Abstract
Parents who refuse or delay vaccines because of vaccine hesitancy place children at increased risk for vaccine-preventable disease. How parental vaccine hesitancy changes as their children age is not known. In 2015, we conducted a follow-up survey of 237 mothers enrolled in a 2-arm clinic-level cluster randomized trial (n = 488) in Washington State that was completed in 2013. We surveyed mothers at their baby’s birth, age 6 months, and age 24 months using a validated measure of vaccine hesitancy. Both mean hesitancy scores (mean 4.1-point reduction; 95% CI, 2.5-5.6; P = .01) and the proportion of mothers who were vaccine hesitant (9.7% at baseline vs 5.9% at 24 months; P = .01) decreased significantly from child’s birth to age 24 months. Changes from baseline were similar for first-time mothers and experienced mothers. Individual item analysis suggested that the decrease may have been driven by increases in maternal confidence about the safety and efficacy of vaccines. Our results suggest that hesitancy is a dynamic measure that may peak around childbirth and may remit as experience with vaccines accumulates.
Parental refusal or delay of childhood vaccines is an important contributor to underimmunization 1,2 and a growing public health problem. 3 As such, vaccine-hesitant parents constitute an important focus of efforts to improve immunization rates. 4 Surveillance of vaccine hesitancy is an essential step toward these efforts. 5
Because vaccine hesitancy denotes a set of attitudes, beliefs, and intentions about vaccines that falls in the middle of the continuum of vaccine acceptance, 6 –8 it is prone to change over time. 9 Vaccine attitudes specifically can be influenced by ever-changing personal, social-environmental (eg, cultural attitudes and norms), and institutional (eg, legislative mandates) factors. 10 In addition, the longitudinal administration of childhood immunizations during several years provides ample opportunity for these factors to have an effect.
How parental vaccine hesitancy changes as their children age, however, is not known. This knowledge gap indirectly affects the design and timing of interventions focused on vaccine-hesitant parents. The objective of this study was to assess the trend in parental vaccine hesitancy during the first 2 years of their child’s life in a cohort of mothers in Washington State.
Methods
In 2015, we conducted a follow-up survey of participants enrolled in a 2-arm clinic-level cluster randomized trial completed in 2013. 11 During the trial, providers at intervention clinics received training in communicating with vaccine-hesitant mothers, but control clinics received no training. The primary outcome was vaccine hesitancy, measured by the validated Parental Attitudes About Childhood Vaccines (PACV) survey. 12,13 The trial showed no difference in vaccine hesitancy between mothers seeking pediatric care at intervention and control clinics at the child’s age of 6 months. 11
For the current study, we readministered the PACV to mothers when their child was aged 24 months. Surveys were conducted by telephone and administered by trained interviewers; participants were given $10. Consistent with previous studies, the PACV is scored from 0 to 100, with ≥50 considered hesitant. 12,13
After confirming that the proportion of parents who were vaccine hesitant was similar in the control group and intervention group of the original study, we combined these 2 groups into a single cohort. Using a complete-case approach, we examined vaccine hesitancy in mothers who completed PACVs at all 3 time points: baseline (child’s birth), age 6 months, and age 24 months. We used the Friedman test to assess the change in the distribution of vaccine hesitancy of each PACV item at the 3 time points. We used linear regression analysis with generalized estimating equations to estimate the mean change in hesitancy at age 6 months and age 24 months relative to baseline (birth), adjusting for education, race, income, and baseline hesitancy. Models assumed an independent working correlation and accounted for correlation between repeated measures from the same mother. We also fit models with interaction terms to test for differences in changes in PACV scores between first-time mothers and experienced mothers at trial enrollment. To assess the robustness of our results, we performed 2 sensitivity analyses, which included all mothers who completed a baseline survey (n = 391) and used 2 different approaches—last value carried forward and multiple imputation—to handle missing data. We completed all analyses using Stata release 13.1. 14
Results
Of 488 mothers enrolled in the trial, 391 completed the baseline survey and 237 responded to surveys at all 3 time points. Among those who responded to all 3 surveys, the mean age was 32.3 years, 186 (78.5%) were college educated, and 190 (80.5%) were white. Survey respondents were more likely to be married, report higher income, and have higher education compared with all mothers enrolled in the trial. We found no significant difference in the proportion of mothers who were vaccine hesitant at baseline among respondents and nonrespondents to the 24-month survey (9.7% vs 13.6% prevalence of vaccine hesitancy at baseline, respectively; P = .23).
