Vaccine Coverage at 36 Months and 7 Years by Parental Birth Country, Washington State Free

We conducted a retrospective cohort evaluation of Washington-born children to compare vaccination coverage between children of US-born parents and children of at least 1 parent born in selected countries outside the United States. We assessed vaccination outcomes at age 36 months, when key developmental milestones are expected to have been met, and 7 years, when Washington school-entry immunization requirements should have been met.

The WAIIS is a mature registry that tracks immunizations for residents over their lifespan and is populated with birth information from Washington Center for Health Statistics Birth Certificate Records (WABC). Children born in Washington from January 1, 2006 through November 12, 2019 with matching WABC and WAIIS records were included. We excluded children with a deceased status or out-of-state residence at the age of vaccine assessment in accordance with Washington State Department of Health Office of Immunization coverage assessment standards. We allowed for a 1-month lag in provider vaccine reporting from the date of WAIIS data query, which was December 13, 2022.

WABC data are captured from the mother, providers, or medical record. From WABC, we abstracted name, birth certificate ID, birthdate, mother’s and father’s birth country, number of prenatal visits, mother’s reported education level, mother’s reported receipt of Women, Infants, and Children (WIC) benefits during pregnancy, and payment source for delivery. From the WAIIS, we abstracted name, birth certificate ID, birthdate, sex, race, ethnicity, current and historical states of residence, record update date, vaccine type, and vaccination date. Race and ethnicity are required fields for provider reporting; however, providers may report them as unknown. To assign categories, Hispanic ethnicity was assessed first. For those not identified as Hispanic, racial groups were assigned as follows: American Indian or Alaska Native, Asian, Black, multiracial, Native Hawaiian or other Pacific Islander, white, other, and unknown. All valid vaccine doses containing measles, mumps, and rubella (MMR), diphtheria, tetanus, and pertussis (DTaP), and poliovirus antigens were included.

The study exposures were Washington births to ≥1 parents born in the following countries: Mexico, India, Ukraine, Philippines, Somalia, Russia, Moldova, Romania, Marshall Islands, and Micronesia. We included the top 4 non-US PBCs during the evaluation period and the top 6 non-US countries with evidence of vaccine barriers. The latter criteria were determined through analysis of coverage with ≥1 doses of MMR among the top 30 PBCs with >1500 births; countries with <80% coverage were included.

To avoid counting children more than once, country categories were classified hierarchically on the basis of a priori hypotheses of MMR vaccine hesitancy (Ukraine > Russia > Moldova > Somalia > Marshall Islands > Micronesia > Romania > Mexico > India > Philippines). For example, a child with 1 parent from Ukraine and 1 parent from Moldova was considered part of the Ukraine group for the purposes of our study. Although many children had high PBC concordance between parents, we conducted sensitivity analyses of PBC classification by checking for differences in vaccination trends using alternate hierarchies, concordant classifications, and non-mutually exclusive classifications. These methods did not meaningfully change results for any vaccine. The reference group included children with 2 US-born parents.

Up-to-date vaccination coverage status was assessed. The primary outcome was receipt of MMR vaccine at 36 months. Secondary outcomes were receipt of DTaP and poliovirus vaccines at 36 months and receipt of each vaccine at 7 years. A child was considered up-to-date for a specific vaccine if they had received valid doses according to minimum ages and between-dose time intervals per Advisory Committee on Immunization Practices recommendations³¹  (Table 1). Vaccination status was assessed at age points to allow equal opportunity for vaccination and increase the sample size by PBC. We chose these vaccines because they are required for Washington school entry and offer protection against conditions with potential for outbreaks. Additionally, prevalent misinformation surrounding MMR vaccine effects on childhood development³²  and the recent detection of poliovirus in the United States for the first time in nearly 10 years³³  underscores the importance of understanding country-level immunization trends.

To understand country-level trends representing potential vaccine delay versus refusal, we conducted a sub-analysis among children eligible at both 36 months and 7 years and classified children as follows to determine changes in MMR vaccination status:

• No documented doses by 7 years old;

• One documented dose by 36 months;

• Receipt of both doses by 36 months or second dose between 37 months and 7 years;

• Receipt of first dose between 37 months and 7 years;

• Receipt of both doses between 37 months and 7 years.

We included the following covariates in the multivariable models to control for socioeconomic status and health care utilization markers:

• Maternal education: high school or less, some college or associate degree, or bachelor’s degree or higher;

• Number of prenatal visits: 0, 1 to 9 or ≥10 visits per the American College of Obstetrics and Gynecology’s recommendations;³⁴ 

• WIC benefits receipt during pregnancy: receipt versus nonreceipt;

• Insurance status: Medicaid, private (self-pay/health maintenance organization/commercial), and charity care.

We linked available WAIIS and WABC records by birth certificate ID. Because of the potential for WAIIS records to be merged or fragmented over time, matches were verified by name and birthdate by using iterative deterministic and fuzzy linkage techniques in SAS Version 9.4.³⁵ 

Descriptive analytic methods were used to quantify the relationship between PBC and vaccine receipt overall and by year of vaccination assessment. Children born in 2017 through 2019 were recommended for vaccination at 36 months just before or during the pandemic era.

We used Poisson regression with robust standard errors to quantify the relationship between PBC and vaccine receipt by utilizing prevalence ratios (PRs) and corresponding 95% confidence intervals (CIs). To determine PBC effects on vaccine receipt over time, we stratified the analysis by vaccine assessment year at 36 months. Data were analyzed by using SAS Version 9.4.

The Washington Department of Social and Health Services Institutional Review Board granted ethical approval.

A total of 902 909 children and 636 746 children met the evaluation inclusion criteria by ages 36 months and 7 years, respectively. Of these, 24% of children (N = 219 378) had at least 1 non-US-born parent (Table 2). Differences by PBC in sex, race, ethnicity, maternal education level, WIC enrollment, prenatal visits, and payment source for delivery are described in Table 2.

Vaccination coverage at 36 months varied widely by PBC ranging from 41.0% to 93.2% for ≥1 doses of MMR and ≥3 doses of poliovirus, and 32.6% to 86.4% for ≥4 doses of DTaP (Table 3). After adjusting for confounding factors, compared with children of US-born parents, MMR vaccination coverage at 36 months was 3% to 10% higher among children of Indian-, Filipino-, and Mexican-born parent(s), 8% to 22% lower among children of Micronesian-, Marshallese-, Romanian-, and Somali-born parent(s), and 35% to 51% lower among children of Moldovan-, Russian-, and Ukrainian-born parent(s) (Table 4).

Patterns of coverage within PBCs were similar for poliovirus and DTaP with some exceptions. Among children of Somali-born parent(s), poliovirus coverage at 36 months was 2% higher (PR, 95% CI: 1.02, 1.01–1.03), whereas DTaP coverage was just 4% lower (PR, 95% CI: 0.96, 0.95–0.98, P <.001), relative to children of US-born parents (Table 4). Children of Marshallese- and Micronesian-born parent(s) had lower coverage for all 3 vaccines relative to children of US-born parents, but the magnitude of difference was largest for DTaP (Table 4).

Vaccination coverage at 36 months declined from 2009 to 2022 among several communities, with declines ranging from 13.0 to 47.2 percentage points among children of Marshallese-, Russian-, Romanian-, Moldovan-, Ukrainian-, and Somali-born parent(s) (Fig 1A). Children of Mexican- and Filipino-born parent(s) had stable or increasing MMR coverage at 36 months from 2009 through 2019 but declining coverage from 2020 through 2022 that ranged from 2.6 to 4.1 percentage points (Fig 1A). Similar but less pronounced country-level trends were observed for poliovirus and DTaP (Fig 1B and 1C, respectively). Among children of Moldovan-, Romanian-, Russian-, Somali-, and Ukrainian-born parent(s), the decline in MMR coverage across vaccine assessment years relative to children of US-born parents became more pronounced over time. From 2009 through 2012, MMR coverage among children of Somali-born parent(s) was higher or comparable relative to children of US-born parents, whereas coverage among children of Moldovan-, Romanian-, Russian-, and Ukrainian-born parent(s) was lower across all birth cohorts. Among the latter PBCs, similar trends were observed for poliovirus and DTaP. Among children of Somali-born parent(s), disparities in DTaP coverage relative to children of US-born parents started in a later cohort (2013) and were much less pronounced than those for MMR (Supplemental Table 7).

