Racial and Ethnic Disparities in Adult Vaccination: A Review of the State of Evidence

Background Adult vaccination coverage remains low in the United States, particularly among racial and ethnic minority populations. Objective To conduct a comprehensive literature review of research studies assessing racial and ethnic disparities in adult vaccination. Search Methods We conducted a search of PubMed, Cochrane Library, ClinicalTrials.gov, and reference lists of relevant articles. Selection Criteria Research studies were eligible for inclusion if they met the following criteria: (1) study based in the United States, (2) evaluated receipt of routine immunizations in adult populations, (3) used within-study comparison of race/ethnic groups, and (4) eligible for at least one author-defined PICO (patient, intervention, comparison, and outcome) question. Data Collection and Analysis Preliminary abstract review was conducted by two authors. Following complete abstraction of articles using a standardized template, abstraction notes and determinations were reviewed by all authors; disagreements regarding article inclusion/exclusion were resolved by majority rule. The Social Ecological Model framework was used to complete a narrative review of observational studies to summarize factors associated with disparities; a systematic review was used to evaluate eligible intervention studies. Results Ninety-five studies were included in the final analysis and summarized qualitatively within two main topic areas: (1) factors associated with documented racial-ethnic disparities in adult vaccination and (2) interventions aimed to reduce disparities or to improve vaccination coverage among racial-ethnic minority groups. Of the 12 included intervention studies, only 3 studies provided direct evidence and were of Level II, fair quality; the remaining 9 studies met the criteria for indirect evidence (Level I or II, fair or poor quality). Conclusions A considerable amount of observational research evaluating factors associated with racial and ethnic disparities in adult vaccination is available. However, intervention studies aimed at reducing these disparities are limited, are of poor quality, and insufficiently address known reasons for low vaccination uptake among racial and ethnic minority adults.


Introduction
Immunizations are a safe and cost-effective strategy to reduce morbidity and mortality resulting from vaccine preventable illness. Despite the proven effectiveness of vaccination, adult immunization rates remain low in the United States. The Centers for Disease Control and Prevention (CDC) estimates that among U.S. adults, there are roughly 4,000 deaths attributable to invasive pneumococcal disease and between 3,000 and 49,000 deaths due to seasonal influenza annually. 1 Racial and ethnic minority adults have higher hospitalization rates due to influenza and lower vaccination coverage than non-Hispanic White (White) adults.
Analysis of seasonal influenza data from 2009-2010 through 2018-2019 shows that the age-adjusted influenza hospitalization rates per 100,000 population among non-Hispanic Blacks (Black) is 68.1 and American Indian/Alaska Natives is 47.5 compared with 38.3 in White adults. 2,3 Notably, Black (39.0%), Hispanic (37.5%), and individuals who identify as other or multiple race (41.4%) have persistently lower influenza vaccination coverage when compared with White (49.3%) adults. 4 A diverse set of barriers contributes to low immunization uptake among the adult population, including lack of physician recommendations for vaccination, misconceptions about vaccination needs, lack of insurance coverage, and incomplete use of evidence-based strategies like standing orders or reminder-recall systems to increase routine adult vaccination. 5,6 Suboptimal vaccination coverage among adults is further compounded by longstanding racial and ethnic disparities in vaccination uptake, reflecting stark health disparities and leading to a disproportionate burden of vaccine preventable diseases among several racial and ethnic minority groups. 7 Social and health inequities, such as poverty and limited health care access, are multifaceted and interrelated issues that influence a wide range of health and quality-of-life outcomes. 8 In addition, distrust of the health care system and history of discrimination in medical research among communities of color are barriers to those seeking health care when needed. 4 It is important to understand and analyze these barriers and challenges to make progress toward overcoming health disparities in adult vaccination.
The purpose of this literature review is to highlight dynamic interactions at the individual, social, and environmental level, and summarize how they influence racial and ethnic disparities in adult immunization. The Social Ecological Model (SEM) is used as a guiding framework to classify factors associated with vaccination disparities. Also, intervention strategies to reduce racial and ethnic disparities in adult immunization are systematically evaluated.

