Search strategy and selection criteria

The search strategies were developed without language restrictions and included the databases of the Medical Literature Analysis and Retrieval System Online (MEDLINE; Ovid), the Excerpta Medica (Embase; Ovid), Virtual Health Library (VHL), Web of Science, Cochrane Library, medRxiv, and PROSPERO from inception to 15 July 2021. We used a search strategy combining terms and synonyms from referenced studies for ‘older adults’, ‘vaccinations’, ‘minority ethnic background’, and ‘views’ (S1 Text). We reviewed reference lists of eligible reports. This study is reported as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (S2 Table). The study protocol is publicly available on PROSPERO (CRD42021237032).

Three independent reviewers systematically screened publications. Studies were eligible if they reported on perceptions, beliefs, and attitudes towards vaccinations in older adults from a minority ethnic background. We included studies from non-White majority populations in their country of origin as these are relevant to the experience of individuals from a minority ethnic background in other settings, since ethnic groups are considered to share a common ancestry, culture, and language [13]. ‘Older adults’ were defined as people aged ≥65 years, consistent with previous studies [11]. Studies of vaccinations in investigation or development, studies focused on specific medical conditions, studies where ethnic background or age group was unidentifiable, systematic reviews, editorials, and conference abstracts were excluded.

Data screening

Records were downloaded into Endnote (version X9), and duplicates were removed. Using a standardised form, 3 independent reviewers (CB, DPG, and UARC) each conducted screening of two-thirds of the total number of titles and abstracts, and full-text articles. The Cohen’s κ statistic using the average across the 3 pairwise combinations of raters addressed inter-rater agreement regarding eligibility. Online systematic review software (Rayyan, QCRI) was used to facilitate literature screening. The publications were initially screened for title and abstract eligibility; full-text articles were then retrieved and screened for eligible publications to be included in data extraction. Discrepancies were resolved through discussion, if necessary with an adjudicator (KW).

Data extraction and analysis

DPG and UARC independently extracted data from each article into a specified data extraction table. Data included study design, analysis method, geographical setting, language, sample size, mean age, ethnic background, sex, vaccination type, views towards vaccinations, and barriers and facilitators influencing vaccination uptake. A consensus meeting was held with CB to finalise data extraction. The risk of bias for studies was assessed by DPG and UARC using the Mixed Methods Appraisal Tool (MMAT), a validated tool for appraising methodological quality for use in systematic mixed study reviews [14]. A convergent synthesis approach informed by Pluye and Hong [15] was used to integrate qualitative and quantitative data. Results from studies that included qualitative, quantitative, and mixed methods data were transformed into qualitative findings using thematic synthesis, and a matrix was built on resultant themes and patterns agreed by CB, UARC, and DPG. Convergent qualitative synthesis seeks to address complex research questions to understand the ‘what, how, and why’ [15] relevant to the study question. Structural determinant categories identified in a previous study examining factors influencing vaccination uptake amongst general older adults [11] were drawn on to group themes. The quality of evidence of collective outcomes was estimated using the Grading of Recommendations Assessment, Development and Evaluation–Confidence in the Evidence from Reviews of Qualitative research (GRADE–CERQual) framework.

Included studies

In total, 3,068 citations were identified by the search, 2,485 citations after duplicates were removed. Following title and abstract screening, 195 potentially eligible articles were retrieved in full text (Fig 1). Of the full-text articles, 167 were excluded: 89 did not examine the intended outcome, 48 were excluded because older adults were unidentifiable in a mixed age population, 12 were excluded because participants were either not from older age groups or not from a minority ethnic background, 7 studies were systematic reviews, and 11 studies were duplicates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

https://doi.org/10.1371/journal.pmed.1003826.g001

A total of 28 relevant studies conducted between 1997 and 2020 were included in the final analysis, of which 17 were quantitative surveys, 8 were focus group or interview studies, 2 were mixed methods studies, and 1 was a case–control study. There was substantial agreement between reviewers at the title and abstract stage (κ = 0.77) and full-text review stage (κ = 0.82) (reported as an average across 3 pairwise combinations of raters at each stage). Meta-analysis was not conducted since there was high heterogeneity across the studies in methods, reporting of outcome, and populations; few papers had sufficient quantitative data for meta-analysis; and this was not an a priori aim.