We observed similar country-level trends for up-to-date vaccination status at 7 years (Tables 5 and 6); however, for children of Somali-born parent(s), the magnitude of difference for MMR coverage relative to children of US-born parents was less pronounced than at 36 months (PR, 95% CI: 0.88, 0.87–0.90, P < .001 vs 0.78, 0.76–0.79, P < .001, respectively; Tables 4 and 6).

Of the 636 746 children who were eligible for evaluation at 7 years, 99% (N = 633 534) were also eligible for analysis at 36 months. Among children of Somali-born parent(s), 15.1% of children received their first dose of MMR vaccine after 36 months and just 14.1% remained unvaccinated at 7 years. Conversely, whereas 12.4% of children of Ukrainian-born parent(s) received their first dose after 36 months old, 41.3% of children remained unvaccinated at 7 years (Fig 2).

In this large evaluation of children born in Washington to immigrant parent(s) spanning several birth cohorts, vaccination coverage at 36 months and 7 years varied widely by PBC, and in many communities, were below levels necessary for herd immunity. This underscores the need for collection and reporting of health outcomes by nativity and other factors that reflect the diverse, lived experiences of local communities. In our evaluation, children of Ukrainian-, Moldovan-, and Russian-born parent(s) had the lowest coverage for all 3 vaccines at 36 months and 7 years, suggesting general parental vaccine concerns. Focus groups among Russian- and Ukrainian-speaking immigrants identified vaccination barriers, including concerns about adverse events, distrust of medical and governmental systems, misinformation spread through social channels, and health care dissatisfaction stemming from language and cultural barriers.³⁶  Similar but less pronounced trends reflecting parental vaccine concerns were observed among children of Romanian-born parent(s). Interviews with Romanian parents in England revealed barriers including a lack of translated vaccination information, difficulties navigating health care systems, and differing health services expectations.^37 ,38  Whether trends among Washington-born children of Romanian-born parent(s) reflect similar barriers requires further understanding.

Disparities in vaccination coverage among Washington-born children of Somali-born parent(s) were unique to MMR vaccine, consistent with previous trends observed among Somali children in Minnesota, Washington,^9 ,20 ,39  and Norway.⁴⁰  Historically, MMR vaccination coverage among Washington-born children of Somali-born parent(s) was high and exceeded that for children of US-born parents. A marked decline in coverage at age 36 months began in 2013, shortly after the Somali community in Minnesota was targeted by antivaccine activists, including Andrew Wakefield,⁴¹  who spread misinformation from his retracted article asserting unsupported associations between MMR vaccine and autism.^42 ,43  Because social networks influence opinions on vaccination^44 –46  and Somali mothers may rely on vaccination experiences of friends and relatives,^47 ,48  it is possible the Washington trend represents the adoption of broader Somali diaspora-based concerns about MMR and autism.⁴⁹  A substantial proportion of Washington-born children of Somali-born parent(s) received ≥1 dose(s) of MMR after age 36 months, which may reflect parental decisions to delay MMR vaccination until after developmental milestones have been met. This is consistent with qualitative interview findings that Somali mothers have unique concerns about the association between vaccination and child development^47 ,48 ,50  and can support clinician approaches to vaccine counseling with Somali parents.