Both mean hesitancy scores (mean 4.1-point reduction; 95% CI, 2.5-5.6; P = .01) and the proportion of mothers who were vaccine hesitant (9.7% at baseline vs 5.9% at 24 months; P = .01) decreased significantly from child’s birth to age 24 months (Table 1). Changes from baseline were similar for first-time mothers and experienced mothers. We found no difference in magnitude, direction, or significance in these results after sensitivity analyses. The proportion of parents who endorsed the more hesitant response to several PACV items decreased over time, but these decreases were not significant (Table 2). These items included concern that a child would have a severe side effect from vaccines (42.2% at baseline to 33.8% at 24 months, P = .12), concern that childhood shots might not be safe (34.2% to 24.9%, P = .26), concern that a vaccine would not prevent the disease (27.4% to 22.8%, P = .13), or agreeing that children get more shots than are good for them (16.5% to 13.1%, P = .07). Agreement that the participants can openly discuss vaccine concerns with their doctor (88.2% to 93.3%, P = .46) and high trust in the child’s doctor (8 to 10 on a 0-10 scale; 92.0% to 92.8%, P = .92) were stable during the 24-month period.
Longitudinal trends in vaccine hesitancy at infant’s birth, age 6 months, and age 24 months in a cohort of mothers enrolled in a cluster randomized trial shortly after childbirth, Washington State, 2013-2015a
Abbreviation: PACV, Parental Attitudes on Childhood Vaccines scale.
aData source: longitudinal survey of mothers (n = 237) conducted during 2013-2015 in western Washington State. Mothers were enrolled shortly after childbirth in hospital birth units and participated in telephone surveys at baby’s age 4 to 6 weeks after birth, baby’s age 6 months, and baby’s age 24 months.
bAdjusted for education, race, and income.
cP value from adjusted linear regression with vaccine hesitancy as the dependent variable and child’s age entered as a linear term.
dThe PACV is scored with a range of 0 to 100. Scores of ≥50 on the PACV are defined as vaccine hesitant.
eChange from baseline is only computed for continuous outcomes.
First-time and experienced mothers’ responses to individual survey items on the Parental Attitudes on Childhood Vaccines (PACV) survey measure (n = 237), Washington State, 2013-2015a
aData source: longitudinal survey of mothers (n = 237) conducted during 2013-2015 in western Washington State. Mothers were enrolled shortly after childbirth in hospital birth units and participated in telephone surveys at baby’s age 4 to 6 weeks after birth, baby’s age 6 months, and baby’s age 24 months.
bNot all mothers responded to all individual survey items; therefore, the number for each response may not sum to 237 for all items.
cFriedman test for differences in distribution of 3-level responses across time points. P < .05 was considered statistically significant.
dScale ranges from 0 (not at all sure) to 10 (completely sure).
eScale ranges from 0 (I do not trust child’s doctor at all) to 10 (I completely trust child’s doctor).
fLimited to mothers who responded to items at baseline (experienced mothers only) (n = 133).
Discussion
We found that mothers’ vaccine hesitancy decreased during the first 2 years of a child’s life. We observed similar trends for both first-time and experienced mothers, a contrast with another cross-sectional study that found more underimmunization in families with >1 child. 15 Responses to individual survey items suggest that the decrease in hesitancy may have been driven by improving maternal confidence in the safety and efficacy of vaccines as their child developed.
Our results suggest that hesitancy may remit over time for many parents as maternal experience with vaccination accumulates. This finding may suggest that pregnancy and soon after birth may be an ideal time to provide mothers with reassuring information and support about early childhood vaccines. 16 Although interventions during pregnancy have primarily focused on maternal vaccination, 17 maternal vaccine acceptance during pregnancy is associated with acceptance of early childhood vaccines. 16
This study is one of the first longitudinal studies to measure vaccine attitudes over time. Limitations included our use of a complete-case approach that excluded mothers who did not complete all surveys. However, sensitivity analyses that included data from all mothers who completed a baseline survey showed similar results. We also found a lower prevalence of vaccine hesitancy at baseline than has been reported elsewhere. 18,19 Replication with maternal populations in other geographic regions would provide further evidence of the generalizability of these findings to other populations.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by the Group Health Foundation. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript.