How inclusive were UK-based randomised controlled trials of COVID-19 vaccines? A systematic review investigating enrolment of Black adults and adult ethnic minorities

• Due to the nature of the pandemic, multiple high-level RCTs were undertaken over a very short period, allowing a ‘snapshot’ of how inclusive trials were across this period and to investigate how effectively NIHR-INCLUDE guidance had been implemented.

• One limitation is that not all studies included recruited from the entirety of the UK. Some were exclusively recruiting from England (n=4), England and Wales (n=1), and England, Wales and Scotland (n=1).

• The use of 2011 census data as a comparator to determine inclusivity is likely to be an underestimate. However, it is currently considered as a gold standard in terms of providing figures for the UK adult ethnic profile, as well as for regional ethnic profiles. Therefore, its use is considered justified within the methodology.

Since its emergence in December 2019, the SARS-CoV-2 COVID-19 pandemic has resulted in over 6 million confirmed deaths worldwide [1]. This public health crisis prompted a race to develop an all-important vaccination - vital to vulnerable groups bearing the brunt of the disease burden [2]. However, it also highlighted how clinical research is not representative of the population it aims to serve [3]. This encompasses ethnic minorities, defined as any ethnic group except White British [4]. For example, Black British adults have an increased risk of death of up to 50% in comparison to their White counterparts when infected with COVID-19 [2]. Despite this, in 2020 the National Institute for Health and Care Research (NIHR) established ethnic minorities constituted only 9.26% of participants in NIHR-supported, UK COVID-19 studies, despite making up (based on NIHR figures) 13.80% of the UK population [3].

The increased threat of COVID-19 towards ethnic minorities and its adverse outcomes arises from a multitude of reasons [5]. Deprivation can exacerbate the effects of COVID-19 [6]. This encompasses house overcrowding, which is eight times more prevalent in Black African households than in White British households [7]. Proximity to infected individuals facilitates the droplet spread of COVID-19 [8], and overcrowded housing provides the appropriate setting for this. This can also be applied when considering where most ethnic minorities are based. Across England and Wales, London is the most densely populated and ethnically diverse region, where 40.2% of residents identified as either Asian, Black, Mixed or Other [9]. Popular urban areas like this create a desirable environment for the virus to spread easily, adding to the risk within these groups.

The risk associated with COVID-19 for ethnic minorities is also associated with lower vaccine uptake rates in comparison to their White counterparts [10]. In the UK, those with ethnicity defined as Black or Black British have the highest rates of vaccine hesitancy, with reasons cited for this being a lack of trust in the vaccine and worries about unknown side effects [11]. This, in turn, creates a vicious cycle, where distrust between ethnic minorities and vaccine manufacturers leads to lower rates of presentation for clinical trials and consequently, low inclusion of these groups [12].

Consequently, as minority ethnicities face greater risk with COVID-19, and that vaccine uptake is generally lower, it is important to ensure that minority ethnicities are represented within COVID-19 vaccine trials. However, underrepresentation in clinical trials compared to population estimates is an issue that not only affects ethnic minorities. Identified under-served groups include, but are not limited to: children, older people and patients with multimorbidity [13]. To address this issue, the NIHR launched the “Innovations in Clinical Trial Design and Delivery for the Under-served” roadmap, also known as the NIHR–INCLUDE [14] (often simply referred to as INCLUDE). INCLUDE is a strategic level overview to provide guidance in order to remove the lack to systemic approach in place to address inequality in research with an overall aim of making research more representative.

Its establishment in 2017 dictates that COVID-19 vaccination trials should be reflective of these guidelines. However, the degree to which this has been implemented is uncertain: no published peer-reviewed literature could be found during the literature search that focussed on the level of Black adult representation, and existing data is reflective of both adult and child ethnic minority participation across observational. interventional and non-randomised studies, as well as randomised clinical trials (RCTs) [3]. It is key to establish whether COVID-19 trials were representative to provide an indication as to whether the provision of INCLUDE [14] guidance was adequate to ensure inclusivity.

Based on the above, alongside the lack of current research on under-served ethnic representation, this study was conducted to investigate whether Black adults and adult ethnic minorities were represented in COVID-19 vaccine trials when comparing to the proportion they make up within the UK population.

A total of 637 studies were identified from a search of OVID Medline, OVID Embase, and The Cochrane Central Register of Controlled Trials. After the removal of duplicates (n=93) and articles that did not align with the eligibility criteria by title, abstract and full text (n= 537), a sum of 7 articles have been included in this review [23‐29] (Fig. 1).

This systematic review included data from 20,437 individuals across 7 COVID-19 vaccine RCTs conducted in the UK. From the studies reviewed, the calculation of Black adult representation revealed a mean of 0.59%. This was significantly different than the population level of Black adults (2.67%) as per 2011 census data. A similar trend among the adult ethnic minority population (8.94%) which was significantly lower than the census data for other ethnic minorities (16.30%). This would suggest that both Black and ethnic minority adults were under-represented in COVID-19 vaccine RCTs.