Search strategy and selection criteria
We conducted a search in PubMed, Cochrane Library, and ClinicalTrials.gov using the following search string: (race OR ethnic* OR racial) AND (disparit* OR differen* OR gap) AND (vaccin* OR immuniz*) using the filter ''Adult: 19 + years.'' Studies in languages other than English were not included and publication year was restricted to 2000-2020. The search was conducted in June 2020 and returned 969 Englishlanguage abstracts; all articles were downloaded into EndNote X9.3.3 (Philadelphia, PA).
Preliminary abstract review was conducted by two authors (M.C.L. and C.J.G.), with full article reviews completed by at least one author to identify relevant studies for inclusion. Following complete abstraction of articles using a standardized template, abstraction notes and determinations were reviewed by all authors; disagreements regarding article inclusion/exclusion as well as SEM categorization were resolved by majority rule. The references of all included articles were reviewed to identify relevant articles not found in the initial literature search.
Of the 969 abstracts returned by our search, 758 were excluded after preliminary review for the following reasons: no abstract; duplicate abstract; epidemiological study (e.g., studies estimating prevalence, morbidity, and mortality); unlicensed (e.g., clinical trials of HIV vaccines) or nonroutine vaccines recommended for travel or limited at-risk groups (e.g., cholera, anthrax, smallpox, etc.); methodological study; no within-study comparison of different race/ethnic groups; not a study of differences in vaccine receipt (e.g., studies of vaccine-preventable disease, immunogenicity, or nonvaccination health behaviors); adults are not the subject of the study (e.g., studies of parents' willingness to vaccinate their children) or separate data on adults was not available; study not based in the United States; and study did not report original data (e.g., review articles and multiple publications of the same analysis). Many excluded abstracts fell into multiple exclusion categories.
The full text of 211 abstracts was obtained and assessed for eligibility. After review, 132 additional studies were excluded: 95 studies not evaluating factors associated with documented disparities, and 37 studies excluded for the reasons noted above. An additional three intervention studies were ineligible for the PICO (patient, intervention, comparison, and outcome) questions generated specifically for the intervention portion of the systemic review. Another 16 studies not captured by the original search were identified by the authors through manual bibliography review of all articles that met study inclusion criteria following full assessment.
As a result, 95 studies were included in the final analysis and summarized qualitatively within 2 main topic areas: (1) factors associated with documented racialethnic disparities in adult vaccination and (2) interventions aimed to reduce disparities or to improve vaccination coverage among racial-ethnic minority groups (Fig. 1). 9

Conceptual framework
The SEM, based on Brofenbrenner's ecological theory of development, which posits that behavior is influenced by multiple individual and environmental factors, 10-12 is frequently used as a framework for prevention in health care. [13][14][15] For this review, the authors applied an adjusted model that defines four interaction levels: individual, interpersonal, community, and environmental. 16 The individual level comprises a person's attitudes, beliefs, and personal characteristics, while the interpersonal level is concerned with how the person interacts with others. The community level examines the settings and institutions in which the person interacts, and the environmental level looks at the high-level context of these interactions, including the policies that may shape behavior.
To organize literature review findings, articles focused on assessing factors associated with racial-ethnic disparities in adult vaccination were categorized using the four SEM levels above. Within each SEM level, articles are further grouped by themes regarding personal characteristics that influence vaccination decision making, interactions with health care and/or health care providers, characteristics of health care settings, and cultural or physical structural barriers to receiving vaccination services.

Systematic review of intervention studies
We used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to examine studies evaluating interventions to reduce adult vaccination disparities or improve vaccination uptake among racial-ethnic minority groups. 9 Included studies assessed immunization of U.S. adult populations 19 years of age or older using either single-level or multilevel interventions for any adult vaccination recommended by CDCs Advisory Committee on Immunization Practices, 17 as well as reported vaccination uptake or changes in uptake over time by race/ ethnicity.
Articles were not restricted by study type (e.g., randomized controlled trials, cohort studies, or longitudinal studies), intervention type, population evaluated, or whether a priori intention to reduce racial and ethnic disparities in adult vaccination was evaluated.
Included articles were categorized into three intervention-specific PICO questions ( Table 1). The PICO questions were designed with the aim of including racial-ethnic minority and nonminority participants eligible for vaccination and measuring disparities either through vaccination uptake or difference in vaccination uptake.
Two coauthors (T.C.J. and A.A.) independently reviewed each target article and abstracted the following information: authors, year of publication, funding source, study design and setting, timing, study population, vaccine of interest, intervention, comparison groups, outcomes, results, strengths and weaknesses of study, and level of evidence. To assess risk of bias, we independently graded each study according to the U.S. Preventive Services Task Force system to evaluate quality factors, including study design and methodological features such as assessment of potential confounders and outcome ascertainment. 18 Due to the heterogeneity of study designs and reporting of outcomes, we did not analyze data quantitatively or compute summary measures.