The MMAT risk of bias and GRADE–CERQual appraisals are summarised in S1 Fig and S1 Table, respectively. The evidence included is supported by high or moderate certainty.

Study characteristics

Study characteristics of the 28 included studies are summarised in Table 1. The majority were US studies in English or Spanish, except for 6 studies set in Hong Kong, 2 studies in Japan, 1 study in Brazil, and 1 multi-centre study (including China, Indonesia, Turkey, South Korea, Europe, Greece, UK, Brazil, and Nigeria). In total, 28,262 individuals with an estimated mean age of 69.8 years were included, of whom 63.2% were female. Six studies explored views in participants from African American backgrounds, 2 studies in participants from Hispanic or Latinx backgrounds, 6 studies in participants from Hong Kong Chinese backgrounds, 2 studies in participants from Japanese backgrounds, 1 study in participants from Brazilian backgrounds, and 12 studies in mixed groups of older adults from minority ethnic backgrounds. Most studies explored views related to either the influenza or pneumococcal vaccine.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of studies.

https://doi.org/10.1371/journal.pmed.1003826.t001

Synthesis

The following sections outline the themes that were identified as barriers to and facilitators of vaccination uptake amongst older adults from minority ethnic backgrounds. Themes were grouped according to previously established structural determinants: (1) healthcare provider and system related, (2) patient related, and (3) policy and operational [11] (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Barriers to and facilitators of vaccination uptake amongst older adults from minority ethnic backgrounds.

SARS, severe acute respiratory syndrome.

https://doi.org/10.1371/journal.pmed.1003826.g002

Barriers to vaccination uptake

Facilitators of vaccination uptake

Comparison to other literature

We have grouped themes under important structural determinants identified by Nagata et al. [11]. Similar sub-themes relevant to older adults from minority ethnic backgrounds emerged in our study: fear and mistrust of modern medicine amongst older adults from African American backgrounds, language and literacy barriers, and cultural beliefs that natural, healthy lifestyles are preferable to vaccinations [11].

A systematic review by Bish et al. [44] examining factors associated with influenza vaccination uptake in the general population found that the degree of fear related to the 2009 influenza pandemic outbreak, positive social pressure, and less fear of side effects correlated with increased intentions of vaccination uptake, consistent with our findings. This suggests these factors are common to the general population. Other factors identified in our study—including lack of information on vaccines, the belief that vaccines cause disease, the perception that vaccines are irrelevant when in good health, and access issues as barriers to vaccination uptake and fear of the disease as a facilitator of vaccination uptake—are likely more common in older adults from minority ethnic backgrounds and important to consider in COVID-19 age-based vaccination policies.

Sheldenkar et al. [45] recently conducted a systematic review of general adult influenza vaccine acceptance in Asia. Similar to our findings, there was a significant lack of studies in South Asia compared to East Asian countries (with Hong Kong having the greatest number of publications). This highlights a pressing gap in the research. South Asians are a sizable ethnic minority in many European countries [46], previous studies have identified vaccine hesitancy amongst Pakistani and Bangladeshi ethnic groups [12], and these groups are at increased risk of COVID-19 mortality [4].

Strengths and limitations

This is the first systematic review to our knowledge seeking to understand perceptions and beliefs influencing vaccination uptake amongst older adults from minority ethnic backgrounds. It provides insights that are pertinent to the COVID-19 pandemic, with vaccination policies being rolled out worldwide. The findings also inform existing national vaccination schedules and future policies in epidemics, to ensure optimal impact and equity. A further key strength of this study is the focus on older adults. Despite being the most vulnerable to severe illness from communicable disease [3], older adults are often excluded from clinical research [47]. The main limitation is the limited number and range of studies available, with the vast majority being set in the US, the findings of which may not be applicable to ethnic minority groups in other settings. We included a handful of studies from majority populations in East Asia, which may not represent the views of those groups in settings where they are an ethnic minority, but we felt these studies were still likely to be of relevance to these populations. There were insufficient papers across diverse countries and ethnicities to examine cross-country differences in majority populations.

