Parental Vaccine Hesitancy and Vaccination Disparities in a Safety-Net System

We surveyed English- and Spanish-speaking parents bringing children aged 2 years to 3 pediatric clinics for well or sick care from August 2019 through February 2020. Parents were eligible if (1) English or Spanish was their first language, (2) their child turned 2 years in the last 12 months, and (3) the child had 1 previous well visit in the last 18 months. We excluded parents of children with medical contraindications to vaccination. A bilingual research assistant recruited dyads in waiting rooms; parents completed the survey on an iPad and received $5. The Colorado Multiple Institutional Review Board approved this study.

We measured parental VH with the Parent Attitudes about Childhood Vaccines (PACV) tool because of its previous validation studies,^6–8  brevity and simplicity, and rigorous Spanish translation.⁹  We transformed responses into scaled (0–100) and dichotomized (<50 = not hesitant and ≥50 = hesitant) scores, as described in previous validation work.⁷  We collected parent demographics and extracted child vaccination data from electronic medical records. Our independent variable of interest was parental VH. Our outcome variable was child ADU at age 2 years (ie, 730 days), which calculates the difference between when vaccine doses are given and when they are due and ranges from 0 (immunized on time) to 638 (unvaccinated).¹⁰  We compared parents by VH status on independent variables and performed multivariable linear regression analyses to assess parental VH with child ADU associations. For a 2-sided α of .05, power of 0.80, population ADU mean of 36 ± 89,¹⁰  VH effect size of 55 ADU,^8,9  and parental VH prevalence of 10%,^8,9  we estimated needing 259 parents (24 hesitant and 235 not hesitant). Analyses were conducted by using R (R Foundation for Statistical Computing, Vienna, Austria).

Of 293 parents approached, 263 (90%) participated. Two surveys were incomplete, and 6 parents responded twice. In Table 1, we describe our final sample of 255 predominantly Hispanic, English-speaking dyads, stratified by parental VH status. Thirty-three (13%) parents were hesitant; 4 (4%) Spanish-speaking parents were hesitant versus 29 (19%) English-speaking parents (P < .01). PACV item responses also differed by language (Supplemental Table 3). Among all children, the mean ADU at age 2 was 66 days (SD 151). In adjusted analyses, children of English-speaking VH parents had a 114-day greater ADU compared with children of English-speaking non-VH parents (P = .02). Parental VH did not associate with changes in child ADU for Spanish-speaking dyads (P = .61). Children of Black parents (referent to non-Hispanic white) were significantly undervaccinated, independent of parental VH (Table 2).

In a safety-net system sample, we found VH in 1 in 8 parents of children aged 2 years, with differential impacts by language, race, and ethnicity. In previous studies, researchers described VH in wealthier, privately insured clinics; in our study, we found VH in clinics serving minority, publicly insured, and Spanish-speaking children, with a prevalence (13%) similar to that previously reported (8%–25%).^8,9,11  Nearly 1 in 4 Black parents had VH, and Black children aged 2 years were additionally undervaccinated, independent of parental VH. Parental VH should be explored as an important vaccination barrier along with others identified for Black toddlers, such as public insurance, access, distrust of health care, and missed opportunities.^12–15  Parental VH was not associated with ADU changes among Spanish-speaking families, perhaps because of respeto, the cultural value by which Spanish-speaking parents may deferentially subjugate their concerns to clinical expertise.¹⁶  This finding bolsters existing work showing how language and culture impact pediatric care.¹⁷ 

As a single-site retrospective study of parents of children aged 2 years, our conclusions should be explored in other systems and with other age and language groups. Furthermore, trust between staff and patients may have caused a social desirability bias to participate, and we were unable to recruit dyads who failed to present for care. Still, our work highlights the need to explore VH by language and culture and pursue implementation studies to adapt evidence-based interventions for families in safety-net systems.

We acknowledge Jason Glanz, PhD, and the Kaiser Permanente Institute for Health Research for providing us with the algorithm for ADU and instructing us in its use.

ADU

average days undervaccinated

PACV

Parent Attitudes about Childhood Vaccines

VH

vaccine hesitancy