Children of Marshallese- and Micronesian-born parent(s) also experienced lower vaccine coverage compared with children of US-born parents, but the magnitude of difference was lower than that for the aforementioned communities. The authors of previous studies of health outcomes among these communities have documented barriers including language, transportation, health care system navigation barriers, differences in preventive care utilization, and perceived discrimination by health professionals.^51 –53  Trust in US medical and government systems may be seriously undermined because these communities continue to suffer the lingering impact of the unconsented US nuclear testing program conducted across the islands between 1946 and 1958.^54 ,55 

Children of Filipino-, Mexican-, and Indian-born parents had higher vaccine coverage compared with children of US-born parents. Facilitators for increased vaccine coverage among these communities are scant in the published literature and likely warrant further research.

Vaccination disparities are widening in certain communities with steady, precipitous declines in coverage at 36 months from 2009 through 2022 for all 3 vaccines among children of Ukrainian-, Moldovan-, Russian-, and Romanian-born parent(s) and for MMR among children of Somali-born parent(s). Relative to children of US-born parents, whose vaccination coverage rates decreased slightly among children recommended for vaccination during the pandemic era, the differences still became more pronounced among children recommended for vaccination among these communities during the pandemic era. It is unclear whether these trends reflect exacerbation of existing inequities to health care access or the emergence of new concerns associated with childhood vaccinations. The trends add PBC-level characterization of childhood vaccination inequities during the pandemic era to existing reports of inequities by race, ethnicity, and socioeconomic status^{25 ,26 ,29 ,30}  and suggest the need for qualitative research with communities to better understand emerging barriers or concerns.

These observed country-level vaccination trends reflect unique, community-specific determinants of vaccination and reveal underlying parental reasons for obtaining, delaying, declining, or missing their child’s vaccinations. Many local ethnic health boards and stakeholders in Washington have developed best practices to promote holistic health of their local communities.^56 ,57  Low and declining childhood vaccination coverage trends that vary by PBC underscore the importance of community-specific partnerships to build trust and redress vaccine inequities. Community-specific strategies could include linguistically and culturally appropriate vaccine information, linguistically-concordant care, the enlistment of trusted community messengers to educate and leverage social norms around vaccines, and increased opportunities for patient–provider dialog about vaccines.^{39 ,48 ,58 –63} 

There are several limitations to this evaluation. First, PBC is just 1 layer of granularity to better understand community-level vaccination trends. Although we controlled for markers of socioeconomic status and heath care access at the time of birth, there is potential for misreporting of these birth certificate fields. In addition, these and other within-PBC determinants of health, like preferred language and parental reasons for migrating and acculturation, were unavailable at assessment ages. Additionally, we were unable to fully account for out-of-state migration, which could result in denominator inflation and an underestimation of vaccination coverage, a limitation that is ubiquitous across state Immunization Information Systems.⁶⁴  However, compared with historical WAIIS provider organization-level patient status methods, the new WAIIS geographical-level patient status methods are more sensitive for the identification of under-immunized communities with health care barriers.

Future analyses should broaden the scope of PBCs included. In addition, given the gaps in WAIIS race and ethnicity data, the exploration of vaccine coverage rates by race and ethnicity versus PBC could highlight potential masking of trends using broad race and ethnicity categories and help inform future vaccine coverage reporting practices.

In this large, statewide evaluation, the linkage of public health data sources to disaggregate data at the PBC-level resulted in enhanced characterization of childhood immunization outcomes. Many communities are at risk for VPD outbreaks because of low and declining vaccination coverage levels. The PBC-level trends reflect heterogeneity in underlying facilitators and barriers to vaccination and provide actionable information about community-level determinants of vaccination, support tailored clinical care and community-informed interventions, guide decision-making around resource allocation, and strengthen information sharing and proactive partnership development with communities. The collection and reporting of information by preferred language would inform even more specific barriers and assets underlying the immunization trends and tailored interventions to support vaccine coverage.

CI

confidence interval

DTaP

diphtheria-, tetanus-, and pertussis-containing vaccines

MMR

measles-, mumps-, and rubella-containing vaccines

PBC

parental birth country

Poliovirus

poliovirus-containing vaccines

PR

prevalence ratio

VPD

vaccine-preventable disease

WABC

Washington Center for Health Statistics Birth Certificate Records

WAIIS

Washington State Immunization Information System

WIC

Women, Infants, and Children