Overall, how hesitant about childhood shots would you consider yourself to be?” in both pandemic time periods compared with the prepandemic time period (pandemic prevaccine: aOR = 0.42; 95% CI = 0.23–0.75); pandemic postvaccine: aOR = 0.48; 95% CI = 0.27–0.84).

In this study of a large and diverse population of parents with young children, there was no evidence of a change in the proportion of parents hesitant toward routine childhood vaccines after the onset of the COVID-19 pandemic compared with the prepandemic levels. However, there were significant changes in parental attitudes about specific vaccine hesitancy topics that contribute to overall vaccine hesitancy. Additionally, there were notable differences in vaccine hesitancy associated with birthing parent sociodemographic factors, including race, preferred language, health insurance status, parity, and education level.

The overall level of vaccine hesitancy in this study mirrors a 2019 National Immunization Survey report finding that approximately 20% of parents are vaccine hesitant.¹³  Despite anecdotal reports that vaccine hesitancy toward childhood vaccines has increased during the COVID-19 pandemic, in this study the degree of vaccine hesitancy during the pandemic was not significantly different than prepandemic levels. Other studies demonstrating worsening parental childhood vaccine hesitancy during the COVID-19 pandemic in the United States were unable to compare levels of hesitancy directly and systematically to the prepandemic levels.^4–6  The results of this study are in line with recent national data suggesting that persistent pandemic-related drops in vaccination coverage are due primarily to access rather than hesitancy.^14,15  In recent national polls from the Kaiser Family Foundation and Pew Research Center, a majority (85% to 88%) of American adults agreed the benefits of childhood vaccines, such as the measles, mumps, and rubella vaccine, outweigh the risk, and these data are unchanged from a similar Pew Research Center poll in 2019.^8,16,17  Furthermore, the results this study presented here are similar to another study performed earlier in the pandemic, which demonstrated parental vaccine attitudes improved early in the pandemic and returned to prepandemic baseline levels by December 2020.⁷ 

Although this study did not find changes in overall vaccine hesitancy, there were significant changes in survey questions that address different aspects of vaccine hesitancy. Notably, this included decreasing trust in the information parents receive about shots, a disagreement with early and then reversal later in the pandemic in the belief that it is better for children to develop immunity by getting sick than with a vaccine and increased polarization of how hesitant parents are about childhood vaccines, with fewer parents being unsure about their vaccine hesitancy during the pandemic time periods. Decreasing trust in vaccine information may be associated with decreasing trust toward COVID-19 vaccines and science in general during the pandemic time period.^18–20  Initial disagreement with the idea that it is better to develop immunity by getting sick than with a vaccine early in the pandemic may have reversed direction later in the pandemic as a spillover effect of changes in perceived COVID-19 disease risk, vaccine safety, and COVID-19 vaccine benefits.³  Finally, the increased polarization about how hesitant parents are toward childhood vaccines (as shown in Fig 2, Question 5) is likely the result of many different contextual factors, including the growth of hyperconnected digital platforms and political polarization.²¹  Taken together, parents who are vaccine hesitant may be more certain of their beliefs and thus less likely to follow vaccination recommendations. Further studies to describe these trends are important for a complete understanding of current parental vaccine hesitancy in the United States.

Data on the association between vaccine hesitancy and sociodemographic factors are mixed, complex, and have evolved over time. Differences in vaccine hesitancy among sociodemographic factors described in this study correlate with recent national data highlighting significant sociodemographic disparities in vaccine coverage.^14,15  Similar to this study, in a recent large US population survey both lower level of income and lower educational attainment was associated with increased rates of parental vaccine hesitancy, although that study did not find race or ethnicity associated with hesitancy.²²  Alternatively, an older study of electronic health records and claims data from 2 large US health maintenance organizations found vaccine refusers were more likely to reside in well-educated and higher income areas.²³  Furthermore, several studies have found Black and Hispanic parents are more likely than white parents to have negative attitudes toward vaccines, although these studies were conducted over 15 years ago.^24,25  Observed racial and ethnic differences in parental vaccine attitudes and uptake must be interpreted within the long history of medial mistreatment toward historically marginalized communities and ongoing structural racism.^26,27  Efforts to address vaccine hesitancy and uptake in these communities requires equitable distribution of immunization resources and exhaustive work to build trust in medicine and public health.