Many studies were excluded as the perspectives of older participants were unidentifiable in a mixed age population. However, a brief overview of these studies indicated that the main themes were broadly similar, though views may be more diverse amongst younger populations.

All studies identified in this review were on influenza and pneumococcal vaccinations. Whilst these vaccinations are largely comparable to COVID-19 vaccination, some recommendations may not be directly applicable to the COVID-19 context. We identified some studies related to COVID-19 vaccination and minority ethnic groups in our search; however, these were all excluded due to age criteria or their focus on vaccinations in development. Recent studies that have explored COVID-19 vaccination hesitancy amongst minority ethnic groups of all ages have reported similar barriers to those we have identified: perceived risk of getting infected with COVID-19, concerns about side effects and safety, and medical mistrust amongst Hispanic and African Americans [48,49].

The studies included in our systematic review focused on minority ethnic populations, but smaller intra-group differences may have been obscured. Ethnic groups are considered to share a common ancestry, culture, and feeling of solidarity with one another [13]. There is wide variation within minority ethnic groups in country of origin, language, religion, socioeconomic characteristics, and experiences that limit the interpretations we can make from this study, but enough shared culture with regards to family structures, identity, and health beliefs to make ethnic identity relevant to health behaviours [13]. Our outcome focused on the perceptions of and attitudes towards vaccinations amongst minority ethnic groups. We acknowledge that whilst the themes cover some social determinants that influence vaccination uptake, all potential social determinants that pose barriers to vaccination in minority populations (including broader economic, social, and cultural factors) may not have emerged in the data.

The majority of studies did not report on the immigration status or citizenship of the individuals in the minority ethnic groups, which is a further limitation. In our search, we identified 2 studies examining perceptions of vaccinations in migrant groups [50,51]; however, these were excluded due to age criteria. Whilst some of the barriers in migrant groups reported in these studies mirrored those identified in this review (fear of side effects, mistrust in the healthcare system, and access), others were unique to this population (fear of facing immigration checks, lack of information in an appropriate language, and incompatibility with migrants’ religion) [50,51]. Language barriers were not identified as an emergent theme in our review. The majority of non-US studies were conducted in either the native language of the minority ethnic group and/or indicated that a translator was used. Some US studies covering Hispanic populations were conducted in Spanish or used a culturally appropriate translator; however, 11 US studies including African American and Hispanic participants either stated the study language was English or language was not specified, which is an important limitation. Clear reporting on language and use of appropriate methods is recommended for future studies.

Recommendations for practice

This review recommends, on a broad level, that efforts to provide adequate information and dispel misconceptions around vaccines by healthcare providers are fundamental to facilitating vaccine acceptance. Older adults from minority ethnic backgrounds who perceive themselves as healthy may benefit most from targeted intervention to increase uptake. Intervention is likely to be most effective through long-term trusted relationships with HCPs, tailored conversations (including family and friends), and a compassionate exploration of patient-related health beliefs. This can be supported by healthcare-system-level actions such as vaccination reminders and translated written information. Policy- and governance-level actions should focus on increasing access. This should include addressing transport issues, providing access for older adults living in rural communities, ensuring adequate availability of vaccination centres, providing home-based vaccination for frail older adults, incorporating vaccination sites into community facilities, and addressing financial barriers and existing inequity in access to preventative healthcare. Healthcare providers and policy makers should seek to tailor these recommendations to the needs and patient-related factors specific to the older ethnic minority communities they serve. Future research should address the gap in studies seeking to understand attitudes to vaccinations amongst older adults from ethnic minority groups in countries outside the US, including Europe and Australasia, and in majority groups in South Asia and Africa.