Temporal trends and characteristics associated with racial, ethnic, and sex representation in COVID-19 clinical trials: A systematic review and meta-analysis

Background: Early in the pandemic, extensive attention was cast on limited inclusion of historically underrepresented patient populations in COVID-19 clinical trials. How diverse representation improved following these initial reports remains unclear. Methods: PubMed, Embase and the Cochrane Library were searched (through April 2024) for US-based COVID-19 trials. Utilizing random-effects, we compared expected proportions of trial participants from racial and ethnic groups and of female sex between trials enrolling primarily in 2020 versus primarily 2021–2022. Meta-regression was performed to assess associations between trial characteristics and group representation. Results: We retrieved 157 studies comprising 198,012 participants. White (2020: 63.1% [95% CI, 60.8%–67.3%]; 2021–2022: 73.8% [95% CI, 71.5%–76.0%]) and female representation (2020: 46.1% [95% CI, 44.7%–47.4%)]; 2021–2022: 51.1% [95% CI, 49.3%–52.8%) increased across enrollment periods. Industry-sponsored trials were associated with higher White (coefficient, 0.10 [95% CI, 0.03–0.18]) and Hispanic or Latinx representation (coefficient, 0.16 [95% CI, 0.08–0.25]) and lower Asian (coefficient, −0.03 [95% CI, −0.06– –0.003]) and female representation (coefficient, −0.03 [95% CI, −0.07– –0.002]). Outpatient trials were associated with higher White (coefficient, 0.20 [95% CI, 0.13–0.26]) and female representation (coefficient, 0.16 [95% CI, 0.13–0.18]), and lower Black representation (coefficient, −0.10 [95% CI, −0.10– –0.08]). Conclusions: Despite improved female representation in COVID-19 trials over time, there was no clear increase in non-White representation. Trial characteristics such as primary sponsor, clinical setting, and intervention type correlate with representation of specific demographic groups and should be considered in future efforts to improve participant diversity.