Reasons and associations
The longstanding problem of racial-ethnic disparities in adult vaccination coverage is multifactorial. This section of the review discusses broad categories of factors that research suggests may be causally associated with differences in vaccination coverage by race/ ethnicity, grouped by the SEM level at which they occur (Fig. 2).  [19][20][21][22][23][24][25] ; however, a minority also assess understanding the need for vaccination 26 and vaccine recommendations. 27,28 Studies have found both Hispanics and Blacks are less likely to be knowledgeable about HPV vaccine and/or disease, 29,30 shingles vaccination, 31  disease. 27,32 Measured differences in general health literacy by race/ethnicity have also been associated with lower vaccine uptake among younger and older adults, [33][34][35] although some studies find no effect of health literacy on vaccination disparities. 36 Attitudes and beliefs regarding vaccination, including believing vaccines may cause harm, also contribute to observed racial-ethnic disparities in vaccination. [37][38][39][40][41][42][43][44][45][46][47][48] Black adults have been shown to prefer social distancing, hand hygiene, or culturally specific home remedies in lieu of vaccination to prevent influenza. 44,46 Differences in insurance status may mediate disparities in vaccination coverage, although available evidence is conflicting 14,49-55 and racial-ethnic differences in vaccination persist among insured persons. 50 Similarly, findings about how health care utilization affects vaccination disparities are mixed 56,57 ; differences in utilization by race/ethnicity might reflect access barriers or other individual-level factors. All of these elements may contribute to lower vaccine-seeking behavior [58][59][60] or vaccine acceptance 60,61 by racial-ethnic minority adults compared with White adults.

