Background: The COVID-19 pandemic has highlighted health disparities affecting ethnic minority communities. There is growing concern about the lack of diversity in clinical trials. This study aimed to assess the representation of ethnic groups in UK-based COVID-19 randomised controlled trials (RCTs). Methods: A systematic review and meta-analysis were undertaken. A search strategy was developed for MEDLINE (Ovid) and Google Scholar (1st January 2020–4th May 2022). Prospective COVID-19 RCTs for vaccines or therapeutics that reported UK data separately with a minimum of 50 participants were eligible. Search results were independently screened, and data extracted into proforma. Percentage of ethnic groups at all trial stages was mapped against Office of National Statistics (ONS) statistics. Post hoc DerSimonian-Laird random-effects meta-analysis of percentages and a meta-regression assessing recruitment over time were conducted. Due to the nature of the review question, risk of bias was not assessed. Data analysis was conducted in Stata v17.0. A protocol was registered (PROSPERO CRD42021244185). Results: In total, 5319 articles were identified; 30 studies were included, with 118,912 participants. Enrolment to trials was the only stage consistently reported (17 trials). Meta-analysis showed significant heterogeneity across studies, in relation to census-expected proportions at study enrolment. All ethnic groups, apart from Other (1.7% [95% CI 1.1–2.8%] vs ONS 1%) were represented to a lesser extent than ONS statistics, most marked in Black (1% [0.6–1.5%] vs 3.3%) and Asian (5.8% [4.4–7.6%] vs 7.5%) groups, but also apparent in White (84.8% [81.6–87.5%] vs 86%) and Mixed 1.6% [1.2–2.1%] vs 2.2%) groups. Meta-regression showed recruitment of Black participants increased over time (p = 0.009). Conclusions: Asian, Black and Mixed ethnic groups are under-represented or incorrectly classified in UK COVID-19 RCTs. Reporting by ethnicity lacks consistency and transparency. Under-representation in clinical trials occurs at multiple levels and requires complex solutions, which should be considered throughout trial conduct. These findings may not apply outside of the UK setting.
Journal article
BMC Medicine
01/12/2023
21