One study from this review did show progress in ethnic minority recruitment. The data from Liu et al [26] indicates that adult ethnic minorities had been over-represented by 8.97%, however, this study only aimed to include participants aged 50 and above. When also considering 2 of the 8 study sites used in this trial were based in the South-West of England with a non-white population of <5% [15] you would anticipate the proportion of those enrolled defined as White British would be potentially higher than the national average, as they are the dominant demographic within this age group and location [15]. The success of diversity in the trial study population is explained within the paper via the utilisation of a wide range of recruitment techniques, including the distribution of advertisements in public places (a recommendation suggested by INCLUDE [14]). The study protocol (26, supplement), explicitly aims to encourage recruitment of “those identifying as Black, Asian and Minority Ethnic”. It should also be noted that, of the 8 study sites, 7 were located in highly populated UK cities which are commonly associated with higher proportions of ethnic minority populations [15]. The study which failed to recruit any Black adults [25] covered 12 sites, 9 of which were in the South-West. The only site with located in an area with an above-average Black population (London) only recruited 15 (2.2%) of patients.

Outside of Liu et al, studies failed to show representative enrolment. Overall, 17.24% of participant demographics retrieved for this review were obtained from sites located in London (15 of 88 sites assessed, Supplement Figure 1), with a Black population of 13.3% [15]. The overall low representation of Black adults suggest a failure to engage with minority ethnicities in areas where there is a sufficient pool to recruit from. Likewise, a substantial proportion of participants randomised were from sites based in regions with the lowest representation of minority ethnicities. For example, South-West England (n=17) (Fig. 4), only has a Black African population of 0.9%, and an overall minority ethnic population of 4.7% [30]. Whilst it is appreciated that trial location is dependent on variables such as available staffing and funding, performing trials in areas with little minority representation will ensure they remain an under-served group.

Another key aspect of our findings identified the potential importance of inclusive recruitment techniques. Using INCLUDE guidance [14] in conjunction with utilising sites in ethnically diverse areas would be anticipated to have a positive effect on recruitment. However, none of the manuscripts reviewed made specific reference to INCLUDE. However, three of the studies specifically identified Black, Asian and minorities ethnic groups as a priority for recruitment [23, 26, 27] although none utilised patient and public involvement (PPI) groups in their study design to aid this (except for in dissemination of results [25]). Four of the studies included in our review contained insufficient English language level as part of their exclusion criteria [23, 26, 27, 29]. This can act as a deterrent for ethnic minorities - a group which also encompasses migrants, who are more likely to have a poor understanding of a host country’s health system and language [31]. Recruitment via email distribution to individuals who have already given consent to be contacted for clinical trials was identified across all 7 studies. One study primarily recruited participants from the NHS COVID-19 volunteer database [28]; however, only 0.7% of volunteers registered on this server identify as Black, African, Black British or Caribbean [32]. Recruiting primarily from this source dictates that the overall nature of results will not be generalisable as this demographic profile is not reflective of the UK population. This is also contradictory to INCLUDE guidelines [14], which instruct researchers to recruit participants through as wide a range of means as possible.

Online screening processes are associated with all 7 of our studies, where potential participants are asked to report details such as medical history. Whilst an online platform provides an appropriate tool of contact during national and local lockdowns, research has suggested that online and virtual consultations amplify existing inequalities in access to healthcare for many migrants due to a lack of digital literacy and access to technology [33]. All of this solidifies their standing as an under-served group in clinical research, in addition to leaving them at risk of misinformation around vaccines due to a lack of communication with health services.

It is appreciated that more inclusive recruitment processes are difficult to attain, especially in the context of the COVID-19 pandemic where timings for trial set-up and recruitment were vastly reduced, and sufficient public involvement was harder to achieve. Strategies such as the provision of a translator with scientific knowledge in the appropriate language can incur an additional large cost on the trial funders, as well as taking longer to perform [34]. However, language as part of eligibility criteria arguably reduces external validity by creating heterogeneity of participants, where English-speaking participants are over-represented in relation to the overall UK population and so become the standard. Therefore, appropriate steps should be taken to improve inclusivity in research, including the provision of translators to improve the inclusion of under-served groups. Furthermore, this respects NIHR guidelines stating that every eligible person who wishes to take part in research, regardless of background, should be offered the same opportunity [14]. We subsequently recommend that trials not only follow INCLUDE guidance [14], but to refer to the document and explicitly state any guidance taken within trial documentation.

For future research, it would be useful to have corresponding demographic output of excluded participants to establish whether discrepancies in inclusivity are due to design (especially exclusion criteria) or down to an overall failure to recruit. A deeper dive into urban/rural breakdown by site location is possible. Likewise, an assessment of recruitment methods and the extent to which trials follow INCLUDE guidance [14] would assist in steering future conversation on inclusivity in recruitment.

Black adults, and adult ethnic minorities, remain under-served in COVID-19 clinical research, despite being disproportionately affected by the disease. This highlights the importance of projects like INCLUDE [14], which has an invaluable role in raising awareness of under-served groups in clinical research to adequately improve the inclusivity of these groups within future vaccine RCTs in the UK as per population estimates. Without inclusive trial populations, results may not be generalisable and thus the external validity is limited and threatened. Furthermore, the exclusion of under-served groups puts them at risk of worse health outcomes due to the risk of deployment of interventions and therapeutics that are ineffective (or even harmful) to some of the population [35]. As such, more inclusive practices must be developed and widely shared in the recruitment of underserved groups, as it is key to understanding the true impact of COVID-19 on the most vulnerable groups of society, and the population as a whole.

NB: “No. minority ethnicities” denotes the number of minority ethnicity adults enrolled in each study