Interpersonal
Differences in receipt of provider recommendation for vaccination by race/ethnicity have been observed, [62][63][64][65] yet accounting for these differences does not fully eliminate disparities. [66][67][68] Provider recommendations can mitigate lower tendencies to initiate vaccination in racial-ethnic minority adults, 64,67 although recommendations from a provider of the same race/ethnicity as the patient may carry more weight. 69 The relationship of provider recommendations on vaccination uptake is further complicated by racialethnic differences in trust in providers. Studies suggest equal or greater trust by Hispanic relative to White patients, 70,71 but lower trust by Black patients. 69,70 Perceived discrimination among racial and ethnic minority groups is complex; research evaluating the impact of overt or subversive discrimination is limited. [72][73][74][75] In one study, Black adults reporting health care stereotype threats from their provider were less likely to receive influenza vaccine, 72 while another found racial consciousness and discrimination experienced within the health care setting were negatively associated with vaccine-seeking behavior. 74 Potential differences in quality of care by patient race/ethnicity may be mitigated by greater contact with the health care system: disparities in influenza vaccination were lower among Hispanic adults reporting diabetes self-management education, 76 and observed vaccination disparities among Medicare enrollees by race/ethnicity decreased with increasing numbers of health conditions. 77 Some observed disparities may be accounted for by racial-ethnic differences in self-reporting vaccination, although findings are inconsistent regarding the direction of the bias. [78][79][80] Community Ample evidence indicates the demographic composition of health care settings for both residential [81][82][83][84][85][86] and acute 87-89 care facilities substantially affects likelihood of vaccination. Studies consistently find patients in facilities serving higher proportions of racial-ethnic minority patients are less likely to be appropriately vaccinated [84][85][86][87] ; such variability often compounds lower vaccination of minority patients within these facilities. 87 The mechanism driving this association is unclear, although some studies postulate it is driven, in part, by differential use of vaccination promotion strategies such as standing orders. 83,89 The availability of vaccination services in different settings in the community may also influence or widen disparities through use of nonmedical vaccination settings. Specifically, studies evaluating the use of pharmacies as alternative vaccination sites by racialethnic minority patients suggest Black adults have lower acceptance of receiving vaccinations within this setting compared with White patients, with findings for Hispanic adults conflicting. 90,91 Environmental The context in which vaccination occurs, including cultural and linguistic factors, can influence vaccination disparities. Among Hispanic adults, Spanish-preferring adults are less likely than both White and Englishpreferring Hispanic adults to receive influenza 92,93 and pneumococcal 92 vaccinations. Notably, one study showed Spanish-preferring adults in longstanding Hispanic communities or areas of low linguistic isolation had the highest vaccination coverage of all three groups, suggesting the value of social support. 92 Multiple studies also indicate lower knowledge about, interest in, and uptake of vaccination among foreignborn compared with U.S.-born racial-ethnic minority adults [94][95][96] ; nativity may be a proxy for limited English proficiency or may signal cultural differences in perceived value of vaccination.
Findings regarding whether health care resource availability, which is associated with vaccination uptake, fully accounts for observed vaccination disparities between racial-ethnic minority and White adults are inconclusive. 97,98 Racial-ethnic minority adults may be more affected than White adults by limited physician or vaccine supply, 98,99 which has particular implications for racial-ethnic minority adults in rural areas. 98,100 Interventions Our search strategy identified 12 intervention studies that addressed 3 intervention-specific PICO questions ( Table 2). Three studies (Level II, fair quality) addressed the first PICO question and met inclusion criteria for direct evidence. [101][102][103] The remaining nine studies answered the third PICO question meeting criteria for indirect evidence (Level I or II, fair or poor quality). [104][105][106][107][108][109][110][111][112] None of the studies fulfilling the inclusion criteria for interventions addressed the second PICO question.
Three prospective studies provided direct evidence and examined the impact of multilevel interventions on adult vaccination disparities in pneumococcal and/ or influenza vaccination coverage. [101][102][103] Each of the three studies examined adult vaccination rates before and after implementing multilevel interventions using a comparison group; two of these studies demonstrated the interventions improved pneumococcal and/or influenza vaccination rates among racial-ethnic minority groups when compared with the White adult group. 101,102 Details regarding the multilevel interventions implemented and evaluated in each of these two studies were not further described.
The third study demonstrated effectiveness of the intervention that comprised countywide large-scale vaccine distribution and public education activities for H1N1 vaccination in White and racial-ethnic minority participants, but its effect on vaccination disparities was not examined. 103 Therefore, the authors concluded that the increase in vaccination rates was not commensurate with their respective population estimates, highlighting any improvement was inadequate.
Of the nine studies that examined adult vaccination rates after, and not before, an intervention by race/ethnicity, three implemented multilevel interventions, 105,111,112 two implemented provider-specific interventions, 107,110 and four studies implemented system-specific interventions. 104,106,108,109 All three studies implementing multilevel interventions observed a significant increase in vaccination rates among racial and ethnic minority groups. However, the effect of these interventions varied by race/ethnicity, with multilevel interventions more effective for increasing influenza vaccination in Puerto Rican adults (no comparison to White adults reported) 105 and hepatitis A and B vaccination series completion in unhoused Black adult populations compared with Whites. 111,112 Similarly, a study focusing on provider-specific interventions demonstrated that Black race was a significant predictor of pneumococcal vaccine refusal when compared with Asians and Whites 107 ; however, a similar study showed no difference in odds of influenza vaccine acceptance among White and Black populations. 110 The remaining four studies utilizing system-specific interventions, including reminderrecall and standardized vaccination protocols, did not demonstrate increases in the likelihood of getting vaccinated or the overall vaccination rates among racial and ethnic minorities. 104,106,108,109 The strengths of the fair quality studies providing direct evidence and that address PICO question 1 include the following: conducted at multiple sites, implemented multilevel interventions in different racial and ethnic minority groups, and analyzed and reported results for each minority group. These studies also used medical records to assess vaccination status at baseline and post-intervention.
Notably, some studies were not designed a priori to evaluate racial disparities in vaccination and did not clearly describe the multilevel interventions implemented. Furthermore, indirect studies that answered PICO question 3 were found to be poor quality evidence, with included interventions conducted at a single site and used self-report to document vaccination status. Overall included studies evaluating interventions to reduce racial and ethnic minority disparities in adult vaccination are insufficient and do not adequately address system-specific or provider-specific interventions.