The results of this study have several strengths and limitations. First, this study is one of the few studies that compared parental vaccine hesitancy in the United States during the COVID-19 pandemic to the prepandemic time period and utilizes one of the few data sources available to make this comparison. Additionally, we used a tool to measure hesitancy that has strong psychometric properties and has been validated and used previously to reflect vaccination behavior.²⁸  Furthermore, this was a large survey with a demographically and geographically diverse population, although it was limited to birthing parents of young children in a single state and results may not be generalizable to other populations or nonbirthing parents. A limitation of this study was the inability to assess changes in attitudes to specific vaccines. As other studies outside the US have demonstrated, the COVID-19 pandemic may have caused differential changes in parental attitudes toward specific childhood vaccines, such as influenza vaccines, which should be explored further.^29–31  Additionally, inherent to all survey studies, attitudes toward childhood vaccines may have differed between survey respondents and nonrespondents, although the vaccine hesitancy questions in this study were incorporated into a larger survey about a variety of health topics. Key demographic differences between respondents and nonrespondents were adjusted for through sample weighting, which helps mitigate nonresponse bias. American Indian or Alaska Native, mixed race, and other nonwhite race categories had small frequencies (n < 30) in the bivariable analysis (Table 1), which limited inferences on differences among hesitancy groups for these categories. Finally, although vaccination behavior was not evaluated directly in this study, parental vaccine hesitancy has been shown to be predictive of vaccination behavior.⁹  Future studies are warranted to investigate whether pandemic related changes in vaccine attitudes resulted in changes in routine childhood vaccination behavior.

In light of these findings, what interventions should be emphasized to improve childhood vaccination uptake following the COVID-19 pandemic? First, it remains important to address ongoing childhood vaccine hesitancy. Despite changes during the pandemic, physicians and other healthcare professionals remain among the most trusted professions in the United States and continue to play a critical role in improving childhood vaccine confidence.³²  Although strong evidence for the most effective methods to address vaccine hesitancy is lacking, there is evidence that a strong presumptive recommendation and effective communication from a trusted healthcare professional, standing orders, provider reminders, patient reminder or recall systems, and strong preschool and school vaccination requirements all promote vaccine uptake.²⁸  Although there is some evidence for the effectiveness of public vaccine confidence messaging through mediums such as social media,^33,34  more work needs to be done to determine the best messaging approaches to increase childhood vaccine confidence.²⁸  Furthermore, efforts to improve childhood vaccine uptake following the COVID-19 pandemic should also focus on improving access and addressing racial, ethnic, and socioeconomic disparities in childhood vaccine uptake. This includes ensuring equitable access to all childhood vaccines, reducing out-of-pocket costs, and improving trust in healthcare and public health professionals within communities who have experienced extensive histories of medical mistreatment. Finally, vaccine hesitancy is an ongoing, evolving, and heterogenous public health threat, yet the United States lacks robust vaccine hesitancy surveillance mechanisms. Efforts should be made to develop surveillance tools to monitor and understand vaccine hesitancy to inform work aimed at improving vaccine confidence and improve progress toward vaccine equity.

In this study, the COVID-19 pandemic was not associated with a change in the proportion of parents who were hesitant toward routine childhood vaccines. However, the COVID-19 pandemic was associated with decreased trust in vaccine information and a polarization of vaccine attitudes. Additionally, there were notable differences between sociodemographic factors including race, preferred language, health insurance status, and education level. Thus, to improve childhood vaccine uptake following the COVID-19 pandemic, it remains important to address ongoing vaccine hesitancy using evidence-based approaches and address significant vaccination disparities.

One of the authors of this study (David M. Higgins, MD, MPH) was supported in part by the Health Resources and Services Administration of the US Department of Health and Human Services under grant number D33HP31669. Survey data were supplied by the Center for Health and Environmental Data of the Colorado Department of Public Health and Environment, which specifically disclaims responsibility for any analyses, interpretations, or conclusions which it has not provided.

COVID-19

coronavirus disease 2019

PACV5

Parent Attitudes about Childhood Vaccines short scale