Discussion
A considerable amount of research regarding racial and ethnic disparities in adult vaccination has been published over the last two decades, with an overwhelmingly large proportion of studies focused on observational findings and a more limited set concentrated on addressing observed disparities. Many studies highlight individual-level factors associated with vaccination disparities; however, factors at the interpersonal, community, and environmental levels are  Influenza vaccine. Implemented the following activities: (1) Patient tracking to identify eligible patients and monitor vaccination status.
(2) Provider reminders included in medical charts.
(3) Educational sessions for clinical staff.     (1) Create regional advisory group to provide support and advocate for flu activities and access.
(2  simultaneously present and likely contribute to observed differences. The limited research evaluating interventions aimed at addressing these observed disparities further underscores the lack of progress in identifying and implementing actionable strategies to advance health equity in adult immunization. Determinants of racial and ethnic disparities in adult vaccination are multifactorial; however, most available research is predominantly centered on factors that occur at the individual level. Application of the SEM in prevention research is widely accepted; a systematic review of multilevel prevention models for vaccination and screening found application of this model to be effective in determining vaccination behaviors when a multisystem approach was used. 113 Similarly, a comprehensive study of influenza vaccination uptake concluded variables at the individual, interpersonal, and institutional levels had the greatest impact on influenza vaccination uptake. 14 In this review, we identified numerous studies that suggest characteristics at every SEM level are associated with low vaccination coverage in racial and ethnic minority populations. Although actions to increase vaccine confidence and coverage are currently being implemented at the federal level, additional stakeholder efforts that address barriers at multiple levels of the SEM are needed to make any tangible progress in addressing these disparities.
Multilevel interventions addressing multiple SEM levels have been shown to be effective in reducing adult vaccination disparities. 114 In this review, most of the included intervention studies with fair quality evidence demonstrated effectiveness in increasing vaccination rates among racial and ethnic minority groups through use of multilevel interventions. These multilevel interventions include strategies that improve vaccine demand through patient education activities and health care provider interactions, and ensure equitable vaccine access in health care settings.
The Community Preventive Services Taskforce also recommends implementing two or more interventions that focus on increasing community demand and/or access to vaccinations to improve vaccination rates in targeted populations and reduce disparities. 115 However, only a limited number of included studies implemented provider-level interventions focusing on health care provider recommendations and hospitalbased strategies. In addition, very few studies evaluated community-or environmental-level strategies such as use of extended clinical hours 116 or pharmacy-based initiatives. 117 These findings underscore the need for additional research to evaluate interventions that address known factors associated with immunization disparities in racial and ethnic minority groups to develop meaningful strategies for increasing vaccine equity across various traditional and nontraditional health care settings.
One of the primary factors identified by this review as driving racial and ethnic vaccination disparities at the individual level was patient attitudes toward vaccination, in particular, the belief that vaccines are harmful. The higher prevalence of these attitudes in racial and ethnic minority adults relative to White adults could be due to differences in risk tolerance by race or ethnicity, such that avoiding the relatively remote chance of a serious adverse event following immunization is valued more highly than the projected benefits of vaccination.
Mistrust of health care providers and conventional medicine in general may influence these attitudes: studies suggest medical mistrust is higher among racial and ethnic minority adults, particularly Black adults, and higher medical mistrust is associated with reduced intent to receive vaccines. 118,119 However, medical mistrust is an incomplete explanation that ignores perceptions and experiences of racism in medical settings [120][121][122] ; research suggests that physician communication with Black adults may be less patient centered or engaged. 123 Addressing health care provider communication with racial and ethnic minority patients, as well as increasing racial concordance and medical workforce diversity, 124 may reduce vaccine hesitancy in these populations.
This literature review has several limitations. First, the literature search string was not inclusive of additional health equity terms such as ''inequities'' or ''inequalities,'' which may have introduced study imprecision. Although the authors conducted a manual bibliography review of all articles meeting inclusion criteria, omission of these terms may have resulted in incomplete capture of available research. Second, assessment of within-population heterogeneity as factor impacting vaccination disparities was not possible due to most included studies using broad race/ethnicity definitions. Finally, the themes described in each SEM level are reflective of available research only; therefore, additional interactions relevant to vaccine decision making and practices among racial and ethnic populations are not discussed.

Conclusions
This literature review presents a comprehensive examination of published racial and ethnic disparity research in adult vaccination over the past 20 years. Although we identified few studies focused on how to decrease racial and ethnic disparities in adult vaccination, our findings suggest broader use of multilevel interventions addressing patient concerns, as well as structural and contextual barriers to access, may help increase routine vaccination uptake in communities of color. Implementation of this type of intervention at the scale needed to evaluate and demonstrate change will likely require dedicated resources and focus from federal, state, and local public health officials, health care providers, and community partners.
Importantly, such activities should be guided by early and ongoing investment from the affected community to maximize the benefits of the intervention. 125 Only through these concerted efforts and partnerships can tangible advancement in reducing longstanding adult vaccination inequities resulting from a national legacy of structural racism be accomplished.
Authors' Contributions M.C.L. is responsible for conceptualizing the literature review, completing the abstract review, completing article review of include observational studies, supporting manual review of additional articles, and drafting the final article; C.J.G. is responsible for completing the abstract review, completing the article review of included observational studies, leading the manual bibliography review, leading the narrative review analysis, and drafting the final article; T.J. is responsible for completing the article review of included interventions studies, supporting manual review of additional articles, leading the systematic review analysis, and drafting the final article.
A.F.A. is responsible for completing the article review of included interventions studies, supporting the systematic review portion of the review, and drafting the final article; and K.J.-J. is responsible for supporting manual review of additional articles and for final review of the article.